Annals of Oncology - ESMO - PDF Free Download

Annals of Oncology Abstract Book of the 42nd ESMO Congress (ESMO 2017) Madrid, Spain 8–12 September 2017

Guest Editors:

ESMO 2017 Congress Scientific Committee

Annals of Oncology Official Journal of the European Society for Medical Oncology and the Japanese Society of Medical Oncology Editor-in-chief J.-C. Soria, Villejuif, France

S. Pignata, Naples, Italy

Associate editors

Hematological malignancies K. Tsukasaki, Saitama, Japan P. L. Zinzani, Bologna, Italy

Urogenital tumors G. Attard, Sutton, UK M. De Santis, Birmingham, UK Gastrointestinal tumors D. Arnold, Lisbon, Portugal A. Cervantes, Valencia, Spain J. Tabernero, Barcelona, Spain Breast tumors F. Andre´, Villejuif, France C. Sotiriou, Brussels, Belgium Thoracic tumors T. Mitsudomi, Osaka, Japan J. F. Vansteenkiste, Leuven, Belgium Head and neck tumors A. T. C. Chan, Shatin, Hong Kong E. Cohen, San Diego, California, USA Gynecological tumors B. Monk, Phoenix, Arizona, USA

BioTechnologies E. Mardis, St. Louis, Missouri, USA

Melanoma G. Long, Sydney, Australia

Onco-Immunology G. Coukos, Lausanne, Switzerland A. Snyder, New York, New York, USA Neuro-oncology A. Ibdaih, Paris, France

Supportive care K. Jordan, Heidelberg, Germany

Statistics M. Buyse, Brussels, Belgium

Epidemiology P. Boffetta, New York, New York, USA P. Lagiou, Athens, Greece

Molecular and surgical pathology I. I. Wistuba, Houston, Texas, USA

Sarcoma and clinical pharmacology O. Mir, Villejuif, France

Annals of Oncology online C. Ferte´, Villejuif, France

Early drug development J.-C. Soria, Villejuif, France

Industry corner: perspectives and controversies K. Dhingra, New York, New York, USA

Preclinical and experimental science T. U. E. Helleday, Stockholm, Sweden

Editors emeriti

Precision medicine C. Swanton, London, UK

F. Cavalli, Bellinzona, Switzerland D. J. Kerr, Oxford, UK J. B. Vermorken, Edegem, Belgium

Bioinformatics N. McGranahan, London, UK

Editorial board M. S. Aapro, Genolier, Switzerland M. Alsina, Barcelona, Spain Y. Ando, Nagoya, Japan P. Autier, Lyon, France H. A. Azim Jr, Brussels, Belgium I. Barnes, Oxford, UK J. Baselga, New York, USA J. Bellmunt, Boston, Massachusetts, USA B. Besse, Villejuif, France J. Beyer, Zurich, Switzerland P. Bierman, Omaha, Nebraska, USA C. Bokemeyer, Hamburg, Germany N. Boku, Tokyo, Japan E. Bria, Verona, Italy E. F. Burgess, Charlotte, USA P. G. Casali, Milan, Italy F. Ciardiello, Naples, Italy A. Comandone, Turin, Italy P. G. Corrie, Cambridge, UK G. Curigliano, Milan, Italy A. Di Leo, Prato, Italy R. Dienstmann, Barcelona, Spain T. Dorff, Los Angeles, California, USA H. Dosaka-Akita, Sapporo, Japan A. Eniu, Cluj-Napoca, Romania T. Fenske, Milwaukee, Wisconsin, USA S. Galbraith, Cambridge, UK G. Giamas, Brighton, UK R. Glynne-Jones, Northwood, UK B.-C. Goh, Singapore A. Goldhirsch, Milan, Italy P. Grimison, Sydney, Australia

S. I. Ou, Orange, California, USA X. Paoletti, Paris, France C. Pezaro, Melbourne, Australia P. Pfeiffer, Odense, Denmark S. Postel-Vinay, Villejuif, France A. Psyrri, New Haven, Connecticut, USA D. Quinn, Los Angeles, California, USA S. S. Ramalingam, Atlanta, Georgia, USA M. Reck, Grosshansdorf, Germany B. Rini, Cleveland, Ohio, USA R. Rosell, Badalona, Spain A. D. Roth, Geneva, Switzerland R. Salazar, Barcelona, Spain M. Scartozzi, Ancona, Italy N. Schmitz, Hamburg, Germany H.-J. Schmoll, Halle, Germany I. Sekine, Tsukuba, Japan Q. Shi, Rochester, Minnesota, USA A. F. Sobrero, Genoa, Italy G. Sonpavde, Birmingham, Alabama, USA S. Takahashi, Tokyo, Japan M. Toi, Kyoto, Japan R. Turck, Montclair, New Jersey, USA B. A. Van Tine, St. Louis, Missouri, USA E. Vilar, Houston, Texas, USA Y.-L. Wu, Guangzhou, China J. C.-H. Yang, Taipei, Taiwan S. Yano, Kanazawa, Japan T. Yoshino, Chiba, Japan A. X. Zhu, Boston, Massachusetts, USA

A. Grothey, Rochester, Minnesota, USA S. Halabi, Durham, North Carolina, USA D. G. Haller, Philadelphia, Pennsylvania, USA K. Hotta, Okayama, Japan I. Hyodo, Tsukuba, Japan M. Ignatiadis, Brussels, Belgium D. H. Ilson, New York, New York, USA H. Iwata, Aichi, Japan F. Janku, Houston, Texas, USA N. Katsumata, Kawasaki, Japan N. Kiyota, Kobe, Japan C. La Vecchia, Milan, Italy P. N. Jr Lara, Sacramento, California, USA J. M. Larkin, London, UK S. Loi, Melbourne, Australia S. Loibl, Neu-Isenburg, Germany F. Lordick, Leipzig, Germany Y. Loriot, Villejuif, France D. Lorusso, Milan, Italy S. Lutzker, San Francisco, California, USA T. Macarulla, Barcelona, Spain M. Martin, Madrid, Spain S. Matsui, Tokyo, Japan J. Maurel, Barcelona, Spain G. McArthur, Melbourne, Australia S. Michiels, Villejuif, France H. Minami, Kobe, Japan Y. Nishimura, Osaka-Sayama, Japan K. Nishio, Osaka-Sayama, Japan M. Ogura, Gifu, Japan A. Ohtsu, Kashiwa, Japan I. Okamoto, Fukuoka, Japan

Executive editor: Lewis Rowett Editorial office: Vanessa Marchesi, Paola Minotti Bernasconi, Giovannella Porcu, Annals of Oncology, Via Luigi Taddei 4, CH-6962 Viganello-Lugano, Switzerland R

Annals of Oncology is covered in C.A.B. International, Current Clinical Cancer, Current Contents/Clinical MedicineV, Current Contents/Life Sciences, Elsevier BIOBASE/Current Awareness in Biological Sciences, EMBASE/Excerpta Medica, IBIDS, Index Medicus/MEDLINE, The International Monitor in Oncology, Medical Documentation Service, Science Citation IndexV and Science Citation Index Expanded. R

Subscriptions

Piscataway, NJ 08854. Periodicals postage paid at Piscataway, NJ and additional mailing offices.

A subscription to Annals of Oncology comprises 12 issues plus supplements in each volume. Prices include postage, and for subscribers outside the UK delivery is by Standard Air.

Subscription records are maintained at Oxford University Press, Oxford, UK.

Annual Subscription Rate (Volume 28, 12 issues, 2017) Institutional - Academic / Non profit only Print and Online £1655.00/$3310.00/f2482.00 Online Only £1204.00/$2408.00/ f1806.00 Print Only £1525.00/$3050.00/f2288.00 Institutional - Corporate Print and Online £2069.00/$4137.00/f3103.00 Online Only £1505.00/$3010.00/f2258.00 Print Only £1906.00/$3813.00/f2860.00 Personal Print and Online £1237.00/$2476.00/f1856.00 Please note: US$ rate applies to US & Canada, Euros applies to Europe, UK£ applies to UK and Rest of World. There may be other subscription rates available, for a complete listing please visit https://academic.oup.com/annonc/subscribe Full prepayment, in the correct currency, is required for all orders. Orders are regarded as firm and payments are not refundable. Subscriptions are accepted and entered on a complete volume basis. Claims cannot be considered more than FOUR months after publication or date of order, whichever is later. All subscriptions in Canada are subject to GST. Subscriptions in the EU may be subject to European VAT. If registered, please supply details to avoid unnecessary charges. For subscriptions that include online versions, a proportion of the subscription price may be subject to UK VAT. Personal rate subscriptions are only available if payment is made by personal cheque or credit card and delivery is to a private address. The current year and two previous years’ issues are available from Oxford Journals. Previous volumes can be obtained from the Periodicals Service Company at http://www.periodicals.com/oxford.html or Periodicals Service Company, 11 Main Street, Germantown, NY 12526, USA. Email: psc@periodicals. com. Tel: (518) 537-4700. Fax: (518) 537-5899 For further information, please contact: Journals Customer Service Department, Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, UK. Email: [email protected]. Tel (and answerphone outside normal working hours): þ44 (0)1865 353907. Fax: þ44 (0)1865 353485. In the US, please contact: Journals Customer Service Department, Oxford University Press, 2001 Evans Road, Cary, NC 27513, USA. Email: [email protected]. Tel (and answerphone outside normal working hours): 800 852 7323 (toll-free in USA/Canada). Fax: 919 677 1714. In Japan, please contact: Journals Customer Service, Oxford University Press, 4-5-10-8F Shiba, Minato-ku, Tokyo 108-8386, Japan. Tel. þ81 3 5444 5858. Fax. þ81 3 3454 2929. E-mail: custserv. [email protected] DOI: For information about DOIs and to resolve them, please visit http://www. doi.org Methods of payment. (i) Cheque (payable to Oxford University Press, to Oxford University Press, Cashiers Office, Great Clarendon Street, Oxford OX2 6DP, UK) in GB£ Sterling (drawn on a UK bank), US$ Dollars (drawn on a US bank), or EUf Euros. (ii) Bank transfer to Barclays Bank Plc, Oxford Group Office, Oxford (bank sort code 20-65-18) (UK), overseas only Swift code BARC GB 22 (GB£ Sterling to account no. 70299332, IBAN GB89BARC20651870299332; US$ Dollars to account no. 66014600, IBAN GB27BARC20651866014600; EUf Euros to account no. 78923655, IBAN GB16BARC20651878923655). (iii) Credit card (Mastercard, Visa, Switch or American Express). Annals of Oncology (ISSN 0923-7534) is published monthly by Oxford University Press, Oxford, UK and distributed in the USA by Central Mailing Services c/o UKP Worldwide, 1637 Stelton Road B2, Piscataway, NJ 08854. The US annual print subscription or price is $3310.00. Airfreight and mailing in the USA by agent named Central Mailing Services c/o UKP Worldwide, 1637 Stelton Road B1-2,

Supplements, reprints and corporate sales For requests from industry and companies regarding supplements, bulk article reprints, sponsored subscriptions, translation opportunities for previously published material, and corporate online opportunities, please email: special. [email protected], fax: þ44 (0) 1865 353774 or visit http://www.oupmediainfo.com/#!/reprints-eprints.

Permissions For information on how to request permissions to reproduce articles/information from this journal, please visit https://academic.oup.com/journals/pages/access_ purchase/rights_and_permissions.

Advertising Advertising, inserts and artwork enquiries should be addressed to Advertising and Special Sales, Oxford Journals, Oxford University Press, Great Clarendon Street, Oxford, OX2 6DP, UK. Tel: þ44 (0)1865 354767; Fax þ44 (0)1865 353774; E-mail: [email protected]

Environmental and ethical policies Oxford Journals, a division of Oxford University Press, is committed to working with the global community to bring the highest quality research to the widest possible audience. Oxford Journals will protect the environment by implementing environmentally friendly policies and practices wherever possible. Please see https://academic.oup.com/journals/pages/about_us/ethical_policies for further information on environmental and ethical policies.

Notice The content of the abstracts contained in this Abstract Book is subject to an embargo. Abstracts accepted for presentation at ESMO 2017 as Proffered Paper (oral presentation), Poster Discussion and Poster will be published online on the ESMO website at 00:05 CEST on Thursday, 31 August 2017. Late-breaking abstracts and abstracts selected for the Press Programme will be made public at 00:05 CEST (Central European Summer Time) on the day of the official Congress session during which they are presented.

Disclaimer No responsibility is assumed by the organizers for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of the rapid advances in medical sciences, we recommend that independent verification of diagnoses and drug dosages should be made. Every effort has been made to faithfully reproduce the abstracts as submitted. However, no responsibility is assumed by the organizers for any omissions or misprints. C The European Society for Medical Oncology 2017 V

All rights reserved; no part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without prior written permission of the Publishers, or a licence permitting restricted copying issued in the UK by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London WIP 9HE, or in the USA by the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. Typeset by Cenveo Publisher Services, Bangalore, India. Printed by Bell and Bain Ltd., Glasgow, UK.

Annals of Oncology Official Journal of the European Society for Medical Oncology Volume 28, 2017 Supplement 5 Abstract Book of the 42nd ESMO Congress (ESMO 2017) Madrid, Spain 8–12 September 2017

Guest Editors: ESMO 2017 Congress Scientific Committee

Annals of Oncology Official Journal of the European Society for Medical Oncology Volume 28, 2017 Supplement 5 Abstract Book of the 42nd ESMO Congress (ESMO 2017) 8–12 September 2017, Madrid, Spain About ESMO About EACR ESMO Executive Board EACR Executive Board ESMO 2017 Congress Officers Acknowledgements Submitted abstracts Basic science Biomarkers Breast cancer, early stage Breast cancer, locally advanced Breast cancer, metastatic CNS tumours Developmental therapeutics Endocrine and neuroendocrine tumours Gastrointestinal tumours, colorectal Gastrointestinal tumours, non-colorectal Genitourinary tumours, prostate Genitourinary tumours, non-prostate Gynaecological cancers Haematological malignancies Head and neck cancer, excluding thyroid Health economics Immunotherapy of cancer Melanoma and other skin tumours New diagnostic tools NSCLC, early stage NSCLC, locally advanced NSCLC, metastatic Palliative care Prevention and screening Psycho-oncology Public health Sarcoma SCLC Supportive care Thoracic malignancies, other Translational research Tumour biology and pathology

v1-v21 v22-v42 v43-v67 v68-v73 v74-v108 v109-v121 v122-v141 v142-v157 v158-v208 v209-v268 v269-v294 v295-v329 v330-v354 v355-v371 v372-v394 v395-v402 v403-v427 v428-v448 v449-v452 v453-v456 v457-v459 v460-v496 v497-v501 v502-v506 v507-v510 v511-v520 v521-v538 v539-v542 v543-v567 v568-v572 v573-v594 v595-v604

Note: Abstract suffixes “O” indicates a submitted abstract accepted for Proffered Paper (oral) presentation “PD” indicates a submitted abstract accepted for Poster Discussion presentation “P” indicates a submitted abstract accepted for Poster Presentation “TiP” indicates a submitted abstract accepted for Poster Presentation, Trial in Progress _PR indicates a submitted abstract selected for inclusion in the ESMO Press Programme Volume 28 | Supplement 5 | September 2017

The European Society for Medical Oncology (ESMO) ESMO is the leading professional organisation for medical oncology, with the overarching goal of improving outcomes for cancer patients everywhere. We are the society of reference for oncology education and information, and are committed to supporting our members to develop and advance in a fast-evolving professional environment. Founded in 1975, ESMO has European roots with a global reach: we welcome oncology professionals from around the world. We are a home for all oncology stakeholders, connecting professionals with diverse expertise and experience, and speaking with one voice for our discipline. Our education and information resources support an integrated multi-professional approach to cancer care, from a medical oncology perspective. We seek to erase boundaries in cancer care, whether between countries or specialities, and pursue our mission across oncology, worldwide. The ESMO community brings together more than 16,000 oncology professionals from over 130 countries. Drawing on 40 years of experience and around 500 expert committee members, ESMO serves its members and the oncology community through: • • • • •

Post-graduate oncology education and training Career development and leadership training for the next generations of oncologists International congresses and workshops to share expertise and best practice, learn about the most up-to-date scientiﬁc advances, and connect with colleagues in related disciplines. Continuously reviewed, evidence-based standards for cancer care in Europe Advocacy and consultation to foster a favourable environment for scientiﬁc research

Cancer care is rapidly becoming more integrated and more specialised; whether their field is research, diagnosis, treatment, care, or advocacy, oncology professionals need to both build their specialist knowledge and connect with the best practitioners in other disciplines worldwide. ESMO membership makes this possible. Please visit esmo.org to learn more. Across Oncology. Worldwide.

The European Association for Cancer Research (EACR) The European Association for Cancer Research is a professional membership society for cancer researchers with more than 10,000 members worldwide. The EACR was founded in 1968 and has one guiding aim: ‘The advancement of cancer research for the public benefit’. Membership is open to anyone actively working or studying in cancer research. Our members work across the full spectrum of the field, from basic through to translational and clinical research, and range from postgraduate students to winners of the Nobel Prize. Researchers who are members of one of the 14 EACR affiliated ‘National Societies’ automatically become members of the EACR as part of the wider benefits of belonging to their national society. The membership fee for active members is just 40 Euros per annum or 120 Euros for 4 years, and special reduced membership fees are available to postgraduate students and those with less than 4 years’ postdoctoral experience. We facilitate communication and collaboration within the cancer research community. We also set out to raise the profile of cancer research in Europe and to make the case for sustained political and economic support. We organise scientific conferences of the highest quality, open to members and non-members. Our Conference Series of small, focused meetings is highly regarded for its focus on the latest research topics and for the provision of opportunities for interaction between speakers and participants. In 2018 we will celebrate our 50th anniversary and invite you to join us at EACR25, our biennial Congress to be held in Amsterdam, Netherlands, 30 June - 03 July 2018. Find out more about the EACR at www.eacr.org

ESMO Executive Board President President-Elect Past President and Membership Committee Chair Treasurer Educational Committee Chair National Representatives and Membership Committee Chair EU Policy Committee Chair Global Policy Committee Chair Guidelines Working Group Chair Press and Media Affairs Committee Chair Board Member Board Member Board Member Board Member Board Member

Fortunato Ciardiello, Naples, Italy Josep Tabernero, Barcelona, Spain Rolf A. Stahel, Zurich, Switzerland Emile Voest, Amsterdam, Netherlands Andre´s Cervantes, Valencia, Spain Fatima Cardoso, Lisbon, Portugal Paolo G. Casali, Milan, Italy Alexandru Eniu, Cluj-Napoca, Romania George Pentheroudakis, Ioannina, Greece Solange Peters, Lausanne, Switzerland Dirk Arnold, Lisbon, Portugal Susana Banerjee, London, United Kingdom Alexander M. M. Eggermont, Villejuif, France Sumitra Thongprasert, Bangkok, Thailand Christoph Zielinski, Vienna, Austria

EACR Executive Board President President Elect Past President Secretary General Treasurer Member Member Member Member Member Member Representative for Early Career Researchers Representative for Early Career Researchers

Anton Berns, Amsterdam, Netherlands Alberto Bardelli, Turin, Italy Richard Marais, Manchester, UK Joan Seoane, Barcelona, Spain Christof von Kalle, Heidelberg, Germany Carlos Caldas, Cambridge, UK Leonor David, Porto, Portugal Caroline Dive, Manchester, UK Daniel Peeper, Amsterdam, Netherlands Yardena Samuels, Rehovot, Israel Emmy Verschuren, Helsinki, Finland Eli Pikarsky, Jerusalem, Israel Therese Sørlie, Oslo, Norway

ESMO 2017 Congress Officers ESMO and Congress President Fortunato Ciardiello, Naples, Italy Scientific Committee Co-Chairs Alberto Sobrero, Genoa, Italy (ESMO) Richard Marais, Manchester, UK (EACR) Educational Chair Andre´s Cervantes, Valencia, Spain Press Officer Solange Peters, Lausanne, Switzerland Local Officer Miguel Martin, Madrid, Spain

Scientific Committee Basic science Chairs: Anton Berns, Amsterdam, Netherlands and Christof von Kalle, Heidelberg, Germany Alberto Bardelli, Candiolo, Italy Raffaele Califano, Manchester, UK Karin de Visser, Amsterdam, Netherlands Kristian Helin, Copenhagen, Denmark David Huntsman, Vancouver, BC, Canada Clare Isacke, Cambridge, UK Richard Marais, Manchester, UK Ultan McDermott, Cambridge, UK Pier Guiseppe Pelicci, Milan, Italy David Tuveson, Cold Spring Harbor, NY, USA

Breast cancer, early stage Chair: Nadia Harbeck, Munich, Germany Herve Bonnefoi, Bordeaux, France Fatima Cardoso, Lisbon, Portugal Marco Colleoni, Milan, Italy Javier Cortes, Madrid, Spain Carlos Caldas, Cambridge, UK Carsten Denkert, Berlin, Germany Angelo di Leo, Prato, Italy Doug Easton, Cambridge, UK Michael Gnant, Vienna, Austria Jos Jonkers, Amsterdam, Netherlands Olivia Pagani, Bellinzona, Switzerland Elzbieta Senkus, Gda nsk, Poland Andrew Tutt, London, UK

Breast cancer, metastatic Chair: Fabrice Andre´, Villejuif, France Joan Albanell, Barcelona, Spain Thomas Bachelot, Lyon, France Fatima Cardoso, Lisbon, Portugal Rob Clarke, Philadelphia, PA, USA Sherene Loi, Melbourne, Australia Sibylle Loibl, Neu-Isenburg, Germany Frederique Penault-Llorca, Clermont-Ferrand, France Aleix Prat, Barcelona, Spain Roman Rouzier, Paris, France Peter Schmid, London, UK Britta Weigelt, New York, NY, USA Lucy Yates, Cambridge, UK

CNS tumours Chair: Michael Weller, Zurich, Switzerland Carmen Bala~ na, Barcelona, Spain Alba Brandes, Bologna, Italy Louis Chesler, London, UK Pim French, Rotterdam, Netherlands Emilie Le Rhun, Lille, France Roger Henriksson, Stockholm, Sweden Stefan Pfister, Hedielberg, Germany Matthias Preusser, Vienna, Austria Jo¨rg-Christian Tonn, Munich, Germany Martin van den Bent, Rotterdam, Netherlands Patrick Wen, Boston, MA, USA Wolfgang Wick, Heidelberg, Germany

Developmental therapeutics Chair: Jan Schellens, Amsterdam, Netherlands Alex A. Adjei, Rochester, MN, USA Udai Banerji, London, UK Christian Blank, Amsterdam, Netherlands Anthony Chalmers, Glasgow, UK Maria Grazia Daidone, Milan, Italy Antoine Italiano, Bordeaux, France Ulrik Lassen, Copenhagen, Denmark Ignacio Melero, Pamplona, Spain Ruth Plummer, Newcastle upon Tyne, UK Jordi Rod on, Houston, TX, USA Teresa Troiani, Naples, Italy Ian Waddell, Manchester, UK

Gastrointestinal tumours, colorectal Chair: Volker Heinemann, Munich, Germany Richard Adams, Cardiff, UK Dirk Arnold, Lisbon, Portugal Andre´s Cervantes, Valencia, Spain Alfredo Falcone, Pisa, Italy Gunnar Folprecht, Dresden, Germany Rob Glynne-Jones, Northwood, UK Thomas Gruenberger, Vienna, Austria Rienk Offringa, Heidelberg, Germany Per Pfeiffer, Odense, Denmark Owen Sansom, Glasgow, UK Julien Taieb, Paris, France Vincenzo Valentini, Rome, Italy

Gastrointestinal tumours, non-colorectal Chair: Eric Van Cutsem, Leuven, Belgium Fatima Carneiro, Porto, Portugal Stefano Cascinu, Modena, Italy David Cunningham, Sutton, UK Manuel Hidalgo, Madrid, Spain Claus Jorgensen, Manchester, UK Yoon-Koo Kang, Seoul, Republic of Korea Florian Lordick, Leipzig, Germany Christophe Mariette†, Lille, France Arnaud Roth, Geneva, Switzerland Dieter Saur, Munich, Germany Thomas Seufferlein, Ulm, Germany Marcel Verhej, Amsterdam, Netherlands Marc Ychou, Montpellier, France

Genitourinary tumours, prostate Chair: Johann de Bono, Sutton, UK Andrea Alimonti, Bellinzona, Switzerland Esther Baena, Manchester, UK Himisha Beltran, New York, NY, USA Winald Gerritsen, Nijmegen, Netherlands Silke Gillessen, St. Gallen, Switzerland Nicholas James, Birmingham, UK Vesa Kataja, Jyv€askyl€a, Finland David Lorente, Valencia, Spain David Olmos, Madrid, Spain Stephane Oudard, Paris, France Scott Tomlins, Ann Arbor, MI, USA David Waugh, Belfast, UK

Genitourinary tumours, non-prostate Chair: Bernard Escudier, Villejuif, France Cory Abate-Shen, New York, NY, USA Laurence Albiges, Villejuif, France Aris Bamias, Athens, Greece Axel Bex, Amsterdam, Netherlands Petri Bono, Helsinki, Finland Fernando Calais da Silva, Lisbon, Portugal Maria De Santis, Coventry, UK Karim Fizazi, Villejuif, France Vincent Khoo, London, UK Antonio L opez-Beltran, Lisbon, Portugal Andrea Necchi, Milan, Italy Camillo Porta, Pavia, Italy Michael Staehler, Munich, Germany

Gynaecological cancers Chair: Domenica Lorusso, Milan, Italy Frederic Amant, Leuven, Belgium Susana Banerjee, London, UK Antonio Gonzalez-Martın, Madrid, Spain Philipp Harter, Essen, Germany Mansoor Mirza, Copenhagen, Denmark Remi A Nout, Leiden, Netherlands Sandro Pignata, Naples, Italy Jamie Prat, Barcelona, Spain Isabelle Ray-Coquard, Lyon, France Ignace Vergote, Leuven, Belgium

Haematological malignancies Chair: Mariano Provencio, Madrid, Spain Christian Buske, Ulm, Germany Maria del Mar C ordoba, Madrid, Spain Andre´s Ferreri, Milan, Italy Juan Fernando Garcia, Madrid, Spain Paolo Ghia, Milan, Italy Michele Ghielmini, Bellinzona, Switzerland Jose´ Gomez Codina, Valencia, Spain Marco Ladetto, Alessandria, Italy Steven Le Gouill, Nantes, France Frederic Peyrade, Nice, France Tim Somervaille, Manchester, UK Evangelos Terpos, Athens, Greece Andreas Trumpp, Heidelberg, Germany

Head and neck cancer Chair: Jean-Pascal Machiels, Brussels, Belgium Amparo Cano, Madrid, Spain Isabel Fonseca, Lisbon, Portugal Vincent Gre´goire, Brussels, Belgium Joel Guigay, Nice, France Christophe Le Tourneau, Paris, France Lisa Licitra, Milan, Italy Marco Merlano, Cuneo, Italy Ricardo Mesia, Barcelona, Spain Sjoukje Oosting, Groningen, Netherlands Amanda Psyrri, Athens, Greece Hans-Peter Rodemann, Tu¨bingen, Germany Christian Simon, Lausanne, Switzerland

Immunotherapy of cancer Chair: Inge Marie Svane, Herlev, Denmark Mario Colombo, Milan, Italy George Coukos, Lausanne, Switzerland Je´roˆme Galon, Paris, France Helen Gogas, Athens, Greece John Haanen, Amsterdam, Netherlands Guido Kroemer, Paris, France Paul Nathan, Northwood, UK Eli Pikarsky, Jerusalem, Israel Thomas Powles, London, UK Daniel Speiser, Lausanne, Switzerland Fiona Thistlethwaite, Manchester, UK

Melanoma and other skin tumours Chair: Reinhard Dummer, Zurich, Switzerland Paolo Ascierto, Naples, Italy Boris Bastian, San Francisco, CA, USA Matthias Guckenberger, Zurich, Switzerland Paul Lorigan, Manchester, UK Salvador Martin-Algarra, Pamplona, Spain Olivier Michielin, Lausanne, Switzerland Daniel Peeper, Amsterdam, Netherlands Antoni Ribas, Los Angeles, CA, USA Caroline Robert, Villejuif, France Yardena Samuels, Rehovot, Istrael Dirk Schadendorf, Essen, Germany Alessandro Testori, Milan, Italy

NETs and endocrine tumours Chair: Michel Ducreux, Villejuif, France Eric Baudin, Villejuif, France Jerome Bertherat, Paris, France Felix Beuschlein, Munich, Germany Jaume Capdevila, Barcelona, Spain Ian Chau, Sutton, UK Rocio Garcia-Carbonero, Madrid, Spain Dik Kwekkeboom†, Rotterdam, Netherlands Marianne Pavel, Erlangen, Germany Guido Rindi, Rome, Italy Enrico Ruffini, Turin, Italy Jonathan Strosberg, Tampa, FL, USA

Non-metastatic NSCLC and other thoracic malignancies Chair: Pilar Garrido, Madrid, Spain Paul Baas, Amsterdam, Netherlands Mariano Barbacid, Madrid, Spain Anne-Marie Dingemans, Maastricht, Netherlands Marina Garassino, Milan, Italy Ce´cile Le Pe´choux, Villejuif, France Keunchil Park, Seoul, Republic of Korea Luis Paz-Ares, Madrid, Spain Sanjay Popat, London, UK Egbert Smit, Amsterdam, Netherlands Paul Van Schil, Edegem, Belgium Johan Vansteenkiste, Leuven, Belgium Emmy Verschuren, Helsinki, Finland

NSCLC, metastatic Chair: Rafal Dziadziuszko, Gdansk, Poland Benjamin Besse, Villejuif, France Odd Terje Brustugun, Oslo, Norway Caroline Dive, Manchester, UK Enriqueta Felip, Barcelona, Spain Keith Kerr, Aberdeen, UK Tony S.K. Mok, Hong Kong, China Solange Peters, Lausanne, Switzerland Martin Reck, Grosshansdorf, Germany Gaetano Rocco, Naples, Italy Charles Rudin, New York, NY, USA Suresh Senan, Amsterdam, Netherlands Lecia V. Sequist, Boston, MA, USA Jean-Charles Soria, Villejuif, France

Public health policy and health economics Chair: Paolo G. Casali, Milan, Italy Peter Boyle, Ecully, France Harry de Koning, Rotterdam, Netherlands Elisabeth de Vries, Groningen, Netherlands Alex Eniu, Cluj-Napoca, Romania Rosa Giuliani, Rome, Italy Maarten IJzerman, Enschede, Netherlands Yolande Lievens, Ghent, Belgium Jose´ Maria Martin Moreno, Valencia, Spain Richard Sullivan, London, UK Carin Uyl-de Groot, Rotterdam, Netherlands

Sarcoma Chair: Jean-Yves Blay, Lyon, France Sebastian Bauer, Essen, Germany Sylvie Bonvalot, Paris, France A. Paolo Dei Tos, Treviso, Italy George Demetri, Boston, MA, USA Mikael Eriksson, Lund, Sweden Hans Gelderblom, Leiden, Netherlands Rick Haas, Amsterdam, Netherlands Heikki Joensuu, Helsinki, Finland Robin Jones, London, UK Javier Martin Broto, Seville, Spain Poul Sorensen, Vancouver, BC, Canada Silvia Stacchiotti, Milan, Italy

Supportive and palliative care Chair: Matti Aapro, Genolier, Switzerland Andreas Charalambous, Limassol, Cyprus Pere Gascon, Barcelona, Spain Joern Herrstedt, Odense, Denmark Jane Hopkinson, Cardiff, UK Karin Jordan, Heidelberg, Germany Stein Kaasa, Oslo, Norway Bernardo Rapoport, Johannesburg, South Africa Tina Saarto, Helsinki, Finland Florian Scotte´, Paris, France Anton Snegovoy, Moscow, Russian Federation Kazuo Tamura, Fukuoka, Japan Declan Walsh, Dublin, Ireland

Translational Research Chairs: Charles Swanton, London, UK and Joan Seoane, Barcelona, Spain Gerhardt Attard, Sutton, UK Carlos Caldas, Cambridge, UK Rodrigo Dienstmann, Barcelona, Spain Caroline Dive, Manchester, UK Raffaella Giavazzi, Milan, Italy Monika Hegi, Epalinges, Switzerland Thomas Helleday, Stockholm, Sweden Johanna Joyce, Lausanne, Switzerland Ilaria Malanchi, London, UK Sergio Quezada, London, UK Sohrab Shah, Vancouver, BC, Canada Sheila Singh, Hamilton, ON, Canada Yinyin Yuan, London, UK The European Society for Medical Oncology wishes to express its appreciation and gratitude to all of the above experts for their major effort in reviewing and selecting the content of this Abstract Book.

ACKNOWLEDGEMENT ESMO gratefully acknowledges the valuable contribution of the following organisations in the ESMO 2017 Congress through their presence in the exhibition hall, organisation of satellite symposia sessions, or via sponsorship activities (as of August 2017). AbbVie, Inc.

Chiltern International Ltd

LeCancer.fr

Pﬁzer Oncology

Accelovance

Clinerion

Life Length

PharmaMar, S.A.

Advanced Accelerator Applications

Clinical Research Malaysia (CRM)

Lilly Oncology

Pharmanovia A/S

Clovis Oncology

Linical

Phillips Molecular Pathway Dx

Color Genomics

Lunit

Pierre Fabre Medicament

Cyagen Biosciences

M3 Global Research

Pivotal SL

Daiichi Sankyo Europe GmbH

Macrogenics

Polyplus-transfection

Deﬁniens AG

medac GmbH

Precision Oncology

Delcath Systems Limited

Median Technologies

Predicine

Diatech Pharmacogenetics

MediBeacon GmbH

prIME Oncology

Eisai Europe Limited

Medscape LLC

QuintilesIMS

Elsevier B.V.

Merck / Pﬁzer Alliance

RBW Consulting

EMJ - European Medical Journal

Merck

Samsung Bioepis Co., Ltd

Erytech Pharma

Molecular Health GmbH

Sandoz International GmbH

Europital

MSD

Sanoﬁ Genzyme

EUSA Pharma (UK) Ltd

Mylan

Seattle Genetics

F.Hoffmann-La Roche

NanoString Technologies

Servier

Fresenius Kabi Deutschland GmbH

Natera, Inc.

Shire

Baxter Healthcare SA

NeoGenomics

Sirtex Medical Europe GmbH

Bayer

Frontiers

Nestle Health Science

Sophia Genetics SA

Beacon Oncology Limited

Genomic Health International Sarl

Nexcelom

Sotio a.s.

Noona Healthcare Oy

Spandidos Publications

Northwest Biotherapeutics Inc.

Springer Healthcare

Novartis

Springer Nature

Novella Clinical, a Quintiles company

Streck

Advanced Targeting Systems Agatha Agendia N.V. Agilent Technologies Amgen (Europe) GmbH) ArcherDX Arog Pharmaceuticals Astellas Pharma Europe Ltd. AstraZeneca Asuragen, Inc. Axience Axon Libreria, S.L. B. Braun Melsungen AG Bavarian Nordic

Berg Health Biocartis NV Bioclinica Biosimilar Medicines, a Medicines for Europe sector group BMJ Publishing Group Ltd. Boehringer Ingelheim Braster Bristol-Myers Squibb Caris Life Sciences Cascadian Therapeutics, Inc. Catalysis S.L. Celgene International Celldex Celltrion Healthcare Co., Ltd.

HalioDx Halozyme, Inc. Helsinn Healthcare SA Histalim HTG Molecular Diagnostics IBM Watson Health

Nutricia Advanced Medical Nutrition

ID Solution

Odonate Therapeutics

ImaBiotech

OncLive

Imedex

OncoDNA s.a.

IncellDx

Oncology Central

Incyte Corporation

Oxford University Press

InnoMedica

Paxman Coolers Ltd.

Ipsen

PeerVoice

Janssen Pharmaceutica NV

PER Global, LLC

KoNECT

Peregrine Pharmaceuticals Inc.

Synlab Sysmex Inostics GmbH Takeda Oncology Tesaro Teva Pharmaceuticals Europe BV The JAMA Network Thermo Fisher Scientiﬁc Varian Medical Systems International AG Wisepress Ltd

Thank you

Annals of Oncology 28 (Supplement 5): v1–v21, 2017 doi:10.1093/annonc/mdx361

BASIC SCIENCE 1O

Wild-type KRAS mediates growth inhibition and resistance to MEK inhibitors through dimerization with mutant KRAS in lung adenocarcinoma

C. Ambrogio1, J. Kohler1, Z. Zhou2, H. Wang3, R. Paranal1, M. Capelletti1, C. Caffarra1, S. Li1, Q. Lv3, D. Santamaria4, K.D. Westover2, P.A. J€anne1 1 Medical Oncology, Dana Farber Cancer Institute, Boston, MA, USA, 2Biochemistry and Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, USA, 3 School of Life Science and Technology, Tongji University, Shanghai, China, 4U1218, ACTION Laboratory, University of Bordeaux, INSERM, Pessac, France Background: Mutations in KRAS are the most frequent RAS alterations in human cancers and the prevalent driver event in lung adenocarcinoma (LUAD). There are no effective targeted therapies for KRAS-driven LUAD and chemotherapy remains the standard of care. Small-molecule inhibitors of the MAPK pathway, one of the prominent downstream KRAS mediators, show minimal clinical activity either as single agents or in combination with chemotherapy. Recently, wild-type KRAS (KRASWT) was shown to enhance tumor fitness in KRAS mutant AML and CRC cell lines while concomitantly increasing sensitivity to MEK inhibition. We hypothesized that dimerization between KRAS proteins could be a key regulator for lung adenocarcinoma biology and determinant of treatment response. Methods: To study the role of wild-type KRAS in the context of KRAS-driven cancer cells, we used genetically inducible models of KRAS loss of heterozigosity (LOH). We developed an isogenic KRASMUT inducible system that lacks endogenous HRas/NRas but harbors conditional CREERT2-controlled KRaslox alleles. Furthermore, we reconstituted KRASWT and dimerization-deficient KRASD154Q in KRAS-driven murine and human LUAD cell lines lacking the wild-type KRAS allele and evaluated the in vitro and in vivo impact on tumor progression and response to MEK inhibition. Results: KRASWT decreased in vitro and in vivo fitness of human and murine KRAS mutant LUAD tumor cells. However, this phenotype was reverted upon MEK inhibition, with KRAS LOH cells being more sensitive than KRASWT expressing cells. Interestingly, both effects were dependent on wild-type/mutant KRAS dimerization and not observed with the dimerization-deficient KRASD154Q. We provide a mechanistic model of the ambivalent function of KRASWT, linking its tumor suppressor function with increased MEK inhibitor resistance through dimerization with mutant KRAS. Conclusions: • KRASWT affects cellular fitness in KRAS-driven LUAD • KRASWT impairs response to MEK inhibitors in KRAS-driven LUAD • KRASWT inhibitory effect is dependent on dimerization with mutant KRAS • Impaired wild-type/mutant KRAS dimerization restores sensitivity to MEK inhibitors in vivo. Legal entity responsible for the study: Dana Farber Cancer Institute Funding: Stand Up To Cancer - American Cancer Society Lung Cancer Dream Team Translational Research Grant (Grant Number: SU2C-AACR-DT17-15) Disclosure: J. Kohler: Received consultant honoraria from Boehringer Ingelheim for writing and publishing two review articles on afatinib and travel grants from Roche, Amgen and Lilly. K.D. Westover: Reports receiving a commercial research grant from Astellas Pharmaceuticals. P.A. J€anne: Stock Ownership: Gatekeeper Pharmaceuticals Advisory Role: AstraZeneca, Boehringer Ingelheim, Pfizer, Chugai Pharmaceutical, ARIAD Pharmaceuticals, Merrimack Pharmaceuticals, Roche, Genentech, Loxo Oncology, Ignyta Research Funding: Astellas Pharma, AstraZeneca. All other authors have declared no conflicts of interest.

2O

CDCP1 initiates tumorigenesis and cooperates with PTEN loss to promote senescence evasion and prostate cancer progression

A. Alajati1, J. Chen1, A. Alimonti2 Cancer, IOR - Institute of Oncology Research, Bellinzona, Switzerland, 2Molecular Oncology, IOR/IOSI, Bellinzona, Switzerland 1

Background: Elevated levels of CUB domain-containing protein 1 (CDCP1) have been reported to be associated with poor prognosis in several human malignancies, including prostate cancer. However, its oncogenic role remains unexploited. Taking an advantage of multiple cross species genetic models, we demonstrate that CDCP1 per se is an oncogene. Particularly in mice, overexpression of CDCP1 in prostatic epithelial cells initiates hyperplasia, which eventually develops high-grade prostatic intraepithelial neoplasia. The functional importance of CDCP1 in tumuorigenesis was further fortified based on the evident display of invasive and metastatic prostate tumors upon its overexpression concomitantly with loss of Pten in mouse prostates. Mechanistically, we demonstrate that CDCP1 leads to the activation of SRC that further enhances the level of the transcription factor c-Myc in vivo. In turn, enhanced Myc triggers transcriptional activation of Cyclin D1 and COUP transcription factor II (COUPTF-II) that bypasses the TGF-b-dependent checkpoint and senescence barrier driven by Pten-loss. Following on, we demonstrate that targeting CDCP1 antagonizes c-Myc expression to block tumourigenesis by reactivating cellular senescence in human prostate cancer cells. C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

Methods: CDCP1, transgenic mouse model, PTEN, prostate cancer. Results: -Conditional overexpression of CDCP1 promotes tumourigenesis in different transgenic model systems -CDCP1 cooperates with Pten-loss in driving full malignant prostate tumourigenesis -Overexpression of CDCP1 overcomes Pten- loss induced cellular senescence by activating Myc-Targeting CDCP1 induces cellular senescence and growth arrest by downregulating Myc. Conclusions: In sum, our findings highlight a crucial role for CDCP1 in 1) driving tumourigeneis in several transgenic models and 2) inducing full malignant progression of Pten-null prostate tumourigenesis by eliciting an oncogenic and tumour suppressive network that results in senescence evasion and metastasis. As CDCP1 is a cell surface protein which can be targeted by currently available monoclonal antibodies, our study also put forth a novel therapeutic strategy to target SRC and MYC tumours in metastatic human prostate cancer. Legal entity responsible for the study: IOR-Bellinzona- Andrea Alimonti Funding: None Disclosure: All authors have declared no conflicts of interest.

3PD

New targets in triple negative breast cancer: Role of oncostatin M receptor pathway

M. Mu~ noz Caffarel1, A. Araujo1, C. Lawrie1, I. Alvarez L opez2, R. Rezola3, A. Abaurrea1 Oncologia, Asociaci on Instituto Biodonostia, San Sebastian, Spain, 2Medical Oncology, Hospital Universitario Donostia, Donostia, Spain, 3Department of Pathology and Anatomy, Onkologikoa, San Sebastian, Spain

1

Background: Triple Negative (TN) breast tumours have poor prognosis, lack of targeted therapies and are often refractory to conventional chemotherapy treatments. Therefore, finding new therapeutic targets for those tumours is an unmet need with high clinical impact. In this context, Oncostatin M receptor (OSMR) is a promising therapeutic target as it is over-expressed in this tumour subtype and its activation promotes invasiveness1,2. We previously showed that OSMR is frequently copy-number gained and over-expressed in squamous cell carcinoma, where it induces migration, invasion and metastasis3,4,5. We now investigate the role of OSMR in breast cancer progression. 1. Guo L, et al. (2013) Oncogene 32: 5272–5282. 2. West NR, et al. (2014) Oncogene 33: 1485-1494 3. Caffarel MM, Coleman N. (2014) Journal of Pathology 232:386-90 4. Caffarel MM, et al (2013) Journal of Pathology 231:168-79 5. Kucia-Tran, et al. (2016) Brit J Cancer 115:212-222. Methods: To address this issue we use a wide array of tools including in vitro cell cultures and in vivo models. The expression of OSMR pathway was analysed in FFPE samples and large datasets of publicly available breast cancer samples (METABRIC, n ¼ 1462; and TCGA, n ¼ 547). Results: OSMR and its ligand Oncostatin M (OSM) are over-expressed in basal tumours, where they associate with shorter overall survival (p ¼ 0.015). While OSMR is expressed by breast cancer cells, the main source of OSM seems to be the tumour stroma, primarily cancer associated macrophages and fibroblasts. OSM treatment of breast cancer cells induces the expression of important mediators of angiogenesis and invasion. Importantly, OSMR activation accelerates tumour onset, tumour growth and metastasis in orthotopic xenografts in nude mice. Conclusions: Our results support that OSMR pathway may have an important role in the initiation and progression of breast cancer and that it could be a promising candidate for therapeutic targeting in Triple Negative Breast Cancer. OSMR could be blocked by antibody based inhibition, strategy that has had a major impact on breast cancer. Legal entity responsible for the study: Molecular Oncology group, Biodonostia Health Research Institute Funding: This work was funded by MINECO (PI15/00623) and Beca SEOM/FontVella 2015, and supported by Roche. Disclosure: All authors have declared no conflicts of interest.

4PD

GGNBP2 suppresses tumor growth and cancer stem cell load of triple negative breast cancer by controlling STAT3 pathway

J. Zhang The 3rd Department of Breast Cancer, Tianjin Medical University Cancer Institute & Hospital, Tianjin, China Background: Gametogenetin binding protein 2 (GGNBP2) is encoded in human chromosome 17q12-q23, a region known as a breast and ovarian cancer susceptibility locus. GGNBP2 has a single C2H2 zinc finger and a consensus LxxLL nuclear receptor (NR) binding motif. We have reported that GGNBP2 suppresses ERa-positive breast tumorigenesis by acting as a nuclear receptor co-repressor to restrain ERa activity.

abstracts However, the detailed molecular mechanisms of GGNBP2 and its role in triple negative breast cancer (TNBC) remain largely unclear. Methods: A human breast cancer tissue array containing 138 human breast tumor tissues were utilized to examine GGNBP2 expression in breast cancer samples by IHC. To address the potential anti-breast tumor activity of GGNBP2 in vitro, we expressed exogenous GGNBP2 in TNBC cells, including MDA-MB-231 and Cal51 cell lines. Cell proliferation and cell cycle were assessed by cell growth curve/EdU assays and flow cytometry after propidium Iodide staining. Apoptosis was determined by flow cytometry after annexin V staining, by caspase 3/7 and caspase 9 activity assays. Cancer stem cell properties were determined by expression of CD44/CD24/ALDH1 markers. The levels of phosphorylated STAT3 and total STAT3 were determined by western blot. Quantitative PCR and Western blot were carried out to evaluate the effects of GGNBP2 overexpression on STAT3 target genes, CCND1, Mcl-1, survivin, bax and bim expression. Results: GGNBP2 expression is down-regulated in TNBC cells and patient tumors and it is associated with poor patient survival. Overexpression of GGNBP2 significanly induces cell cycle G0/G1 phase arrest and apoptosis in TNBC cell lines. Expression of cancer stem cell markers also decreased in GGNBP2-overexpressed TNBC cells. GGNBP2 reduces the expression levels of CCND1, Mcl-1 and survivin, promotes the expression levels of bax and bim proteins. Importantly, overexpression of GGNBP2 inhibits STAT3 phosphorylation and STAT3 downstream garget gene expression, including CCND1, Mcl-1 and survivin. Conclusions: GGNBP2 serves as a critical nuclear negative regulator of STAT3mediated gene expression and tumorigenesis. Legal entity responsible for the study: Jin Zhang Funding: None Disclosure: All authors have declared no conflicts of interest.

5PD

The acquired resistance to the combination of the anti-EGFR cetuximab and the MEK-inhibitor refametinib in KRAS mutated colorectal cancer cell lines depends on PI3K-signalling

P.P. Vitiello1, G. Martini1, C. Cardone1, D. Ciardiello1, V. Belli1, N. Matrone1, T. Troiani1, S. Napolitano1, V. Sforza1, G. Papaccio2, V. Desiderio2, V. De Falco1, E. Giunta1, F. Morgillo1, M.R. Diadema1, P. Vitale1, N. Zanaletti1, F. Ciardiello1, E. Martinelli1 1 Clinical and Experimental Medicine “F. Magrassi”, University of Campania “Luigi Vanvitelli”, Naples, Italy, 2Experimental Medicine, University of Campania “Luigi Vanvitelli”, Naples, Italy Background: Previous studies showed that the combination of an anti-epidermal growth factor (EGFR) and a selective MEK-inhibitor displays a significant anti-tumour activity in RAS-wild type colorectal cancers (CRCs), while the same combination partially reverts anti-EGFR primary resistance in KRAS mutated colorectal cancer cell lines. However, mechanisms of resistance to this combination are still unexplored. Methods: We generated KRAS mutated CRC cell lines (HCT15 and HCT116) resistant to a combination of cetuximab (an anti-EGFR antibody) and BAY86-9766 (refametinib, a selective MEK1/2-inhibitor) after continuous exposure to increasing concentration of the drugs for 8 months. Resistant clones had an IC50 20-100-fold higher than the parental cells. We evaluated by Western Blot (WB) analysis and quantitative Reverse Transcriptase Polymerase Chain Reaction (qRT-PCR) the expression and activation status of a panel of receptor tyrosine kinases (RTKs) and intracellular transducers. We further analysed by MTT assay the sensitivity of these cetuximab-MEKi resistant (CMres) cell lines to GDC-0941 (pictilisib, a selective PI3Ka inhibitor) and afatinib (BIBW 2992, an irreversible pan-HER inhibitor) either used alone or in combination. Results: We found consistent hyperactivation of the PI3K-AKT pathway and concurrent inactivation of the MAPK pathway, coupled to the activation of multiple RTKs of the HER family such as HER2 and HER3 in resistant cells when compared to parental cells. Treatment with GDC-0941 was able to partially restore the sensitivity to the drug combination, suggesting a central role for this pathway in mediating resistance in this setting, while afatinib was not capable of reverting the resistant phenotype when used alone but showed synergistic activity when combined to GDC-0941. Conclusions: These preliminary results suggest that PI3K activation plays a central role in the acquired resistance to the combination of anti-EGFR and MEK-i. PI3K activation depends at least in part by the activation of the HER family of RTK, but it can also be activated by other receptors. In vivo experiments on mice are currently ongoing. Legal entity responsible for the study: University of Campania “L. Vanvitelli” Funding: Associazione Italiana per la Ricerca sul Cancro (AIRC) Disclosure: All authors have declared no conflicts of interest.

2 | Basic science

Annals of Oncology 6PD

FPA150, a novel B7-H4 therapeutic antibody with checkpoint blockade and ADCC activities

C.D. Kaplan1, D. Houser1, F. Kemp2, N. Nielson3, A. Hsu4, K. Legris5, G. Brattich6, H. Xiang7, A. Ahene4, U. Jeffry8, D. Bellovin2, L. Borges1 1 Immuno-Oncology Research, Five Prime Therapeutics, San Francisco, CA, USA, 2 Pharmacology, Five Prime Therapeutics, San Francisco, CA, USA, 3Flow Cytometry, Adimab, Lebanon, NH, USA, 4Bioanalytics, Five Prime Therapeutics, San Francisco, CA, USA, 5Molecular Biology, Five Prime Therapeutics, San Francisco, CA, USA, 6Protein Chemistry, Five Prime Therapeutics, San Francisco, CA, USA, 7Clinical Pharmacology, Five Prime Therapeutics, San Francisco, CA, USA, 8Toxicology, Five Prime Therapeutics, San Francisco, CA, USA Background: B7-H4, a member of the B7 family of immune modulators, negatively regulates both T cell immune responses and anti-tumor immunity. While B7-H4 is highly expressed in a range of solid tumors, expression in heathy tissues is limited. Hence, we sought to generate a therapeutic antibody that both blocks the T cell inhibitory checkpoint activity of B7-H4 and mediates potent antibody-dependent cell-mediated cytotoxicity (ADCC) against B7-H4-expressing tumor cells. Methods: Fully human B7-H4 antibodies raised by screening Adimab’s yeast-based platform were evaluated for protein and cell binding, epitope specificity, and species cross-reactivity. We assessed these B7-H4 antibodies for checkpoint blockade activity using our proprietary in vitro assay, comprised of primary human T cells and B7-H4expressing artificial antigen presenting cells. We used primary human cell-based and reporter cell line-based assays to assess the in vitro ADCC activity of B7-H4 antibodies against B7-H4-expressing target cell lines. We assessed the in vivo anti-tumor efficacy of our lead B7-H4 antibody in mice bearing syngeneic tumor cell lines engineered to display mouse B7-H4. Results: After generating a panel of B7-H4 antibodies, we found that 12 out of 41 antibodies reverse B7-H4-mediated inhibition of T cell proliferation and IFNg production in vitro. Importantly, 11 of these antibodies with checkpoint blockade activity belong to the same epitope bin, recognize the B7-H4 IgV domain, and fully cross-react with cynomolgus monkey and rodent B7-H4, suggesting that these antibodies bind and block an evolutionarily conserved functional domain. When glycoengineered for enhanced FcgRIIIa binding, selected checkpoint blockade antibodies also mediate potent ADCC activity against cells exhibiting a range of B7-H4 expression. In a murine tumor model expressing B7-H4, our selected therapeutic candidate FPA150 significantly impairs tumor growth in a dose-dependent manner. Conclusions: We generated a therapeutic candidate B7-H4 antibody, FPA150, which possesses both T cell immune checkpoint blockade and ADCC activity. We initiated IND-enabling studies and plan to file an IND application by the end of 2017. Legal entity responsible for the study: Five Prime Therapeutics Funding: Five Prime Therapeutics Disclosure: C.D. Kaplan: Currently employed by and own stock in Five Prime Therapeutics. D. Houser, A. Hsu, K. Legris, G. Brattich, H. Xiang, A. Ahene, U. Jeffry, D. Bellovin, L. Borges, F. Kemp: Currently employed by Five Prime Therapeutics N. Nielson: Currently employed by Adimab.

7P

GSKJ4, an H3K27me3 demethylase inhibitor, effectively suppresses the breast cancer stem cells

N. Yan1, Y. Cao2, F. Zhang1,3, L. Xu2 Department of Oncology, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China, 2Prevention and Cure Center of Breast Disease, Third Hospital of Nanchang, Nanchang, China, 3Department of Oncology, Suzhou Kowloon Hospital, Shanghai Jiaotong University School of Medicine, Suzhou, China

1

Background: Breast cancer stem cells (BCSCs) are responsible for breast cancer metastasis and treatment failure. Hence, eliminating BCSCs poses possibility to eradicate breast cancer. Recently, studies have been suggested that H3K27me3 is implicated in maintenance of cancer stem cells (CSCs), however, the roles of H3K27me3 in BCSCs remain poorly investigated. Hence, we aimed to explore the functionality of H3K27me3 in BCSCs. Methods: Here, we firstly determined the global level of H3K27me3 in mammospherederived cells. Then, we detected the effect of GSKJ4 in cell viability through CCK8 assay. Next, we tested the impact of GSKJ4 on BCSCs expansions with CD44þCD24- phenotype and ALDH1-positive via flowcytometry. In addition, the impact on self-renewal capacity of BCSCs were also asked using mammosphere formation assay and colony formation assay. Further, western blot and Q-PCR were conducted to explore the effect of GSKJ4 in the expression level of stemness-related markers. Finally, we determined the influence of GSKJ4 on tumorigenicity using a xenograft model and investigated the underlying mechanisms. Results: We identify H3K27me3 as a negative modulator of stemness of BCSCs and suggest GSKJ4 is a promising drug targeting BCSCs. we show that the H3K27me3 level is decreased in mammosphere-derived BCSCs. In breast cancer cells, we demonstrate that GSKJ4 could markedly inhibit the proliferation. Strikingly, we show that GSKJ4 could effectively suppress BCSCs including its expansion, self-renewal capacity, and the expression of stemness-related markers. Additionally, our xenograft model confirms that GSKJ4 is able to effectively inhibit the tumorigenicity of MDA-MB-231. Mechanistically, the inhibition effect of GSKJ4 in BCSCs is via inhibiting JMJD3 and UTX thus causing increment of H3K27me3 level, which results in suppressing stemness factors including NANOG, SOX2 and OCT4.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Conclusions: Our results provide strong supports that epigenetic modification is associated with maintenance of properties of BCSCs and reveal that GSKJ4 is capable to be a prospective agent targeting BCSCs. Legal entity responsible for the study: N/A Funding: None Disclosure: All authors have declared no conflicts of interest.

8P

The impact of rotenone-modulated oxidative stress on the survival of human breast cancer stem cells (CD24-/CD441)

S.I. Wanandi1, R.A. Syahrani2, S.W.A. Jusman1 1 Biochemistry and Molecular Biology, Faculty of Medicine, Universitas Indonesia, Jakarta, Indonesia, 2Master Program in Biomedical Sciences, Faculty of Medicine, Universitas Indonesia, Jakarta, Indonesia Background: Cancer stem cells (CSCs) have been proven to be tumorigenic and may be responsible for the resistance to chemo-radiation therapy, disease recurrence and metastasis. Chemo-radiation therapy modulates oxidative stress in cancer cells, leading to cellular adaption response including modulation of cell survival and antioxidant defense mechanisms. However, the redox status alteration of breast CSCs is not yet clearly understood. The aim of this study was to elaborate the impact of rotenone-modulated oxidative stress on the survival of human breast CSCs (CD24-/CD44þ) which might be beneficial to understand the underlying mechanism of chemo-radiation therapy resistance. Methods: Human breast CSCs (CD24-/CD44þ) and non-CSCs (CD24-/CD44-) were treated with rotenone and DMSO (vehicle) for 6 hours, respectively. The effects of rotenone on oxidative stress were assessed by analysing intracellular reactive oxygen species (ROS) level using dihydroethidium assay, as well as mRNA expression and specific activity of MnSOD antioxidant. Finally, cell survival was determined using MTS assay, as well as through analysis of survivin mRNA expression. Results: Our results showed that rotenone could not modulate the superoxide level of human breast CSCs (CD24-/CD44þ), in contrast to that of non-CSCs (CD24-/CD44-). Albeit MnSOD synthesis in human breast CSCs has been excessively enhanced following rotenone treatment, the enzyme activity was still lower than in non-CSCs. Importantly, the cell viability of CSCs was higher than that of non-CSCs, which related to the increase of survivin. Conclusions: We conclude that human breast CSCs (CD24-/CD44þ) could survive better than their counterpart non-CSCs (CD24-/CD44-) when treated with rotenone. This impact might be associated with the increase of antioxidant MnSOD expression and survivin mRNA expression. Legal entity responsible for the study: Faculty of Medicine, Universitas Indonesia Funding: None Disclosure: All authors have declared no conflicts of interest.

9P

Furthermore, knockdown KMO decreased the xenografted tumor growth of MDAMB-468 cells, suggesting its oncogenic role in TNBC. Conclusions: Our data highlight the novel and critical roles of KMO in TNBC progression and metastasis. Legal entity responsible for the study: Taipei Veterans General Hospital Funding: Taipei Veterans General Hospital Disclosure: All authors have declared no conflicts of interest.

Kynurenine-3-monooxygenase (KMO) protein promotes triple negative breast cancer progression

C-Y. Liu1, T-T. Huang1, C-H. Lee1, W-L. Wang1, H-C. Lee2, L-M. Tseng3 Oncology, Taipei Veterans General Hospital, Taipei, Taiwan, 2Department and Institute of Pharmacology, National Yang-Ming University, Taipei, Taiwan, 3Surgery, Taipei Veterans General Hospital, Taipei, Taiwan

10P

PIM1 kinase promotes cell migration via SHP2 in triple-negative breast cancer

F. Braso Maristany1, J. Quist1, C. Wells2, A. Grigoriadis1, P. Marra1, A. Tutt1 Breast Cancer Now Research Unit, Division of Cancer Studies, King’s College London, London, UK, 2Cancer Cell Biology & Imaging, Division of Cancer Studies, King’s College London, London, UK

1

Background: Triple-negative breast cancers (TNBCs) are aggressive and associated with poor prognosis. We have recently demonstrated that PIM1 regulates cell death, tumour growth and chemotherapy response in TNBC. This study aims to further explore the molecular mechanisms by which PIM1 promotes malignant phenotypes in TNBC, in particular cell migration. Methods: The HumanHT-12 v4 expression array was used to interrogate changes in gene expression upon PIM1 knockdown in TNBC cell lines. Transwell migration assays and time-lapse live-cell imaging were used to study the role of PIM1 in cell migration. To assess the morphology of TNBC cells we stained F-actin with 488-phalloidin. Phospho-kinase arrays were used to elucidate the pathway by which PIM1 may control cell migration. Results: Gene expression analysis revealed PTPN11 as the most downregulated gene upon PIM1 knockdown in TNBC cell lines. These results were validated by qRTPCR in 3 TNBC cell lines. PTPN11 encodes for the phosphatase SHP2, known to be relevant for the migration of TNBC cells. We therefore studied whether PIM1 was also required for this phenotype in TNBC cell lines. PIM1 knockdown led to a defect on 2D-transwell migration in MDA-MB-231 and SUM159 cells, similar to that observed upon SHP2 knockdown. Interestingly, SHP2 knockdown did not affect short-term cell population growth of TNBC cells, suggesting that PIM1 exerts its role in cell population growth via different mechanisms, as demonstrated previously. Upon PIM1 knockdown, MDAMB-231 showed lower motility persistence, increased circularity and a reduction of Factin filaments. To understand the common downstream targets of PIM1 and SHP2 and elucidate the pathway by which PIM1 may control cell migration, we used phospho-kinase arrays. These revealed decreased phosphorylation of PLCg1, FAK and PYK2, proteins involved in cell migration, upon either PIM1 or SHP2 knockdown. Conclusions: These data suggest that PIM1 regulates cell migration by controlling PTPN11/SHP2 expression and provide further evidence for PIM1 as a target for TNBC therapy, not only to induce apoptosis and prevent tumour growth, but also to prevent TNBC migration. Legal entity responsible for the study: King’s College London/Breast Cancer Now Funding: Breast Cancer Now Disclosure: All authors have declared no conflicts of interest.

1

Background: Triple-negative breast cancer (TNBC) remains a difficult-to-treat cancer and the biology beneath TNBC is a research interest. Tryptophan-kynurenine metabolism plays an important role in epithelial-mesenchymal transition (EMT), cancer stem cells (CSCs) and immune escape. Previous studies have focused on the expression and function of the first step and the rate-limiting enzyme in tumor cells, whereas the second step catabolic enzyme kynurenine 3-monooxygenase (KMO) was rarely addressed in tumorigenesis. Hence, we sought to investigate of the mechanism and functions of KMO in TNBC carcinogenesis. Methods: KMO gene alteration and mRNA transcripts were analyzed from the The Cancer Genome Atlas (TCGA) database. MDA-MB-231 and MDA-MB-468 TNBC cell lines were used for in vitro studies. Cell proliferation, colony formation, transwell migration/invasion assays and tumorsphere forming ability were used for functional study. Signal transduction pathways were assessed by Western blot, quantitative realtime PCR and reporter assays. The effect of KMO on tumor growth was tested in nude mice with breast cancer xenografts. Results: TCGA analysis showed high-frequency of KMO amplification alterations, which was related to poor overall survival in breast cancers. KMO transcripts were upregulated in the tumor tissues of breast cancers, especially in TNBC. The functional assays showed that ectopic KMO expression promoted tumorigenesis, including cell growth and abilities of colony formation, migration, invasion, and tumorsphere formation. Moreover, western blot analysis revealed expressions of epithelial marker E-cadherin were decreased and mesenchymal markers N-cadherin, and Twist were increased by KMO overexpression in MDA-MB-468 cells. Interestingly, the mRNA and protein levels of pluripotency genes including CD44, Nanog, Oct4, and SOX-2 were also suppressed by KMO knockdown. Data of reporter gene assay showed that the activities of Nanog, Oct4, and SOX-2 promoters were enhanced by KMO overexpression.

Volume 28 | Supplement 5 | September 2017

11P

SHP-1 agonist SC-43 enhanced the anti-tumor effect of docetaxel through suppressing p-STAT3 in triple negative breast cancer cells

K-F. Chen1, C-Y. Liu2, T-I. Chao1, P-Y. Chu3, T-T. Huang2, W-L. Wang2, C-H. Lee2, K-Y. Lau2, W-C. Tsao2, C-W. Shiau4, L-M. Tseng5 1 Medical Research, National Taiwan University Hospital, Taipei, Taiwan, 2Oncology, Taipei Veterans General Hospital, Taipei, Taiwan, 3Pathology, Show Chwan Memorial Hospital, Changhua City, Taiwan, 4Biopharmaceutical Sciences, National Yang-Ming University, Taipei, Taiwan, 5Surgery, Taipei Veterans General Hospital, Taipei, Taiwan Background: Triple negative breast cancer (TNBC) is an aggressive cancer and its prognosis remains poor. Combinational therapies are a promising strategy for enhancing treatment efficacy. Blockade of STAT3 signaling have been shown to enhance the response of cancer cells to conventional chemotherapeutic agents. Here we used a SHP-1 agonist SC-43 to dephosphorylate STAT3, thereby suppressing oncogenic STAT3 signaling, and tested it in combination with docetaxel in TNBC cells. Methods: TNBC cell lines (HCC-1937, MDA-MB-468, MDA-MB-231) were used for in vitro studies. Cell viability was examined by MTT assay. Combination index was determined using Calcusyn anaysis. Apoptosis was examined by flow cytometry and western blot. Signal transduction pathways in cells were assessed by western blot. In vivo efficacy of SC-43 in combination with docetaxel was tested in nude mice with breast cancer xenografts. Results: To exam expression of SHP-1 in clinical samples, we analyzed mRNA expression of SHP-1 gene (ptpn6) in a public TNBC dataset (The Cancer Genome Atlas, TCGA). We found that higher SHP-1 mRNA expression is associated with better overall survival in TNBC patients. Sequential combination of docetaxel and SC-43 in vitroshowed enhanced anti-proliferation and apoptosis associated with decreased p-STAT3 and decreased STAT3-downstream effector cyclin D1 in the TNBC cell lines MDA-MB-231, MDA-MB-468 and HCC-1937. Ectopic expression of STAT3 reduced

doi:10.1093/annonc/mdx361 | 3

abstracts theincreased cytotoxicity induced by the combination therapy. In addition, this sequential combination showed enhanced SHP-1 activity compared to SC-43 alone. Furthermore, siRNA against SHP-1 reduced apoptosis induced by the combination treatment. Moreover, by ectopic expression of SHP-1 mutants that caused SC-43 to lose its SHP-1 agonist capability, the SC-43-induced p-STAT3 signaling inhibition was reduced in the cells subjected to the combination treatment, suggesting SHP-1 plays a crucial role in docetaxel-SC-43-mediated TNBC cell apoptosis, Importantly, combination of docetaxel and SC-43 showed enhanced anti-tumor growth compared to singleagent therapy in MDA-MB-231 xenografted tumor mice. Conclusions: SHP-1 agonist SC-43 enhanced the anti-tumor effect of docetaxel by SHP-1 dependent STAT3 inhibition in human TNBC cells. We suggest a therapeutic potential of SHP-1 agonist in combination with docetaxel for TNBC. Legal entity responsible for the study: National Taiwan University Hospital Funding: NHRI-EX106-10608BI Disclosure: All authors have declared no conflicts of interest.

12P

Delineating the mechanisms of resistance to panHER inhibitors in HER21 breast cancer cells

O. Kutuk1, H. Basaga2, Y. Ozmay1, B. Karakas2 Department of Medical Genetics, Baskent University Faculty of Medicine Adana Uygulama Ve Arastirma Mer., Adana, Turkey, 2Molecular Biology, Genetics and Bioengineering Program, Sabanci University, Istanbul, Turkey

1

Background: Despite the increase in patient survival rates promoted by increased screening and prevention efforts, much faster tumor genome sequencing and developed smart targeted therapies, de novo or acquired chemoresistance remains to be a significant factor for treatment failure in breast cancer therapeutics. Conventional chemotherapy, radiotherapy as well as targeted therapies activate mitochondrial cell death machinery to eliminate cancer cells. BCL-2 protein family members regulate mitochondrial cell death pathway by controlling mitochondrial outer membrane permeabilization. Neratinib and dacomitinib are potent and irreversible pan-EGFR inhibitors, which block their autophosphorylation and downstream signaling. Moreover, neratinib and dacomitinib have been shown to activate cell death in HER2-overexpressing cell lines. The aim of our study was to identify molecular pathways responsible for panHER2 inhibitor resistance in HER2þ breast cancer cells. Methods: The expression of EGFR and BCL-2 protein family members was determined by immunoblotting and qPCR. CellTiter-Glo was used to measure cell viability and AnnexinV/PI staining and flow cytometer was used to evaluate apoptotic response. BH3 profiling was used to determine the apoptotic blocks and mitochondrial cell death priming in breast cancer cells. Results: Here we showed that increased MCL-1 and decreased BIM mediate resistance to neratinib in ZR-75-30 and SKBR3 cells while increased BCL-XL and BCL-2 and decreased BIM promoted neratinib resistance in BT474 cells. Cells were also crossresistant to dacomitinib. BH3 profiles of HER2þ breast cancer cells efficiently predicted anti-apoptotic protein dependence and development of resistance to panHER2 inhibitors. Adding specific ERK1/2 inhibitor SCH772984 to neratinib or dacomitinib led to increased apoptotic response in SKBR3 and ZR-75-30 cells, but we did not detect a similar response in BT-474 cells. Reactivation of ERK1/2 was primarily responsible for acquired resistance in SKBR3 and ZR-75-30 cells. Intriguingly, both ERK1/2 and Akt/NFkappaB pathways were responsible for neratinib resistance in BT474 cells. Conclusions: Our results showed that different mitochondrial apoptotic blocks mediated acquired panHER2 resistance in HER2þ breast cancer cell lines as well as highlighted the potential of BH3 profiling assay in prediction of panHER2 resistance in breast cancer cells. Legal entity responsible for the study: Ozgur Kutuk Funding: Baskent University Disclosure: All authors have declared no conflicts of interest.

13P

AXL as a potential primary and secondary trastuzumab resistance mechanism in breast cancer cells with HER2 overexpression

J.M. Cejalvo1, E. Tormo1, A. Adam1, F. Rojo2, B. Pineda1, S. Zazo2, P. Gonzalez-Alonso2, M. Sabbaghi3, E. Alonso1, A. Rovira3, J. Albanell3, B. Bermejo De Las Heras1, O. Burgues4, A. Lluch1, J.A. Perez Fidalgo1, P. Eroles1 1 Oncology, Hospital Clinico Universitario de Valencia, Valencia, Spain, 2Cancer Institute, University Hospital “Fundacion Jimenez Diaz”, Madrid, Spain, 3Medical Oncology Department, Hospital del Mar, Pompeu Fabra University, Barcelona, Spain, 4 Pathological anatomy, Hospital Clınico Universitario de Valencia, Valencia, Spain Background: Breast cancer (BC) is a heterogeneous disease, HER2þ represents between 15-30% of all subtypes. Trastuzumab (T), a monoclonal antibody able to inhibit HER2 activation, has been successfully employed in HER2 amplified tumors both in adjuvant and in metastatic settings, conferring an improvement in DFS, PFS and OS. Despite these results, many patients experience primary or secondary resistance to therapy. The mechanism of resistance is unclear, our aim is to assess AXL, a receptor tyrosine kinases (RTK) implicated in epithelial-to-mesenchymal transition, as potential mechanisms of resistance.

4 | Basic science

Annals of Oncology Methods: We used two cell lines to investigate possible mechanisms of primary and secondary resistance to T in HER2þ and hormone receptor negative BC. AU565 sensitive to T (AU565-S), and HCC1954 a primary T-resistant cell line. A third cell line with acquired resistance to T (AU565-R) was generated by treating AU565-S cells with constant dose of T (15mg/mL) for 4 months. Cell viability was estimated by MTT assay. We explored the expression of AXL by Western blot (WB) and quantitative reverse transcriptase PCR (qRT-PCR). Results: The cell viability analysis at 7 days assay confirmed AU565-S as sensible to T, HCC1954 as primary resistant and the development of a secondary resistance to T (AU565-R) (50% of increased viability from AU565-S). HER2 overexpression in all three cells lines were confirmed by WB and FISH. qRT-PCR indicated an important up-regulation of AXL at mRNA levels in AU565-R and HCC1954 compared to AU565S (p < 0.05). In the same line, WB analyses showed a significantly increase in AXL protein expression in AU565-R and HCC1954 (2.03 and 7.37 fold, respectively). Finally, a selective AXL inhibitor (TP-0903) has demonstrated reduction of viability in all cell lines and significant restoration of sensitivity to T in AU565-R (p < 0.01). Conclusions: Our results suggest: 1) AXL could be a potential mechanism of both primary and secondary resistance to T; 2) combination therapy with AXL inhibitor plus T restored T sensitivity in in vitro model with AXL overexpressed. These results merit further study and to explore this RTK as possible therapeutic targets in case of anti-HER2 treatment failure. Legal entity responsible for the study: Hospital Clinico Universitario of Valencia. Biomedical Research Institute INCLIVA Funding: None Disclosure: J.A. Perez Fidalgo: Received fees from AstraZeneca, Ipsen, Novartis, Pfizer and Roche for participation in speaker bureaus. Had travel/accomodation expenses paid/reimbursed by AstraZeneca, Roche and Sandoz. All other authors have declared no conflicts of interest.

14P

Alterations to trastuzumab-induced antibody-dependent cell-mediated cytotoxicity (T-ADCC) in a lapatinib-resistant HER21 breast cancer cell line model

N. Gaynor1, M. Guibourdenche1, B. Browne1, L. O’Driscoll2, N. O’Brien3, N. O’Donovan1, J. Crown4, D. Collins1 1 National Institute for Cellular Biotechnology, Dublin City University, Dublin, Ireland, 2 School of Pharmacy and Pharmaceutical Sciences, Panoz Institute & Trinity Biomedical Sciences Institute, Trinity College Dublin, Dublin, Ireland, 3Department of Medicine, Division of Hematology/Oncology, The University of California, Los Angeles, Los Angeles, CA, USA, 4Dept of Medical Oncology, St Vincents University Hospital, Dublin, Ireland Background: Lapatinib is a dual targeting (EGFR/HER2) small molecule tyrosine kinase inhibitor (TKI) approved for the treatment of HER2þ breast cancer. Lapatinib treatment can often lead to acquired resistance and refractory disease. There is no data available on the sensitivity of lapatinib-resistant breast cancer cells to immune cellmediated cytotoxicity. To investigate the consequences of a lapatinib resistant phenotype on the immune response, we examined T-ADCC and the expression levels of immune-related proteins in a cell line model of lapatinib-resistant HER2þ breast cancer. Methods: Lapatinib-resistant SKBR3 cells (SKBR3-Lap) were generated by continuous exposure to 250nM lapatinib for 6 months alongside untreated parental cells (SKBR3Par). FACS-based assays employing peripheral blood mononuclear cells (PBMCs) isolated from healthy volunteers were used to measure direct PBMC-mediated cytotoxicity and T-ADCC. Comparative DNA microarray studies (SK-Par vs. SK-LAP) identified differentially expressed immune-related genes which were subsequently examined at the protein level by Western blot. Results: T-ADCC was 45-50% lower in SKBR3-Lap compared to the SKBR3-Par cell line at the three effector to target cell ratios examined - 1:1 (p ¼ 0.01), 5:1 (p ¼ 0.001) and 10:1 (p ¼ 0.024). Direct immune cell-mediated cytotoxicity was low (45 years old. Methylation differences were assessed using Wilcoxon rank sum test. We selected from The Cancer Genome Atlas (TCGA) those genes regulated by significantly different methylated sites and their expression was analysed. MiRNA expression data from TCGA, METABRIC and data previously published from our group was evaluated in a meta-analysis. We then selected those target genes which expression was more affected by miRNA deregulation. Pathway enrichment analysis was performed with most relevant genes from the epigenetic study by Enrichr. Results: We detected a global hypomethylation profile in BCVY samples and hypermetilation of 502 specific CpG sites exclusive of this group of age. Hypomethylated CpG sites were regulating genes mainly involve in neuronal processes, extracellular matrix and cell communication. Whereas specific hypermethylation was located in genes related to immune system, NOTCH signalling, vesicular trafficking, DNA repair and senescence. MiRNA expression meta-analysis revealed a profile of 22 miRNAs significantly deregulated in BCVY. Pathway enrichment analysis of most affected target genes showed an involvement in neural processes, glucose metabolism, vesicular trafficking,

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology DNA repair, histone and chromatin related proteins, apoptosis, cell cycle, response to DNA damage and senescence. Conclusions: Our work highlights the presence of epigenomic profile distinctive of BCVY. Both methylation and miRNAs studies points to deregulation of pathways related to neural processes, vesicular trafficking, DNA repair and senescence. All these processes may lead to cancer development and progression, thus genes in these pathways may be potential candidates for further studies. Legal entity responsible for the study: INCLIVA Research Institute Funding: None Disclosure: S. Sanchis: Funded on a FPU predoctoral Fellowship (FPU13/04976) from MINECO, Spanish Government. G. Ribas: Funded on a Miquel Servet II contract (CPII14-00013) from CarlosIII Health Institute. M.P. Chilet: Funded by private Patients Fundation LeCado. CIBERONC is an initiative of the Carlos III Health Institute.All other authors have declared no conflicts of interest.

24P

Evaluation of cell free circulating DNA in plasma by digital PCR for early diagnosis in Peruvian breast cancer patients

A. Aguilar1, A. Murillo2, J. Ponce3, J. Araujo1, J. Pinto1, C. Vigil3, R. Fujita2, J. Buleje2 Unit of Basic and Traslational Research, Oncosalud SAC, Lima, Peru, 2Facultad de Medicina, Universidad de San Martin De Porres, Lima, Peru, 3Breast Unit, Oncosalud SAC, Lima, Peru

1

Background: New diagnostic tools can be useful and give clinical benefits, including diagnosis, prognosis, treatment and monitoring of the disease. A new non-invasive method is the study of liquid biopsies, performed in fluids containing cell-free circulating DNA (cfDNA). Analyses with digital PCR technique (dPCR) allows to establish the levels of cfDNA as well as the absolute quantification of mutant alleles with accuracy. This system scatters the sample among twenty thousand wells (microfluids on the chip), where amplification reactions occur independently and are recorded by a reader. Methods: Peripheral blood samples were obtained from breast cancer patients and healthy controls. From each sample, the cfDNAs were extracted from plasma using the MagMAXTM Cell Free DNA Isolation Kit and dPCR was done for quantification of samples. We evaluated two genes, PUM1 and RNaseP. For amplification of fragments, master mix 1X (Applied Biosystems) and PCR detection assays of both genes were combined with 1.5 ll of plasma cfDNA, dispersed in the chips and placed in a ProFlex TM thermocycler (Applied Biosystems) following the program pre-established by the manufacturer, with additional five cycles. Finally, quantification data were obtained R 3D AnalysisSuiteTM Cloud Software. Comparison of DNA concenwith QuantStudioV tration in copies per microliter and other statistical calculations were performed with InfoStat 2015. Results: Significant differences were found in the values of cfDNA between patients and controls for PUM1 (p ¼ 0.0001) and RNase P (p ¼ 0.0003). These results allowed to establish cut-off points between groups at 78,995 and 51,154 copies/uL, respectively. These values can be considered in the classification of groups for further analysis of others samples. Statistical support for the use of markers in diagnosis was also evaluated using the ROC curve that favors the PUM1 marker, with a sensitivity of 75% and a specificity of 95.2%. Conclusions: Based on the significant differences found between breast cancer patients and controls, cell free DNA is a good biomarker that can be used in the diagnostic of breast cancer. On the other hand, digital PCR has been established as a good tool to check cfDNA levels from plasma of breast cancer patients. Legal entity responsible for the study: Jose Buleje Funding: Programa Nacional para la competitividad y Productividad (Innovate Per u)No.138-PNICP-PIAP-2015, Universidad de San Martin de Porres, Oncosalud - AUNA Disclosure: All authors have declared no conflicts of interest.

25P

Breast cancer predisposing germline mutations identified by exome sequencing

E.S. Kuligina1, A.P. Sokolenko1, I.V. Bizin2, M.O. Anisimova1, A.A. Romanko1, E. Imyanitov1 1 Department of Tumor Biology, NN Petrov Research Institute of Oncology, Saint Petersburg, Russian Federation, 2Bioinformatics, St. Petersburg State Polytechnical University, Saint Petersburg, Russian Federation Background: A significant portion of hereditary predisposition to breast cancer (BC) is attributed to yet unknown factors. Russian population is characterized by surprisingly strong founder effect, therefore whole exome sequencing (WES) for a limited number of these genetically homogenous patients has a potential to identify novel BCpredisposing genes. Methods: WES was performed for 32 Russian BC cases, which demonstrated strong clinical signs of the hereditary disease (family history, BC bilaterality, young onset) and

Volume 28 | Supplement 5 | September 2017

lacked germline mutations in “canonical” BC genes (BRCA1, BRCA2, CHEK2, PALB2, and NBS1/NBN). Results: Eight patients carried potentially pathogenic mutations in BRCA1 network genes (3 truncations (BLM p.Q548*, RAD51C c.904 þ 1G>A, FANCM p.S497fs) and 5 missense mutations (FANCM (p.R100W, p.Q891P), ERCC4 p.R799W, RAD54L p.R394W, RAD50 p.D515G)). The remaining 24 patients were analyzed for the presence of rare non-silent genetic variants; in total, 15437 alleles had ExAC frequency 45 years old. Global gene expression was filtered and normalized by RMA method. After initial pre-processing we analysed expression in 3,639 mRNAs, 66,457 lncRNA and 3,271 pre-miRNA and differences were assessed using t-test. We performed a meta-analysis with gene expression data from The Cancer Genome Atlas (TCGA) for validation of results. Pathway enrichment analysis was performed by Enrichr. Results: showed a specific transcriptomic landscape in BCVY. Clariom D study revealed 134 significant mRNA with p-value< 0.05 that pointed out towards pathways related with olfactory receptors, GPCR signalling, tight junction and cell-cell communication. After meta-analysis with TCGA gene expression data and own data, 43 genes were statistically significant and 15 of those withstood FDR correction (FDR< 0.05). Among those we found PIK3CB, HOXD10, ZNF654, TMEM204, IRX5, PF4, MAGEA2 and TSR2 deregulated in BCVY compared with older women. Pathway enrichment analyses and GO search highlight pathways related to cell-cell communication, cancer processes, chemokine and PI3K signalling pathways, cell differentiation, extracellular matrix, vesicular trafficking, neuronal processes among others. Conclusions: We find the presence of a distinctive transcriptomic profile of the BCVY samples. Our study points to deregulation of pathways related to cell migration, proliferation and differentiation that promote cancer development and progression. Genes obtained in meta-analyse might be potential target genes for further studies in BCVY which could help to clarify the biological background for the development of the disease in this group of age. Legal entity responsible for the study: INCLIVA Instituto de investigaci on Funding: None Disclosure: S. Sanchis: Funded on a FPU predoctoral Fellowship (FPU13/04976) from MINECO, Spanish Government. G. Ribas: Funded on a Miquel Servet II contract (CPII14-00013) from CarlosIII Health Institute. M.P. Chilet: Funded by private Patients Fundation LeCado. CIBERONC is an initiative of the Carlos III Health Institute. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx361 | 7

abstracts 27P

Integrated system-level analyses of androgen receptor variant networks to identify novel prostate cancer-relevant genes that serve as prognostic biomarkers

F. Magani, E. Bray, S. Peacock, N. Zhao, K. Burnstein Molecular and Cellular Pharmacology, University of Miami Miller School of Medicine, Miami, FL, USA Background: Castration resistant prostate cancer (CRPC) is a rapidly progressing disease state for which there is no cure. The constitutively active androgen receptor (AR) splice variant AR-V7 represents a well-established mechanism of therapeutic resistance and disease progression. This variant lacks the AR C-terminal ligand binding domain and, as such, is not inhibited by androgen deprivation therapy. Designing high-affinity drugs to target the amino terminus of AR and AR-V7 is a major challenge due to the intrinsic disorganized structure of this region. Thus, there is an imperative need to identify novel AR-V7 hub genes in PC that may serve as novel therapeutic targets. Methods: We performed a highly robust gene expression meta-analysis on PC patient samples. We defined gene modules correlated with PC progression using a Weighted Gene-Co-expression Network Analysis (WGCNA), a powerful systems biology approach. Further, we identified AR-V7 downstream target genes using gene expression profiling and mapped the AR-V7 functional interactome for the first time using a novel high-throughput synthetic genetic array screen in yeast. Finally, we combined the results from our three independent system-level analyses with experimental data to identify hub genes that were upregulated in PC patients, upregulated by AR-V7, and that also functionally interacted with AR-V7. Results: The identified genes not only included select genes previously linked to PC, such as members of the topoisomerase and cyclin families, but also novel genes that had not been previously linked to PC progression. The identified gene-signature expression correlated with patients’ Gleason score and had a prognostic value that predicted disease free-survival at the time of patient biopsy in large independent cohorts. Conclusions: In sum, we show here an unbiased integrated system-level analysis of ARV7 networks, where we combined bioinformatic analysis of patient samples and cellbased approaches to identify new candidate genes in CRPC that may serve as novel prognostic markers and future targeted therapies. Legal entity responsible for the study: University of Miami Miller School of Medicine Funding: Sylvester Comprehensive Cancer Center Disclosure: All authors have declared no conflicts of interest.

28P

ODM-208, a novel CYP11A1-inhibitor as a therapeutic approach for the treatment of castration-resistant prostate cancer

R. Oksala1, M. Karimaa1, M. Ramela1, R. Riikonen1, R. Huhtaniemi2, P. Rummakko3, G. Wohlfahrt3, A. Vuorela3, M.V. Mustonen4, P. Kallio1 1 Research and Development, Orion Corporation Orion Pharma, Turku, Finland, 2 Medicine, University of Turku, Institute of Biomedicine, Physiology, Turku, Finland, 3 Research and Development, Orion Corporation Orion Pharma, Espoo, Finland, 4 Research and Development, Orion Corporation Finland, Espoo, Finland Background: Androgen receptor (AR) plays a central role in prostate cancer and continues to be a driver in castration-resistant prostate cancer (CRPC), with increased AR expression in most cases. Approximately half of the men with CRPC respond initially to abiraterone or enzalutamide, but most relapse within 1 to 2 years. Majority of the abiraterone and enzalutamide-resistant tumors have still high AR expression and persistent AR activity. Several precursor steroids, like testosterone (T) and dihydrotestosterone activate AR, are synthesized in adrenal glands and de novo in tumours. CYP11A1 (cytochrome p450scc) is a mitochondrial enzyme catalysing the conversion of cholesterol to pregnenolone (Preg), which is the first rate-liming step in steroid hormone biosynthesis. ODM-208 is a novel, oral, non-steroidal and selective inhibitor of CYP11A1 enzyme and suppresses the synthesis of all steroid hormones and precursors. Methods: The inhibition of CYP11A1 was measured in vitro by detecting the formation of radiolabelled isocapronic acid in a human adrenal cortex cell line (H295R), and further analysing Preg and T formation by ELISA. Inhibition of the adrenal and testicular hormone production in vivo was tested in the intact male rat assay by analysing plasma concentrations of progesterone (P), corticosterone (C) and T (with LS-MS/MS) after single oral dose of ODM-208. The tumor growth inhibition was studied by using androgen dependent VCaP cells, which were subcutaneously grafted to intact male nude mice. When tumor volumes reached on average 200 mm3, mice were castrated, and after regrowth of the tumors, the oral treatment of ODM-208 was started. Results: ODM-208 potently inhibits CYP11A1 enzyme and formation of Preg and testosterone with low nM concentrations in vitro. In male rats, clear decreases of P, C and T concentrations can be detected already after single oral administration of ODM-208. In the murine VCaP CRPC xenograft model ODM-208 significantly inhibited tumor growth. Conclusions: ODM-208 shows promising antitumor activity in preclinical CRPC models and suggests that ODM-208 may have the potential to be an effective treatment in CRPC. Clinical trial in patients with metastatic CRPC is planned to be started in the 2018. Legal entity responsible for the study: Orion Corporation Orion Pharma Orion Corporation Orion Pharma Funding: Orion Corporation Orion Pharma

8 | Basic science

Annals of Oncology Disclosure: R. Oksala, M. Karimaa, M. Ramela, R. Riikonen, P. Rummakko, G. Wohlfahrt, A. Vuorela, M.V. Mustonen, P. Kallio: Employee: Orion Corporation Orion Pharma. All other authors have declared no conflicts of interest.

29P

Determining the role of the ETS factor ELF3 in normal and malignant prostate

L.K. Archer, F.M. Frame, N.J. Maitland Deaprtment of Biology, Cancer Research Unit, University of York, York, UK Background: Aberrations in the ETS transcription factor family members are a common feature of multiple cancers including prostate cancer (PCa), such as the TMPRSS2:ERG fusion. ELF3, also known as ESE-1, is an epithelial-specific ETS transcription factor involved in regulating cell differentiation in various tissues, however, its role in the prostate is controversial. The aim of this study was to identify the function of ELF3 in normal prostate development and to explore its role in PCa. Methods: Three model systems were used: prostate cell lines, primary prostate epithelial cells cultured from patient tissue and paraffin-embedded human tissue sections. The function of ELF3 was investigated using knockdown via siRNA transfection and overexpression via lentivirus transduction. Western blots, immunofluorescence and immunohistochemistry were used to measure protein localisation and levels of expression. Other assays measured cell viability, colony forming ability and migration. Results: ELF3 expression was restricted to the basal layer of the normal prostate epithelium and was not expressed in stroma. Analysis of a prostate tissue microarray indicated that whilst ELF3 is expressed in benign prostate tissue, its expression is lost in low-grade prostate tumours and re-expressed in some more advanced tumours. ELF3 knockdown resulted in decreased migration, cell viability and did not induce stem cell characteristics, whilst promoting basal cell gene expression. ELF3 overexpression increased cell migration. ELF3 was induced in primary prostate epithelial cells following treatment with the clinically approved HDAC inhibitor Vorinostat, which can promote neuroendocrine differentiation. Conclusions: ELF3 expression correlates with the normal prostate epithelial cell differentiation hierarchy, and may have a role in advanced PCa. Analysis of total gene expression following knockdown of ELF3 will give an indication of transcriptional networks that are regulated by ELF3. In addition, a lentivirus that expresses an ELF3 mutant, which alters its localisation, will be used to assess any cytoplasmic function of ELF3. This comprehensive and clinically relevant approach will allow complete elucidation of the role of ELF3 in prostate cell differentiation and PCa. Legal entity responsible for the study: Norman Maitland Funding: Prostate Cancer UK (PCUK) registered charity Disclosure: All authors have declared no conflicts of interest.

30P

Treatment-induced hypoxia attenuates enzalutamide response and promotes resistance in pre-clinical models of prostate cancer

P.J. Maxwell1, M. Labonte1, M. McKechnie1, O. Duddy1, C. Armstrong1, C.W. Ong1, A. Zoubeidi2, J. Worthington3, D. Waugh1 1 Centre for Cancer Research and Cell Biology, Queens University Belfast, Belfast, UK, 2 Vancouver Prostate Centre, University of British Columbia, Vancover, BC, Canada, 3Axis Bioservices, Axis Bioservices, Coleraine, UK Background: Inhibition of androgen signalling remains the therapeutic mainstay in castrate-resistant prostate cancer. Retention of active AR signalling or acquisition of splice variants have been reported as mechanisms of resistance to the anti-androgen Enzalutamide. Other non-AR dependent mechanisms of resistance have also emerged including acquisition of a hypoxic microenvironment. We propose treatment-induced hypoxia and the induction of angiogenesis may define a novel mechanism of relapse to Enzalutamide. Methods: Preclinical experiments were conducted in LNCaP tumors and established human prostate cancer cell lines. Tumour growth, intra-tumoral hypoxia and blood vessel density were measured in vivo. AR expression, activation and target gene expression were measured in vitro. Effects of Enzalutamide on hypoxia-driven, disease-progressing pathways and genes of interest and the role of these genes in resistance to Enzalutamide was investigated. Results: Enzalutamide promoted persistent hypoxia in LNCaP tumours in vivo, followed by increased blood vessel density and restoration of oxygen tension (>14 days). In vitro, hypoxia increased AR expression and transcriptional activity in LNCaP cells and sustained but did not further potentiate high basal AR and ARv7 activity in 22Rv1 cells. Enzalutamide failed to attenuate the concurrent hypoxia-induced HIF-1 and NFjB signalling, resulting in up-regulation of disease-progressing genes and pathways. Administration of neutralizing antibodies to two hypoxia-regulated genes, IL-8 and VEGF prolonged Enzalutamide-mediated LNCaP tumour growth control over 28 days in vivo (p < 0.001) and re-sensitised enzalutamide-resistant LNCaP cells in vitro. Conclusions: Enzalutamide-induced hypoxia upregulates the expression of VEGF and IL-8, whose multi-model signalling effects contribute to microenvironment-promoted resistance in prostate tumours. Legal entity responsible for the study: David Waugh Funding: Prostate Cancer UK

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Disclosure: J. Worthington: Scientific director at Axis Bioservices. D. Waugh: Consulting/advisory role at Almac Diagnostics, Craigavon, Northern Ireland. All other authors have declared no conflicts of interest.

32P

Legal entity responsible for the study: D. Escors Funding: None Disclosure: All authors have declared no conflicts of interest.

Prognostic impact of KRAS mutation in cell-free DNA in patients with pancreatic cancer

M.K. Kim1, S.M. Woo2, B. Park3, K-A. Yoon4, Y.H. Kim5, J. Joo3, S.J. Park2, S-Y. Kong6 Cancer Biomedical Science, Graduate School of Cancer Science and Policy, National Cancer Center, Goyang, Republic of Korea, 2Center for Liver Cancer, National Cancer Center, Goyang, Republic of Korea, 3Biometric Research Branch, National Cancer Center, Goyang, Republic of Korea, 4College of Veterinary Medicine, Konkuk University, Seoul, Republic of Korea, 5Molecular Imaging & Therapy Branch, National Cancer Center, Goyang, Republic of Korea, 67Center for Hematologic Malignancy, National Cancer Center, Goyang, Republic of Korea

1

Background: Cell-free DNA (cfDNA) has been known to be released from tumor cells and evaluated potential biomarkers for therapeutic responses. However, the role of cfDNA in pancreatic cancer has not been well studied. Here we selected KRAS mutation which has been known common over 95% of pancreatic ductal adenocarcinoma (PDA) and evaluated applicability as a prognostic marker through the quantitative analysis of cfDNA and KRAS mutation in the patients with PDA. Methods: Total of 106 PDA patients were enrolled in the study. The concentration and fraction of KRAS mutation were measured by KRAS screening multiplex droplet digital PCR kit (Biorad, USA) in plasma samples. Results: KRAS mutation was detected in 97.4% of tissue samples and the correlation with cfDNA was 0.561 with 80.5% positivity. KRAS mutation concentration and fractional abundance showed the association with poor survival in both PFS (P 90 days. In multivariable analysis factors associated with delays in TTC were: recent year of diagnosis (2011 vs 2001 OR ¼ 1.31; 95%CI 1.03-1.67), older age (76-80 vs 66-70 OR ¼ 1.51; 95%CI 1.33-1.72), black race (OR ¼ 1.35; 95%CI 1.14-1.58), having state buy-in (as an indicator of poverty) (OR ¼ 1.27; 95%CI 1.1-1.47), comorbidities (Charlson score 1 OR ¼ 1.23;95%CI 1.09-1.37; score 2 OR ¼ 1.57; 95%CI 1.37-1.81), mastectomy (OR ¼ 1.49; 95%CI 1.33-1.67), mastectomy with immediate reconstruction (OR ¼ 1.85; 95%CI 1.37-2.48), Oncotype DX testing (OR ¼ 1.68; 95%CI 1.4-2.02), mastectomy >30 days after the initial surgery (OR ¼ 16.91; 95%CI 12.07-23.68), brachytherapy (OR ¼ 4.11; 95%CI 3.17-5.34) and whole breast radiation prior to

doi:10.1093/annonc/mdx362 | 51

abstracts chemotherapy (OR ¼ 31.9; 95%CI 28.05-36.49). After adjusting for potential confounders, patients with TTC >90 days had worse OS (HR ¼ 1.37; 95%CI 1.27-1.48) and BCSS (HR ¼ 1.34; 95%CI 1.19-1.51). Conclusions: A delay in adjuvant chemotherapy administration >90 days is associated with adverse outcomes among older breast cancer patients. Determinants of delays were sociodemographic in nature, related to patient’s characteristics and to treatment received. Every effort should be made to identify vulnerable patients and to administer chemotherapy in a timely manner. Legal entity responsible for the study: The University of Texas MD Anderson Cancer Center Funding: Susan G. Komen, CIPRIT Disclosure: All authors have declared no conflicts of interest.

174P

Adjuvant chemotherapy in pT1ab node-negative triple negative breast carcinomas: Results of a national multi-institutional retrospective study

A. De Nonneville1, A. Gonc¸alves1, C. Zemmour1, M. Cohen2, J.M. Classe3, F. Reyal4, P.E. Colombo5, E. Jouve6, S. Giard7, E. Barranger8, R. Sabatier1, F. Bertucci1, J.M. Boher9, G. Houvenaeghel2 1 Medical Oncology, Institut Paoli Calmettes, Marseille, France, 2Surgical Oncology, Institut Paoli Calmettes, Marseille, France, 3Surgical Oncology, Institut Rene´ Gauducheau, Saint Herbain, France, 4Surgical Oncology, Institut Curie, Paris, France, 5 Surgical Oncology, Val-d’Aurelle, Montpellier, France, 6Surgical Oncology, Institut Claudius Regaud, Toulouse, France, 7Surgical Oncology, Centre Oscar Lambret, Lille, France, 8Surgical Oncology, Centre Antoine Lacassagne, Nice, France, 9Department of Clinical Research and Investigation, Biostatistics and Methodology Unit, Institut Paoli Calmettes, Marseille, France Background: Triple negative breast cancers (TNBC) are considered as associated with poor outcome, but prognosis of subcentimetric, node-negative disease remains controversial and evidence that adjuvant chemotherapy (CT) is effective in these small tumors remains limited. Methods: Our objective was to investigate the impact of adjuvant CT on survival in pT1abN0M0 TNBC. Patients were retrospectively identified from a cohort of 22,475 patients who underwent primary surgery in 15 French centers between 1987 and 2013. Since rare pathological types may display very particular prognoses in these tumors, we retained only the invasive ductal carcinomas of no special type according to the last WHO classification witch is most common TNBC histologic type. End-points were disease-free survival (DFS) and metastasis-free survival (MFS). A propensity score for receiving CT was estimated using a logistic regression including age, tumor size, SBR grade, and lymphovascular invasion. Results: Of a total of 284 patients with pT1abN0M0 ductal TNBC, 144 (51%) received post-operative CT and 140 (49%) did not. Patients receiving CT had more adverse prognostic features, such as tumor size, high grade, young age, and lympho-vascular invasion. Adjuvant CT was not associated with a significant benefit for DFS (Hazard ratio, HR ¼ 0.77 [0.40-1.46]; p ¼ 0.419, Log-rank test) or MFS (HR ¼ 1.00 [0.46-2.19], p ¼ 0.997), with 5-year DFS and MFS in the group with CT vs. without of 90% [81%94%] vs. 84% [74%-90%], and 90% [81%-95%] vs. 90% [83%-95%], respectively. Results were consistent in all supportive analyses including multivariate Cox model and the use of the propensity score for adjustment and as a matching factor for case– control analyses. Conclusions: This study did not identify a significant DFS or MFS advantage for adjuvant CT in subcentimetric, node-negative ductal TNBC. Although current consensus guidelines recommend consideration of adjuvant CT in all TNBC larger than 5 mm, clinicians should carefully discuss benefit/risk ratio with patients, given the yet unproven benefits. Legal entity responsible for the study: SIRIC program (INCa-DGOS-Inserm 6038) Funding: SIRIC program (INCa-DGOS-Inserm 6038) Disclosure: All authors have declared no conflicts of interest. 175P

OHERA: A real world study of cardiac events in > 3700 patients with her2-positive early breast cancer treated with trastuzumab: Final analysis

E. Lidbrink1, J. Erfan2, E. Chmielowska3, B. Otremba4, A. Bouhlel5, S. Lauer6, M. Liste Hermoso5, E. Nu¨esch6, M. Shing7, V. Misra8 1 Karolinska Institute, Karolinska University Hospital, Stockholm, Sweden, 2Josa Andras, Teaching Hospital, Nyiregyhaza, Hungary, 3Regionalne Centrum Onkologii, Bydgoszcz, Poland and Medical University Nicolaus Copernicus, Torun, Poland, 4Onkologische Praxis, Oldenburg, Oldenburg, Germany, 5Global Pharma Development, F Hoffmann-La Roche Ltd, Basel, Switzerland, 6Biostatistics, F Hoffmann-La Roche Ltd, Basel, Switzerland, 7Genentech Inc, Clovis Oncology Inc, San Francisco, CA, USA, 8The Christie Clinic, The Christie NHS Foundation Trust, Manchester, UK Background: As of Sept 2016, > 2.2 million breast cancer (BC) patients (pts) have R ; H) in clinical trial or real world settings. Risk of received trastuzumab (HerceptinV cardiac failure in pts treated with H in real world practice may differ from that observed in a clinical trial setting.

52 | Breast cancer, early stage

Annals of Oncology Methods: OHERA (NCT01152606) is a non-interventional post-approval safety study that investigated incidence of symptomatic congestive heart failure (CHF) and cardiac death in pts with HER2-positive early BC (EBC) receiving adjuvant intravenous H (H IV) in routine clinical practice, per the EU Summary of Product Characteristics (SmPC). Pts with HER2-positive EBC (stage I–IIIb) considered for treatment with H IV per the EU SmPC were enrolled, treated and monitored per local practice. Primary endpoints were incidence of symptomatic CHF (New York Heart Association [NYHA] class II–IV) and incidence of cardiac death. Secondary endpoints included time to onset of CHF. The final analysis included pt data collected for up to 5 years or until death, loss to follow-up or consent withdrawal. Results: Pts were enrolled Aug 2007–Nov 2010 at 199 sites across 9 countries. The safety population included 3733 pts with EBC treated with H IV. Median treatment duration was 11.8 months; median follow-up was 60.1 months. Incidence of symptomatic CHF was 2.8% (n ¼ 106); including severe symptomatic CHF (NYHA class III–IV) in 1.0% (n ¼ 38) pts. Median time to onset of symptomatic CHF was 5.7 months (95% CI 5.3– 6.5) and 77 (72.6%) pts achieved CHF resolution. Incidence of cardiac death was 0.2% (n ¼ 6). 251 pts had a left ventricular ejection fraction (LVEF) drop of 10% from baseline to < 50% and 169/251 (67.3%) achieved LVEF drop resolution. Incidence of CHF was higher in pts with baseline risk factors such as pre-existing cardiac conditions, age 65 years or baseline LVEF 55%. Pts who had left-side radiotherapy at baseline did not have higher CHF incidence. Conclusions: OHERA is the largest study investigating the cardiac safety of adjuvant H IV in a real world setting to date. Final analysis results were consistent with cardiac safety results from previous adjuvant H IV clinical trials in EBC, and the baseline risk factors for CHF reported in the H IV EU SmPC. Clinical trial identification: Protocol number: BO20652/RO 45-2317 ClinicalTrials.gov NCT01152606 Legal entity responsible for the study: F Hoffmann-La Roche Ltd Funding: F. Hoffmann-La Roche Ltd., Basel, Switzerland Disclosure: A. Bouhlel, M. Liste Hermoso: Employee of F Hoffmann-La Roche Ltd. S. Lauer: Contract work for Hoffmann-La Roche Ltd. E. Nu¨esch: Stock ownership: Roche. Employee of F. Hoffmann-La Roche Ltd. M. Shing: Previous employee of Genentech Inc. & current collaboration with Genentech Inc. V. Misra: Membership of an advisory board: Amgen, Eisai, Roche. All other authors have declared no conflicts of interest.

176P

Clinical outcomes of delayed start of trastuzumab treatment in patients with early breast cancer: ml25232 study

S. Beslija1, T. Ceric1, B. Hasanbegovic1, A. Kurtovic-Kozaric2, A. Pasic1, N. Mahic1, M. Kalamujic1, A. Cardzic1, J. Alidzanovic3, I. Marjanovic4, A. Mekic-Abazovic5 1 Oncology, Clinical Center, University of Sarajevo, Sarajevo, Bosnia and Herzegovina, 2 Pathology, Cytology and Human Genetics, Clinical Centre of Sarajevo University, Sarajevo, Bosnia and Herzegovina, 3Oncology, University Clincal Center Tuzla, Tuzla, Bosnia and Herzegovina, 4Oncology, University Clinical Hospital Mostar, Mostar, Bosnia and Herzegovina, 5Oncoogy, Hematology and Radiation Unit, Cantonal Hospital Zenica, Zenica, Bosnia and Herzegovina Background: Breast cancer patients in low and middle income countries have limited access to targeted therapies such as trastuzumab. The discontinuous availability of trastuzumab created waiting lists and subsequent very delayed treatment. Since few studies have systematically analyzed possible deleterious effect of delayed trastuzumab treatment, we designed a study to investigate its consequences on overall survival and disease-free survival. Methods: This was a multicenter cohort study of HER2-positive early breast cancer patients (n ¼ 223) diagnosed between 01/05/2005 and 01/05/2010 in the Federation of Bosnia and Herzegovina. The study began in 01/01/2010, and enrollment was completed in 30/06/2012. Last follow up and cut off date for analysis was 31/03/2015. Results: A total of 223 women (median 55 years; IQR: 49-61 years) were recruited. Since 131 (59%) patients waited for > 6 months after surgery to receive trastuzumab, we categorized our patient cohort into three groups: non-waiting group (n ¼ 92; wait time < 6 months), and waiting group 1 (n ¼ 85; wait time between 7 to 12 months) and waiting group 2 (n ¼ 46; > 13 month wait). OS at 5 years in non-waiting group was 84%, compared to 72% in wait group 1 and 75% in wait group 2 (p > 0.05). DFS at 5 years in the non-wait group was 79%, compared to 65% in wait group 1, and 68% in wait group 2 (p > 0.05). Conclusions: Unfortunate and unique circumstances in developing countries have created waiting lists for trastuzumab treatment—our systematic analysis of 223 women has shown that delayed start of trastuzumab treatment does not have a statistically significant effect on clinical outcomes, but shows a trend towards worse OS and DFS for women with delayed treatment. Thus, trastuzumab treatment has a persistent benefit even when administered with delayed start. Clinical trial identification: ML25232 Legal entity responsible for the study: Roche Funding: Roche Disclosure: T. Ceric: Honoraria: Roche, Novartis, Pfizer. Consulting or Advisory Role: Roche, Novartis, Pfizer. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 177P

Effects of neratinib (N) on health-related quality of life (HRQoL) in early-stage HER21 breast cancer (BC): longitudinal analyses from the phase III ExteNET trial

S. Delaloge1, Y. Ye2, D. Cella3, M. Buyse4, A. Chan5, C. Barrios6, F.A. Holmes7, J. Mansi8, H. Iwata9, B. Ejlertsen10, B. Moy11, G. von Minckwitz12, S. Chia13, M. Gnant14, S. Smichkoska15, A. Ciceniene16, S. Moran17, A.H. Auerbach18, L. Fallowfield19, M. Martin Jimenez20 1 Medical Oncology, Institut Gustave Roussy, Villejuif, France, 2Statistics, Puma Biotechnology Inc., Los Angeles, CA, USA, 3Department of Medical Social Sciences, Northwestern University, Chicago, IL, USA, 4Drug Development, International Drug Development Institute (IDDI), San Francisco, CA, USA, 5Medical Oncology, Breast Cancer Research Centre-Western Australia and Curtin University, Perth, Australia, 6Medical Oncology, Pontiﬁcal Catholic University of Rio Grande do Sul School of Medicine, Porto Alegre, Brazil, 7Medical Oncology, Texas Oncology, Houston, TX, USA, 8Medical Oncology, Guy’s and St Thomas’ NHS Foundation Trust and Biomedical Research Centre, King’s College London, London, UK, 9Medical Oncology, Aichi Cancer Center, Chikusa-ku, Nagoya, Japan, 10Clinical Oncology, Rigshospitalet, Copenhagen, Denmark, 11Medical Oncology, Massachusetts General Hospital Cancer Center, Boston, MA, USA, 12Medical Oncology, German Breast Group, Neu-Isenburg, Germany, 13 Medical Oncology, British Columbia Cancer Agency, Vancouver, BC, Canada, 14 Surgical Oncology, Comprehensive Cancer Centre, Medical University of Vienna, Vienna, Austria, 15Medical Oncology, University Clinic for Radiotherapy and Oncology, Ss Cyril and Methodius University of Skopje, Skopje, Macedonia, 16Medical Oncology, Oncology Institute of Vilnius University, Vilnius, Lithuania, 17Clinical Research and Development, Puma Biotechnology Inc., Los Angeles, CA, USA, 18N/A, Puma Biotechnology Inc., Los Angeles, CA, USA, 19Sussex Health Outcomes Research & Education in Cancer (SHORE-C), Brighton & Sussex Medical School, University of Sussex, n on, Brighton, UK, 20Medical Oncology, Hospital General Universitario Gregorio Mara~ Madrid, Spain Background: The international, randomized, placebo (P)-controlled phase III ExteNET trial (NCT00878709) showed that N for 1 y after trastuzumab-based adjuvant therapy significantly improved 2-y invasive disease-free survival in early HER2þ BC patients (pts) (HR 0.67; 95% CI 0.50–0.91; p ¼ 0.0091) [Chan et al. Lancet Oncol 2016]. Detailed longitudinal evaluation of HRQoL was an exploratory endpoint of ExteNET. Methods: 2840 pts received N 240 mg/d or P for 1 yr. Pts completed FACT-B and EQ5D questionnaires at baseline and months (M) 1, 3, 6, 9, and 12. Changes in scores from baseline were compared between groups using ANCOVA with no imputation for missing values. Sensitivity analyses using alternative methods were applied. Changes in HRQoL scores were considered to be clinically meaningful if greater than minimal clinically important differences (MCID) reported in the literature. Results: 2407 pts were evaluable for FACT-B (N, n ¼ 1171; P, n ¼ 1236), and 2427 for EQ-5D (N, n ¼ 1186; P, n ¼ 1241). Compliance with questionnaires exceeded 85%. N was associated with decreased HRQoL scores at M1 vs P, after which between-group differences diminished (Table). They were consistently less than MCIDs, except for physical well-being (PWB) subscale at M1. BC subscale (BCS) showed small improvements with N at M3–M9, all less than MCIDs. Different sensitivity methods did not alter the results.

Table: 177P Adjusted† mean difference in change from baseline: Scale

FACT-B total TOI-PFB TOI-ESB PWB BCS EQ-5D index EQ health state

MCID range

7–8 5–6 5–6 2–3 2–3 0.09–0.10 7–10

N vs P M1

M3

M6

M9

M12

–2.9* –2.6* 0 –2.4* 0.3 –0.02* –2.7*

0.1 –0.3 1.2* –1.1* 0.7* –0 0.1

–0.6 –0.7 0.6 –1.0* 0.4* –0 –1.3*

–0.5 –0.4 0.5 –0.9* 0.6* 0 –0.7

–0.8 –1.0* 0.4 –1.1* 0.2 0 –0.4

†

For baseline score;

*Statistically signiﬁcant at p < 0.05 without adjustment for multiple testing TOI ¼ trial outcome index; PFB ¼ PWB þ functional WB þ BCS; ESB ¼ emotional WB þ social WB þ BCS.

Volume 28 | Supplement 5 | September 2017

Conclusions: N was associated with decreased HRQoL, in particular in PWB, at M1, possibly due to N-related diarrhea. Based on their small magnitude, differences observed after M1 in PWB favoring P and in BCS favoring N, may not be clinically important. Clinical trial identification: NCT00878709 Legal entity responsible for the study: Wyeth, Pfizer and Puma Biotechnology Funding: Puma Biotechnology Disclosure: S. Delaloge: Grants and personal fees from Roche and GSK. Y. Ye: Employment: Puma Biotechnology Inc. M. Buyse: Employee and shareholder of IDDI. A. Chan: Personal fees for educational meetings from Pfizer, Amgen. Non-financial support from Puma Biotechnology outside the submitted work. C. Barrios: Grants from Amgen, AstraZeneca, Boehringer Ingelheim, Novartis, Pfizer, Roche, Celgene, Sanofi, Lilly, Puma. Personal fees from GSK, Novartis, Pfizer, Roche, Eisai. B. Ejlertsen: Grants to institution from NanoString, Novartis, and Roche, outside the submitted work. Travel support for educational meetings from AstraZeneca and Celgene. G. von Minckwitz: Research funding to the institution from Amgen, Roche, Novartis, Celgene, Teva, AstraZeneca, Myriad Genetics, AbbVie and Vifor Pharma. M. Gnant: Grants from Sanofi-Aventis, Novartis, Roche, GSK, Pfizer, Smith Medical. Personal fees from Novartis, AstraZeneca, Accelsiors, Eisai. S. Moran, A.H. Auerbach: Employment and stock options: Puma Biotechnology Inc. All other authors have declared no conflicts of interest.

178P

Second interim analysis of HerSCin, a German non-interventional study of subcutaneous trastuzumab for HER2-positive early breast cancer in routine clinical practice

K. Lu¨dtke-Heckenkamp1, S. Ku¨mmel2, A. Ruf-Do¨rdelmann3, A. Distelrath4, J. Wacker5, S. Schmatloch6, S. Busch-Liles7, M. Schmidt8 1 Department of Oncology and Hematology, Niels Stensen Kliniken, Franziskus Hospital Harderberg, Georgsmarienhu¨tte, Germany, 2Cancer Center Essen, Klinikum Essen Mitte, Essen, Germany, 3Department of Gynecology and Obstetrics, Municipal Clinic Karlsruhe, Karlsruhe, Germany, 4Oncologic Centre, Clinic Fulda gAG, Fulda, Germany, 5 Gynaekologische Onkologie, Frauenklinik Bruchsal, Bruchsal, Germany, 6Senology, Elisabeth Hospital, Kassel, Germany, 7Medical Affairs, Roche Pharma AG, GrenzachWyhlen, Germany, 8Department of Obstetrics and Gynecology, University Hospital Mainz, Mainz, Germany Background: Compared with IV trastuzumab, subcutaneous trastuzumab (HSC) showed non-inferior outcomes in the HannaH trial and was preferred by patients (pts) and healthcare professionals in the PrefHer study. The ongoing HerSCin study (NCT01959386) is evaluating HSC in routine clinical practice in Germany. Methods: Pts with HER2-positive early breast cancer treated with (neo)adjuvant HSC (investigator’s chosen regimen) in routine oncology practice between Nov 2013 and Nov 2016 were eligible. Pts could be enrolled retrospectively up to 9 wks after starting HSC. Baseline characteristics, treatment, adverse events (AEs), clinical outcomes and quality of life (EORTC QLQ-C30/QLQ-BR23) data are collected prospectively. Primary efficacy endpoints are pCR rate (neoadjuvant setting) and 2-year disease-free survival (adjuvant setting). Results: At the data cut-off for the second planned interim analysis (Nov 2016), 420 of 1003 pts enrolled to date from 103 German centres had completed therapy and were eligible for analysis. The median duration of follow-up was 12.2 (range 3.3–25.8) mo. Baseline characteristics are below. The mean duration of HSC was 8.8 mo (neoadjuvant: 9.2; adjuvant: 8.7). All-grade and grade 3 AEs were reported in 63% and 15% of pts, respectively. The most common all-grade AEs were fatigue (10%), diarrhoea (9%) and arthralgia (7%). AEs led to treatment interruption/withdrawal in 48 pts (11%). Only 1 of the 4 fatal AEs was considered treatment related (cardiac/respiratory failure). The pCR rate (including carcinoma in situ) in the neoadjuvant subgroup was 60.3% (95% CI 48.5–71.2). Efficacy results in the adjuvant subgroup are not mature. Conclusions: The 60.3% pCR rate is consistent with prospective trials of IV trastuzumab and HSC. Tolerability is as expected based on results from randomised trials. HSC is an active, feasible and tolerable treatment for use in routine oncology practice as well as the clinical trial setting. Clinical trial identification: NCT01959386 Legal entity responsible for the study: Roche Pharma AG Funding: Roche Pharma AG Disclosure: S. Ku¨mmel: Membership on advisory board or board of directors: Roche Pharma AG. S. Busch-Liles: Employment: Roche Pharma AG. M. Schmidt: Membership on advisory board or board of directors: Novartis, Pfizer, Pierre-Fabre, Roche. Corporate-sponsored research: Pierre-Fabre. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx362 | 53

abstracts

Annals of Oncology

Table: 178P Parameter, No. of pts (%) Median age, years (range) ECOG performance status

Cardiac conditions at baseline HER2 status by IHC

Histological grade

Subtypea

Positive nodal status Hormone receptor status

0 1 2 Missing Arterial hypertension Coronary heart disease 0/1þ 2þ 3þ Missing 1 2 3 Missing/unknown Ductal Lobular Other ER positive PgR positive ER and PgR negative

All pts (n ¼ 420)

Neoadjuvant subgroup (n ¼ 78)

Adjuvant subgroup (n ¼ 342)

56 (20–90) 258 (61) 116 (28) 10 (2) 36 (9) 134 (32) 14 (3) 3 (1) 104 (25) 310 (74) 3 (1) 7 (2) 184 (44) 222 (53) 7 (2) 343 (82) 24 (6) 54 (13) 161 (38) 280 (67) 234 (56) 127 (30)

52 (20–77) 44 (56) 26 (33) 1 (1) 7 (9) 24 (31) 3 (4) 0 13 (17) 65 (83) 0 0 30 (38) 46 (59) 2 (3) 65 (83) 3 (4) 10 (13) 25 (32) 42 (54) 39 (50) 31 (40)

57 (27–90) 214 (63) 90 (26) 9 (3) 29 (8) 110 (32) 11 (3) 3 (1) 91 (27) 245 (72) 3 (1) 7 (2) 154 (45) 176 (51) 5 (1) 278 (81) 21 (6) 44 (13) 136 (40) 238 (70) 195 (57) 96 (28)

a

One patient (adjuvant setting) recorded as both ductal and lobular. ER¼oestrogen receptor; IHC¼immunohistochemistry; PgR¼progesterone receptor.

179P

Timing of initiation of trastuzumab (T) and long-term outcome of patients (pts) with early-stage (ES) HER2-positive (HER21) breast cancer (BrCa): Impact of neo-adjuvant (NAdj) versus adjuvant (Adj) strategy

G. Gullo1, N. Walsh2, D. Fennelly1, D. Tryphonopoulos3, J. Walshe1, K. O’Mahony4, N. Silva5, L. Hammond4, J. Ballot6, C. Quinn7, C. Buckley8, J. Crown1 1 Medical Oncology Department, St Vincents University Hospital, Dublin, Ireland, 2 National Institute for Cellular Biotechnology, Dublin City University, Dublin, Ireland, 3 Medical Oncology, Agios Savas Hospital, Athens, Greece, 4Pharmacy Department, St Vincents University Hospital, Dublin, Ireland, 5Pharmacy Department, St Vincent’s Private Hospital, Dublin, Ireland, 6Medical Oncology Clinical Research, St Vincents University Hospital, Dublin, Ireland, 7Pathology Department, St Vincents University Hospital, Dublin, Ireland, 8Medical Oncology, St Vincents University Hospital, Dublin, Ireland Background: The optimal schedule of anti-HER2 Tx for HER2þ ESBrCa with respect to chemotherapy and surgery remains undefined. We performed a retrospective analysis of a large, prospectively maintained single institution data base to study the impact of treatment schedule on clinical outcome. Methods: Our database included all pts treated with T for Stage I to III HER2þBrCa who had a minimum follow up (FU) of 3 years. Time-to-first-T (TFT) was calculated from the date of the first diagnostic breast biopsy to the date of the first T. Pts with stage N3b and N3c or inoperable disease were excluded from the study. Results: A total of 506 pts treated between October 2001 and March 2014 were included in the study. T was administered as part of AdjTx in 386 (76%) pts, and of NAdjTx in 120 (24%), 338 (67%) pts had TCH [docetaxel/CBDCA/T] or “TCH-like”, 119 (24%) pts had delayed concomitant (i.e. AC-TH)/sequential T. Median FU is 73.3 months (range 1.4-176.3). Median TFT for the overall cohort was 12 weeks (range 1.9-122.3). Median TFT was significantly shorter in the NAdj than in the Adj cohort: 4.4 vs 14 weeks [p12 weeks, respectively [p < 0.0001]. Pts with TFT >12 weeks had a significantly higher risk of recurrence [HR 1.96; p ¼ 0.008] and death [HR 2.84; p ¼ 0.006] than pts with TFT 12 weeks. Pts with positive lymph nodes (LNþ) and TFT >12 weeks had significantly higher risk of relapse [HR 2.40, p ¼ 0.001] and death [HR 2.10, p ¼ 0.024] than pts with TFT 12 weeks. However, despite NAdj pts having significantly higher rate of LN þ (75% vs 53%, p < 0.0001), DFS and OS were superimposable in the two cohorts. Pts with LNþ had superior DFS when treated with NAdjTx compared with AdjTx [p ¼ 0.006].

54 | Breast cancer, early stage

Conclusions: Our mature data indicate that timing of anti-HER2 Tx significantly affects long-term outcome and T should be initiated 12 weeks from first diagnosis of ES HER2þBrCa. The early institution of T in the NAdj cohort abolished the negative impact of LNþ, thus suggesting that this should be considered the optimal Tx strategy for ES HER2þBrCa. Legal entity responsible for the study: Giuseppe Gullo Funding: None Disclosure: All authors have declared no conflicts of interest. 180P

Adjuvant endocrine treatment: Stop or continue?

E.W. Muller1, A. van de Wouw2, D. van Dreuten1 Internal Medicine, Slingeland Hospital, Doetinchem, Netherlands, 2Internal Medicine, Vie Curie MC, Venlo, Netherlands

1

Background: There is currently a trend towards extending adjuvant endocrine treatment in higher risk breast cancer patients up 10 years. However, a trade off has to be made between persisting side effects of endocrine treatment vs a small advantage in recurrence risk. It is not well known if patients still suffer side effects after 5 years of endocrine treatment, and if these may be reversible. Therefore, we studied the change in side effects of endocrine treatment and overall quality of life during and 3 months after cessation, in patients who completed at least 5 years of treatment. Methods: We included 101 patients from 2 oncological practices who underwent curative treatment for breast cancer and whose adjuvant endocrine therapy ended between 2013 and 2016. Patients willing to cooperate filled out a questionnaire before and 3 months after stopping their endocrine therapy. Hot flashes, joint pain, muscle pain and fatigue were scored as absent, a little, severe or very severe. Overall quality of life was scored on a scale from 0 to 10. Results: 101 patients were included. Average was 61 years. Tumors were T1-T4, N0N3, M0. Most patients received tamoxifen for 2-3 years, followed by an aromataseinhibitor for 3-6 years. The main finding of this survey is that overall quality of life improved significantly after stopping endocrine therapy from 6.9 (range 4-10) to 7.7 (510) (p < 0.01 Wilcoxon paired rank test). 22 women improved 2 points. Patients who scored high on muscle aches and joint complaints improved the most. Conclusions: Even patients who completed at least five years of endocrine treatment suffer side effects up to the end of treatment. After cessation these ameliorate in many, and this improves quality of life significantly. These findings are relevant when deciding

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology on extended adjuvant endocrine treatment in individual patients. Detailed analysis will be presented. Clinical trial identification: Under Dutch law no obligations for protocol submission for this type of survey, only institutional approval. Legal entity responsible for the study: E.W. Muller Funding: None Disclosure: All authors have declared no conflicts of interest.

181P

Use and effectiveness of adjuvant ovarian function suppression (OFS) in premenopausal women with early breast cancer

A.R. Ferreira1, J. Ribeiro2, A. Mayer3, M. Brito4, A. Miranda3, J.P. Fernandes5, J.L. PassosCoelho6, L. Costa1, I. Vaz-Luıs7 1 Medical Oncology, Hospital de Santa Maria and Instituto de Medicina Molecular, Faculdade de Medicina, Universidade de Lisboa, Lisbon, Portugal, 2Medical Oncology, ogico Regional do Sul, Fundac¸~ao Champalimaud, Lisbon, Portugal, 3Registo Oncol Instituto Portugu^es de Oncologia de Lisboa Francisco Gentil, Lisbon, Portugal, 4Medical Oncology, Instituto Portugu^es de Oncologia de Lisboa Francisco Gentil, Lisbon, Portugal, 5Medical Oncology, Hospitais CUF Lisboa, Lisbon, Portugal, 6Medical Oncology, Hospital da Luz and Hospital de Beatriz Aˆngelo, Lisbon, Portugal, 7 Department of Medicine and Unit INSERM 981, Institut Gustave Roussy, Villejuif, France Background: OFS either in association with tamoxifen (TAM) or an aromatase inhibitor (AI) improved disease-free survival in young women (35) and in those premenopausal women at higher risk of recurrence. However, its survival benefit remains largely unknown. In this study we characterize real-world use of adjuvant OFS from 2006 to 2015 and analyze its overall survival (OS) impact. Methods: Retrospective observational cohort study of premenopausal women with Stage I-III hormone receptor-positive (HRþ) breast cancer treated at one of 5 large centers in Portugal and diagnosed from 2006-2015. Study outcomes were use of OFS and OS. Pearson’s Chi2 test, logistic regression and Cox proportional hazards models were used. Results: Of 1717 eligible patients, 304 (17.3%) were treated with adjuvant OFS, of which 271 (15.4%) in combination with TAM and 33 (1.9%) with AI. Baseline characteristics differed by subgroups: patients treated with OFS were younger, had larger, less differentiated (grade III 16% vs 24% for OFS), more frequently HER2 positive (14% vs. 19% for OFS) tumors, and underwent more frequently mastectomy (48% vs 57% for OFS), radiotherapy (25% vs 31% for OFS) and (neo)adjuvant chemotherapy (73% vs 79% for OFS). Adjuvant OFS was used at least since 2006 with an increase in its use from 2014 onward (16% vs 25% since 2014), particularly for the combination with AI (0.4% vs 8% since 2014). In a multivariate model, characteristics associated with use of OFS included younger age and year of diagnosis 2014 (both p < 0.001). Median time on OFS was 25 mo. (interquartile range 21-27). With a median follow-up of 38 mo. (IQR 20-66) and after controlling for age at diagnosis, staging, histologic grade, HER2 status, use of (neo)adjuvant CT, type of surgery and year of diagnosis, patients treated with OFS had a better OS when compared to those not treated with OFS (adjusted-HR 0.44, 95% CI 0.19-0.96; p ¼ 0.040). Absolute benefit at year 5 was 2.1% (93.2 [95% CI 90.8 – 94.9) vs 95.3 [95% CI 89.7 – 97.9]). Conclusions: In the real-world setting, a quarter of premenopausal women with early breast cancer were already treated with adjuvant OFS in 2014. After a median follow-up of 3 years, adjuvant OFS showed an OS benefit. Legal entity responsible for the study: Hospital de Santa Maria, Centro Hospitalar de Lisboa Norte Funding: None Disclosure: All authors have declared no conflicts of interest. 182P

A phase II randomised study of Adjuvant hypo-fractionated radiotherapy with concurrent vs sequential letrozole in postmenopausal women with hormone receptor positive breast cancer: Report of pulmonary toxicity and cosmetic outcome

R. Upadhyay1, D.N. Sharma1, P. Julka2, G.K. Rath3 Radiation Oncology, All India Institute of Medical Sciences, New Delhi, India, 2Medical Oncology, B.R. Ambedkar Institute Rotary Cancer Hospital (AIMS), New Delhi, India, 3 Radiation Oncology, B.R. Ambedkar Institute Rotary Cancer Hospital (AIMS), New Delhi, India 1

Background: The sequence of hormonal therapy with adjuvant radiation(RT) is debated because of anticipated morbidity. We conducted a phase II study to evaluate feasibility and efficacy of concurrent and sequential letrozole along-with hypo-fractionated RT(HFRT). Methods: A total of 50 Post-menopausal women with hormone-receptor positive, Stage I-III Breast cancer received adjuvant HFRT 42.5Gy/16fr/3weeks and were randomly assigned to either concurrent (arm A) or sequential letrozole (arm B). Letrozole was started 3 weeks before RT in the concurrent, and 3 weeks after RT in sequential group. Pulmonary toxicity was assessed by clinical examination, chest x-ray, pulmonary function tests and HRCT chest (if indicated) at baseline, at one and six months(m) post RT. Cosmetic outcome was reported in both arms with six parameters (Table) at 6 m post RT.

Volume 28 | Supplement 5 | September 2017

Results: A total of 48 patients(pts) were followed up for 6 m (25 in Arm A and 23 in Arm B). None of the pts developed acute pulmonary toxicities. Mean (R) FeV1 and FVC values at baseline, 1 and 6 m post RT were 1.8 l (1.6-1.9) and 2.2 l(2.1-2.4), 1.79 l(1.5-1.9) and 2.1 l(1.9-2.4) and 1.85 l(1.6-2) and 2.2 l(2-2.4) respectively, and were comparable. FeV1 and FVC remained within 80 to 120% of the baseline values in 37 pts (20 Arm A vs 17 Arm B, p ¼ 0.5). FeV1 and FVC were reduced by more than 80% at 6 m in 3 pts of Arm A and 5 pts in Arm B, (p ¼ 0.7), while this was improved by over 120% in 5 pts (2 vs 3, p ¼ 1). RTOG grade 2-3 radiation dermatitis was seen in 33 pts (15 vs 18, p ¼ 0.55) while 5 pts had grade 4 toxicity (2 vs 3, p ¼ 1). There was no treatment interruption because of toxicity.

Table: 182P Cosmetic Outcome

Breast Shrinkage Breast Hardness Breast Swelling Change in Skin appearance Self-breast assessment Photographic breast assessment

Mild change

Marked change

ARM A

ARM B

ARM A

ARM B

6 2 1 9 14 15

5 2 0 8 13 15

2 1 0 3 5 3

3 0 0 3 5 2

Overall, 18 pts had excellent cosmesis (7 vs 11, p ¼ 0.4) while 32 had good cosmesis. (18 vs 14, p ¼ 0.4).

Conclusions: HFRT along-with concurrent Letrozole is well tolerated. However, patients are being followed to assess loco-regional disease control and late toxicities. Legal entity responsible for the study: All India Institute of Medical Sciences Funding: None Disclosure: All authors have declared no conflicts of interest.

183P

Intrinsic tumor features underlying clinical subtype discordance in early breast cancer

V. Kotoula1, E. Giannoulatou2, K. Papadopoulou1, I. Tikas1, K. Manousou1, M. Bobos1, S. Lakis1, G. Lazaridis1, I. Efstratiou1, F. Zagouri1, G. Pentheroudakis1, H. Gogas1, C. Christodoulou1, A. Koutras1, A. Psyrri1, C. Papandreou3, P. Papakostas1, D. Bafaloukos1, D. Pectasides1, G. Fountzilas1 1 Data Ofﬁce, Hellenic Cooperative Oncology Group (HeCOG), Athens, Greece, 2 Computational Genomics Laboratory, Victor Chang Cardiac Research Institute, Darlinghurst, Australia, 3Department of Medical Oncology, University Hospital of Larissa, Larissa, Greece Background: Despite efforts for laboratory and method harmonization, discordant clinical subtypes for ER/PgR/HER2 that determine treatment selection for breast cancer patients in clinical practice, still pose a challenge. Methods: We investigated the clinical relevance of discordant clinical subtypes and their clinicopathological and genotype characteristics (60-gene panel) in a series of 1427 breast cancer patients treated within 4 adjuvant trials (2 in the pre- and 2 in the post-trastuzumab era; recruitment period 1997 – 2012). Treatment decisions were based on local laboratory typing; all tumors were re-typed centrally. Disease-free survival was assessed. Results: We observed 340 (23.8%) discordant tumors for ER/PgR and/or HER2, ranging from 30% in the oldest to 19% in the most recent trial (p ¼ 0.004); Cohen’s K was 0.512 for all subtypes, 0.583 for ER/PgR and 0.687 for HER2. ER/PgR discordance was associated with ER (p < 0.001) and PgR (p ¼ 0.017) heterogeneity, basal phenotype, as well as higher grade, TILs and Ki67 labeling (all p < 0.001). HER2 discordant tumors had lower HER2 gene and CEN17 copies, and lower HER2/CEN17 ratios (all p < 0.001). Triple-positive tumors were rarely (0.5%) retyped as triple-negative (TN). ER/PgR discordant tumors had mutation patterns resembling HER2þ and TN, e.g., inversed TP53 and PIK3CA mutation prevalence (p < 0.001). Mutation clustering and phylogenetic analysis distinguished between concordant ERþ/PgRþ/HER2- tumors (73% of all tumors) and all other subtypes, with strong associations between ER6/ PgR6/HER2- and ERþ/PgRþ/HER2 6 (p < 0.001). More relapses were noticed in patients with ER/PgR and HER2 negative-to-positive cases who did not receive hormonotherapy and trastuzumab (multivariate p ¼ 0.048 and p ¼ 0.016, respectively), but not in positive-to-negative cases. Conclusions: Apart from technical considerations, clinical subtype discordance may reflect the genetic background of breast cancers, which appear to evolve by deviating from the ERþ/PgRþ status. Development and reporting of phenotypic surrogates

doi:10.1093/annonc/mdx362 | 55

abstracts

Annals of Oncology

predictive of discordance is needed for increasing diagnostic accuracy and appropriate treatment selection. Legal entity responsible for the study: Hellenic Cooperative Oncology Group (HeCOG) Funding: HeCOG Disclosure: H. Gogas: Advisory or consultancy role: Amgen, Bristol-Myers Squibb, MSD, Novartis, Roche. C. Papandreou: Honoraria and/or Advisory Role: Astellas Pharmaceuticals, AstraZeneca, Janssen Pharmaceutical, Merck S.A., Roche (Hellas) S.A., Sanofi-Aventis, Pfizer Hellas S.A., Merck Sharp & Dohme. G. Fountzilas: Honoraria: AstraZeneca. Consulting or Advisory Role: Pfizer, Sanofi, Roche. Stock ownership (an immediate family member): Ariad. All other authors have declared no conflicts of interest.

184P

Distribution of genomically defined recurrence risk in luminal A and B breast tumors defined by inmunohistochemistry: A retrospective study in Spanish population

S. Perez Ramirez1, M. del Monte-Millan1, S. L opez-Tarruella1, I. Marquez-Rodas1, nez3, J.A. Garcıa-Saenz4, Y. Jerez1, F. Lobo Samper2, Y. Izarzugaza Peron2, N. Martınez Ja~ 7 Lara Alvarez n5, M. Arroyo Yustos6, M.A. no , F. Moreno Ant on4, P. Zamora Au~ E.M. Ciruelos Gil8, L. Manso Sanchez8, M.J. Echarri Gonzalez9, J.A. Guerra Martınez10, C. Jara Sanchez11, V. Valentın Maganto12, M. Martin Jimenez1 1 Medical Oncology, Hospital General Universitario Gregorio Mara~ non. Instituto de Investigacion Sanitaria Gregorio Mara~ n on (IiSGM), Madrid, Spain, 2Medical Oncology, 3 Hospital Universitario Fundaci on Jime´nez Dıaz, Madrid, Spain, Medical Oncology, Hospital Universitario Ram on y Cajal, Madrid, Spain, 4Medical Oncology, Hospital Universitario Clınico San Carlos, Madrid, Spain, 5Medical Oncology, Hospital Universitario La Paz, Madrid, Spain, 6Medical Oncology, Hospital Universitario Prıncipe de Asturias, Madrid, Spain, 7Medical Oncology, Hospital Universitario Infanta Leonor, Madrid, Spain, 8Medical Oncology, Hospital Universitario 12 de Octubre, Madrid, Spain, 9 Medical Oncology, Hospital Universitario Severo Ochoa, Madrid, Spain, 10Medical Oncology, Hospital Universitario de Fuenlabrada, Madrid, Spain, 11Medical Oncology, Hospital Universitario Fundaci on Alcorc on, Universidad Rey Juan Carlos, Madrid, Spain, 12 Former Regional Oncology Coordinator, Ministry of Health, Madrid, Spain Background: Semiquantitative inmunohistochemical (IHC) expression of progesterone receptor (PR) adds prognostic value to the current IHC-based luminal A (LA) definition, such that patients (pts) with LA (Ki67 20%) tumors can be V V spared from adjuvant chemotherapy (CT). Oncotype Dx (ODX) and MammaPrint (MP) assays have been validated as predictors of CT benefit. This study assessed the distribution of recurrence risk in LA and LB breast tumors as defined by Ki67 and PR. Methods: A retrospective analysis was performed in 889 T1-2, N0-Nmic, M0 tumors for which ODX or MP results and local pathology data were available. Ki67 was assessed by IHC (high 14% and low 20%). Histological grade was defined using the Nottingham grading system. Results: Median age 54 years (18-77). All pts had HER2 negative tumors. Median tumor size 15 mm (2-88). Three hundred (33.7%) tumors were classified as LA and 589 (66.3%) as LB. Grade 1 tumors were higher in LA (27%) than in LB (19%) pts (p < 0,001). CT was first recommended in 137 pts (45.7%) with LA vs. 361 pts (61.3%) with LB tumors. ODX was performed in 432 (48.6%) pts and MP in 457 (51.4%). Recurrence risk distribution varied significantly between groups (p < 0.001). ODX: among LA pts, 71.4%, 25.7% and 2.9% had low, intermediate and high recurrence risk respectively; among LB pts respective values were 46.2%, 44.2% and 9.6%. MP: among LA pts, 81.2% and 18.8% had low and high recurrence risk, respectively; among LB pts respective values were 54.2% and 45.8% (Table). After test results CT was recommended to 61 pts (20.3%) with LA vs. 268 pts (45.5%) with LB tumors. R

R

Legal entity responsible for the study: Ministry of Health of the Community of Madrid (Spain) Funding: Ministry of Health of the Community of Madrid (Spain) Disclosure: S. L opez-Tarruella: Advisory role: Novartis, Pfizer, Celgene. Honoraria/ lecturer: Novartis, Pfizer, AstraZeneca, Celgene, Roche. I. Marquez-Rodas: Advisory role: Bristol-Myers Squibb, MSD, Novartis, Roche, Pierre Fabre, Amgen and Bioncotech. M. Martin Jimenez: Holds a patent on the predictive value of the PAM50 gene profile assay. All other authors have declared no conflicts of interest.

185P

T. Kinoshita1, K. Aogi2, M. Takahashi3, K-I. Ito4, T. Oba4, N. Shiroma5, K. Arihiro5, F. Tsukamoto6, S. Shiino1, M. Yoshida7, S. Ohsumi2 1 Division of Breast Surgery, National Cancer Center Hospital, Tokyo, Japan, 2Division of Breast Surgery, Shikoku Cancer Center, Matsuyama, Japan, 3Department of Breast Surgery, Shikoku Cancer Center, Matsuyama, Japan, 4Division of Breast, Endocrine & Respiratory Surgery, Shinshu University School of Medicine, Matsumoto, Japan, 5 Department of Anatomical Pathology, Hiroshima University Hospital, Hiroshima, Japan, 6Department of Breast and Endocrine Surgery, Japan Community Health care Organization (JCHO) Osaka Hospital, Osaka, Japan, 7Department of Pathology and Clinical Laboratories, National Cancer Center Hospital, Tokyo, Japan Background: 95-Gene Classifier (95GC; CurebestTM 95GC Breast) is one of the multigene prognostic assays of estrogen receptor (ER) positive and node negative breast cancer patients, developed 95 gene-set without overlap with that used in another RT-PCR V based product (Oncotype DX ). According to the original paper, the prognostic capability of 95GC has been shown even in the “intermediate” patients by microarraybased simulation model, “Recurrence Online”, of above RT-PCR product. But still 95GC has been validated only using the data from single institute and public database. Here we report the result of the first multi-center validation study for this multi-gene assay. Methods: ER-positive and T1-2/N0/M0 breast cancer patients who received only hormonal therapy, without chemotherapy, in adjuvant were enrolled retrospectively. For each patient, fresh frozen tissue was applied to the assay and the classification result of 95GC, “L” or “H”, was used for the validation on 5 year recurrence free survival (5YRFS). Results: We analyzed 75 eligible cases out of 150 enrolled, and found 47 patients classified as “L” showed 96.5% of 5Y-RFS (95%CI:89.5-98.9) while 28 patients of “H” showed 79.1% of 5Y-RFS (95%CI:63.8-88.5). There was a statistically significant difference between RFS of “L” and “H” groups by Log-Rank test (p ¼ 0.0017). Other factors having significant association with 95GC were histological grade (p ¼ 0.0012), “Recurrence Online” (p < 0.001) and “PAM50” (p < 0.001). And the patients of histological grade 2, of intermediate group by “Recurrence Online” and of Luminal B by “PAM50” could be classified into “L” and “H” by 95GC with different trends of RFS. Conclusions: 95GC was well validated by this first multi-centered retrospective study on 5Y-RFS of ER-positive, node-negative patients who received only hormonal therapy in adjuvant. The result indicates the usefulness of this novel multi-gene assay, as it can classify target patients into 2 groups, “L” and “H”, according to the prognosis of 5YRFS. Legal entity responsible for the study: Takayuki Kinoshita Funding: None Disclosure: All authors have declared no conflicts of interest. R

186P

Table: 184P Recurrence Risk (%) ODX (n ¼ 432)

LA (n ¼ 300)

LB (n ¼ 589)

Low Intermediate High MP (n 5 457) Low High

71.4 25.7 2.9

46.2 44.2 9.6

81.2 18.8

54.2 45.8

Conclusions: There is a wide distribution of recurrence risk results between LA and LB tumors defined by Ki67 and PR which confirms the important role of gene-expression assays in adjuvant decision making. Of note about half of pts with LB tumors had low recurrence risk indicating minimal benefit from adjuvant CT.

56 | Breast cancer, early stage

The first report of multicenter validation study of 95-gene classifier, a multi-gene prognostic assay of estrogen receptor positive and node negative breast cancer patients

Breast cancer PAM50 subtypes: Correlation between RNA-Seq and multiplexed gene expression platforms

A.C. Picornell1, I. Echavarria Diaz-Guardamino2, E.L. Alvarez Castillo1, S. Lopez-Tarruella Cobo2, Y. Jerez2, K. Hoadley3, J. Parker3, M. del Monte-Millan2, R. Ramos Medina1, na1, M. Cebollero4, T. Massarrah2, F. Moreno Ant on5, J.A. GarcıaJ. Gayarre1, I. Oca~ Saenz5, H. Gomez Moreno6, A.I. Ballesteros Garcia7, M. Ruiz Borrego8, C. Perou3, M. Martin Jimenez2 1 Medical Oncology, Instituto de Investigaci on Sanitaria Gregorio Mara~ non (IiSGM), n on, Madrid, Spain, 2Medical Oncology, Hospital General Universitario Gregorio Mara~ 3 Madrid, Spain, Genetics, University of North Carolina - Chapel Hill, Chapel Hill, NC, 4 non, Madrid, Spain, USA, Pathology, Hospital General Universitario Gregorio Mara~ 5 Medical Oncology, Hospital Universitario Clınico San Carlos, Madrid, Spain, 6Medical Oncology, Instituto Nacional de Enfermedades Neoplasicas - INEN, Lima, Peru, 7Medical Oncology, Hospital Universitario de La Princesa, Madrid, Spain, 8Oncologıa, Hospital Virgen del Rocıo, Seville, Spain Background: Gene expression signatures are a key tool for decision-making in breast cancer. In 2000 Perou et al. identified 4 intrinsic subtypes of breast cancer from gene expression data: LumA, LumB, HER2-enriched and Basal-like. These breast cancer subtypes yielded a superior prognostic impact than classical immunohistochemistry factors. From the initial intrinsic subtype, a 50-gene signature was developed for subtype assignment. PAM50 is being successfully used in multiplexed gene expression platR , which is the basis for the ProsignaV R test. The forms such as NanoString nCounterV

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology latter was approved for the risk of distant relapse estimation in postmenopausal women with hormone receptorþ, nodeþ/- early stage breast cancer patients; and is a dailyused tool assessing the need of adjuvant chemotherapy. Methods: The analyses were performed in paraffin embedded tissues (FFPE) from 96 patients recruited in a multicenter, prospective, non-randomized triple negative breast cancer trial (NCT01560663). Pre-treatment core biopsies were performed following clinical practice guidelines and conserved as FFPE for further RNA extraction. PAM50 R and RNA-Seq technologies. Subtype was performed on both NanoString nCounterV assignment was based on the nearest centroid classification following this procedure for both platforms. R /RNA-Seq Results: Subtype calling agreed on 96% of the cases (NanoString nCounterV discordances: 3 Basal-like/HER2-enriched and 1 HER2-enriched/LumA). Both the Spearman correlation to each of the centroids and the risk of recurrence (ROR) were above 0.89 in both platforms. Furthermore, the agreement on proliferation score reached up to 0.97. In addition, 82% of the individual PAM50 genes showed a correlation coefficient >0.80. Conclusions: The RNA-Seq is a fundamental research tool for whole transcriptome analysis. However, it cannot be massively used in the daily clinical practice, due to its processing time requirements and economic costs. We demonstrated that the RNA-Seq R , with the latter protechnology provides similar results to the NanoString nCounterV viding lower cost and more simplicity in its use. Clinical trial identification: NCT01560663 Legal entity responsible for the study: Instituto de Investigaci on Sanitaria Gregorio n (IiSGM) Mara~ no Funding: None Disclosure: All authors have declared no conflicts of interest.

187P

Summary of head-to-head comparisons of patient (pt) risk classifications by the 21-gene recurrence score (RS) assay and other genomic assays for early breast cancer (EBC)

Z. Varga1, P. Sinn2, D. McCullough3, A. Lau3, M.C. Sto¨ppler3, F.L. Baehner3, C. Chao3, A. Seidman4 1 Departement Pathologie und Molekularpathologie, Universit€ atsspital Zu¨rich, Zurich, akopathologie, Pathologisches Institut, Heidelberg, Switzerland, 2Sektionsleiter Gyn€ 3 Germany, Medical Affairs, Genomic Health, Inc., Redwood City, CA, USA, 4Department of Medical Oncology, Memorial Sloan-Kettering Cancer Center, New York, NY, USA

Results: 14 studies were found that compared the RS assay with BCI (1), BCI, EPclin, and ROR (1), EP/EPclin (2), MMP (6), and ROR (4). Overall discordance in risk stratification ranged from 43% to 66% between assays (Table). The RS assay classifies 12% of pts as high risk, compared with EP (63%), EPclin (48%), and MMP (46%), assays with low/high risk groups, and compared with BCI (16%) and ROR (33%), assays that, like the RS assay, use three risk groups. Conclusions: The five most common genomic assays in clinical use for EBC riskstratify pts differently and thus are not interchangeable. Of these, the RS assay classifies the smallest proportion of pts as high risk. Legal entity responsible for the study: Zsuzsanna Varga Funding: Genomic Health Disclosure: Z. Varga: Consultant/advisor: Genomic Health, Roche. P. Sinn: Advisor: Genomic Health. D. McCullough, A. Lau, M.C. Sto¨ppler, F.L. Baehner, C. Chao: Employment and stock ownership: Genomic Health. A. Seidman: Consultant/speaker: Genomic Health.

188P

The impact of Oncotype DX breast cancer assay results on clinical practice: A UK experience

V.E. Crolley1, H. Marashi2, S.K.R. Rawther3, M. Parton4, J. Graham5, A. Vinayan6, S. Sutherland7, A. Rigg8, A. Wadhawan9, C. Harper-Wynne10, E. Spurrell11, H. Bond12, F. Raja13, J. King14 1 Oncology, Royal Free Hospital, London, UK, 2Oncology, Beatson West of Scotland Cancer Centre, Glasgow, UK, 3Oncology, Shefﬁeld Teaching Hospitals NHS Foundation Trust, Shefﬁeld, UK, 4Oncology, Royal Marsden NHS Foundation Hospital, London, UK, 5 Oncology, South Tees NHS Foundation Trust, Northallerton, UK, 6Oncology, East And North Hertfordshire NHS Trust, Stevenage, UK, 7Oncology, Mount Vernon Cancer Centre, London, UK, 8Oncology, Guys and St Thomas NHS Foundation Trust, London, UK, 9Oncology, Velindre NHS Trust, Cardiff, UK, 10Oncology, Maidstone and Tunbridge Wells NHS Trust, Maidstone, UK, 11Oncology, The Whittington Hospital NHS Trust, London, UK, 12Oncology, Royal Cornwall Hospitals NHS Trust, Truro, UK, 13Oncology, University College London Hospitals NHS Foundation Trust, London, UK, 14Medical Oncology, Royal Free Hospital, London, London, UK Background: Gene expression profiling is increasingly being used by clinicians to help determine whether or not to offer adjuvant chemotherapy. Oncotype DX is a 21 gene panel developed to predict the risk of tumour recurrence in patients with oestrogen receptor (ER) positive, HER2 negative breast cancer. NICE has recommended its use for patients at intermediate risk of recurrence, where this information would help clinicians to assess the potential benefit of chemotherapy versus endocrine therapy alone. Our aim was to see how oncologists are interpreting Oncotype DX tests in their clinical practice. Methods: Data on patient and tumour characteristics (size, grade, ER/PR/HER2/nodal status), Oncotype DX recurrence score, treatment options offered and treatment outcomes were collected from 14 cancer centres across the UK. Results: Of the 628 patients tested, 317 (50%) were in the low risk category (recurrence score 50K pts treated based on 21-gene RS results have shown that pts with low RS EBC can safely forgo chemotherapy. Because of its extensive validation and wide clinical use, the RS assay is a common comparator in head-to-head studies with other assays. Methods: Published/presented studies of the RS assay performed on same tumor samples with Breast Cancer Index (BCI), EndoPredict (EP) or EPþclinical features (EPclin), MammaPrint (MMP), and/or Prosigna (ROR) assays were reviewed. Study findings were summarized descriptively.

Table: 187P Discordancea Between the RS Assay and Other Assays BCI

ROR

Studyb

1-level

2-level

Overall

Sestak 2016 Bartlett 2016c Alvarado 2015 Dowsett 2013 Sinn 2017 Varga 2013 Clough 2013 Denduluri 2011 Maroun 2015 Shivers 2013

37%

5%

42%

EP/EPclin

1-level

2-level

Overall

40% 37% 41% 45%

10% 10% 3% 20%

50% 46% 43% 66%

MMP

1-level

2-level

Overall

29%/29%

18%/21%

47%/50%

1-level

2-level

Overall

38% 34% 31% 26%

19% 25% 22% 19%

57% 58% 53% 44%

a. Overall¼any discordance in risk classiﬁcation between the RS assay and other; 1-level¼discordance of one risk category (low $ intermediate or intermediate $ high); 2-level¼discordance of two risk categories (low $ high). b. Four studies lacked risk classiﬁcation information appropriate for inclusion in this table. c. Study used nonstandard RS cutoffs for the RS vs. MMP comparison.

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx362 | 57

abstracts recommended for 101 patients (45.9%), the option of chemotherapy was discussed/ offered to 31 (14.1%) and 88 (40.0%) were not offered chemotherapy. Overall, 160 patients (25.9%) received chemotherapy. Where oncologists recommended chemotherapy to patients (n ¼ 231), 59.7% of patients went on to receive chemotherapy. Where oncologists had offered or discussed chemotherapy as an option (n ¼ 58), 27.6% of patients went on to receive it. The most common regimes were FEC75x6 (23.9%), ECx6 (13.8%) and ECx4 (9.4%), with 13.2% of patients receiving 3rd generation chemotherapy (FEC/T, TC or EC/taxane); other regimes included ACx4, TCx4 and weekly paclitaxel. Conclusions: Throughout the UK, about half of patients tested had low risk Oncotype scores and the majority (74.1%) of patients tested did not receive chemotherapy. The widest variation in clinical practice was observed in interpreting intermediate risk Oncotype results, and in the chemotherapy regimens offered. Legal entity responsible for the study: Judy King Funding: None Disclosure: All authors have declared no conflicts of interest.

189P

Enhancing decision-making about adjuvant chemotherapy in ER1, HER2- early breast cancer (EBC) following EndoPredict testing

L. Fallowfield1, S. May1, L. Matthews1, V. Jenkins1, J. Mackay2, A. Arbon3, B. Hack3, J. Hall4, C. Harper-Wynne5, S. Hinde6, A. Moss7, E. Thanopoulou8, S. Westwell9, D. Wlaszly3, R. Simcock3, G. Patel3, D. Bloomfield3 1 SHORE-C/BSMS, University of Sussex, Brighton, UK, 2Oncology, University College London Cancer Institute, London, UK, 3Oncology, Brighton&Sussex University Hospitals, Brighton, UK, 4Oncology, Dartford & Gravesham NHS Trust, Dartford, UK, 5Oncology, Kent Oncology Centre, Maidstone, UK, 6Centre for Health Economics, University of York, York, UK, 7Oncology, Western Sussex Hospitals NHS Foundation, Worthing, UK, 8 Oncology, Surrey and Sussex HEalthcare NHS Trust, Redhill, UK, 9Oncology, East Sussex Healthcare NHS Trust, Eastbourne, UK Background: Chemotherapy side-effects can be substantial. There is increasing recognition that with surgery, radiotherapy and hormone treatment (tmt), many patients (pts) derive no benefit from chemotherapy and experience only iatrogenic harm. Gene expression profiling tests can help refine recurrence risk and likely chemotherapy benefit. EndoPredict is a multigene test which includes clinico-pathologic parameters to produce an EPclin score classifying risks of distant recurrence as low or high for ERþve HER2–ve pts treated with adjuvant endocrine tmt alone. We compared tmt decisions pre and post EndoPredict test results, pts’ anxiety, decisional conflict and oncologists’ confidence about decisions made. Methods: 14 oncologists in 7 UK hospitals saw 149 pts judged to have equivocal indications for chemotherapy. Pts and oncologists discussed provisional tmt decisions based on usual prognostic factors. These decisions were reconsidered when EPclin results were available. Pre and post-test pts completed Spielberger’s State/Trait Anxiety inventory (STAI) and a decision conflict scale (DCS). Oncologists additionally recorded:basic clinical details, their agreement with, and confidence about tmt decisions (endocrine (E) therapy þ/- chemotherapy(C)). Results: 66.7% pts with an initial E alone decision and a high risk result upgraded to EþC. 9.4% pts with initial EþC decisions and high risk results down-graded to E alone. None of 46 pts initially favouring E alone who were low risk changed decisions. 82.8% who initially wanted EþC and had low risk scores downgraded to E alone. Endopredict results increased oncologists’ confidence (8% ‘strongly agreed’ pre-test, 50% post-test). Oncologists neither agreeing nor disagreeing with decisions fell (24% to 5%). Anxiety was stable in pts with unchanged decisions. Pts whose tmt was downgraded had significantly lower anxiety scores (p < 0.01); those whose tmt was upgraded had increased scores (p < 0.001). Likewise overall uncertainty on DCS fell post-test (p < 0.023). Conclusions: EndoPredict scores increased oncologists’ and pts’ decision-making confidence, generally improved the matching of risk with therapy decisions. Clinical trial identification: ISRCTN69220108 Legal entity responsible for the study: David Bloomfield Funding: Myriad Genetics Disclosure: L. Fallowfield, D. Bloomfield: Educational grant from Myriad Genetics. J. Mackay: Consultancy role to Myriad Genetics and travel expenses. S. Hinde: Educational grant from Myriad Genetics. A. Moss: Consultancy for Myriad Genetics. All other authors have declared no conflicts of interest.

190P

Circulating ESR1 mutations at the end of aromatase inhibitor adjuvant treatment and after relapse in breast cancer patients

V. Allouchery1, L. Beaussire2, A. Perdrix1, D. Sefrioui1, L. Augusto3, C. Guillemet1, N. Sarafan-Vasseur2, F. Di Fiore4, F. Clatot4 1 Medical oncology, Centre Henri Becquerel, Rouen, France, 2Institute for Biomedical Research and Innovation, Inserm U1245, University of Rouen, Rouen, France, 3Medical oncology, Henri Becquerel, Rouen, France, 4Oncologie Me´dicale, Centre Henri Becquerel, Rouen, France Background: Detection of ESR1 circulating mutations is associated with acquired resistance to aromatase inhibitor (AI) in metastatic breast cancer. Until now, the presence of ESR1 circulating mutations at the end of the adjuvant treatment by AI in early breast

58 | Breast cancer, early stage

Annals of Oncology cancer had never been clearly established. In this context, the aim of the present study was to evaluate the ESR1 circulating mutation frequency at the end of adjuvant treatment in patients with a subsequent local or metastatic relapse. Methods: This monocentric retrospective study was based on available stored plasmas and included all early breast cancer patients who completed at least 2 years of AI adjuvant treatment and experienced a documented relapse at least 6 months after the end of their treatment. ESR1 circulating mutations (D538G, Y537S/N/C) were detected by droplet digital PCR in plasma samples taken both at the end of adjuvant treatment and on AI progression in patients re-exposed to AI during the metastatic phase. Results: A total of 39 patients were included, with a median adjuvant AI exposure of 60 months (range 41-85). One patient (2.6%) had a local relapse only, while all the others (97.4%) had a metastatic relapse during follow-up. Median delay between the end of adjuvant treatment and relapse was 25 months (range 6-71). No ESR1 circulating mutation was detectable at the end of AI adjuvant therapy. In contrast, among the 25 patients (64%) who progressed on AI during the metastatic setting, 17 plasma samples were available and 7 patients (41,2%) had a detectable mutation. Conclusions: Our results highlighted that there is no emergence of circulating ESR1 mutation at the end of an AI-based adjuvant treatment in hormone receptor positive breast cancer patients. In contrast, and as expected, we showed that re-exposure to AI in the metastatic setting induced circulating mutation detection in a significant fraction of the patients. Our present findings point out the low interest in ESR1 circulating mutation detection during the adjuvant setting, even for patients that will relapse. Legal entity responsible for the study: Centre Henri Becquerel Funding: None Disclosure: All authors have declared no conflicts of interest.

191P

Pathological proliferation score to predict genomic risk categories in early stage breast cancer

A. Bulbul1, D. Tsao-Wei2, E.A. Mino1, A. Mustafa3, S. Rashad4, H. Abboud5, S. Chouial1, T. Braik1, K. Masoud1, D. Tripathy6 1 Hematology/Oncology, Kymera Independent Physicians, Carlsbad, NM, USA, 2 Biostatistics, University of Southern California Norris Comprehensive Cancer Center, Los Angeles, CA, USA, 3Anesthesiology, Acharya Shri Chander College of Medical Sciences and Hospital, Jammu, India, 4Medicine, All Saints University of Medicine, Roseau, Dominica, 5Medicine, Windsor University, Monee, IL, USA, 6Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA Background: Five of the 16 cancer-related genes used to calculate the Recurrence score (RS) are proliferative genes. Appropriate utilization of an expensive test is important especially in areas of limited resources. A relatively inexpensive ‘Pathological Proliferative score’ (PrS) of a tumor may help group patients in risk categories correlating with the RS. Methods: We retrospectively studied 205 patients with Lymph node negative, hormone receptor (HR) positive, HER2 negative status (ODX candidates) between1990-2015 treated across three rural community oncology practices. Proliferation score was calculated by combining tumor grade, visual mitotic score and Ki67 immunostaining (on a scale of 1-3, lowest score of 3; highest score of 9). Log-rank test was used for survival analysis. Results: PrS correlated with ODX risk recurrence (p < 0.001, Fischer’s Exact test) [Table]. PrS predicted FFP (p ¼ 0.014) at 10 years with PrS (3-4) 96%62%, PrS (5-7) 91%65% and PrS > (7-9) 75%61%. It did not predict PFS(p ¼ 0.77), OS(p ¼ 0.84). Type of adjuvant treatment or none did not affect Low Prs(3-4) 10 yr PFS (p ¼ 0.18 and OS (p ¼ 0.33). Int/High PrS (5-9) showed benefit with adjuvant hormonal therapy compared to none at 10-year OS (p ¼ 5 years after diagnosis (i.e. late recurrence). Clinical trials report contradicting results on the effect of extended endocrine therapy > 5 years to reduce late recurrence risk. Using publically available breast cancer gene expression profiles, we aimed to gain insight into the biology that increases the risk for late recurrences. Methods: Gene expression profiles of primary ERþ/HER2- tumors of breast cancer patients were collected with disease-free survival (DFS) data, defined as time of diagnosis to local recurrence or distant metastasis. We defined (i) a group containing all patients (n ¼ 2,231), (ii) a group that received 5 years of endocrine therapy only (n ¼ 591), and (iii) a group that received no systemic therapy (n ¼ 497). For each group, genes were ranked on their association with DFS as determined by multivariate Cox regression with age, tumor size, grade, lymph node status, and therapy as covariates. Gene set enrichment analysis (GSEA) was performed on these gene lists with the Hallmark collection from the Molecular Signatures Database. Within each group, associations with early recurrence were studied in all patients with censoring at 5 years if no event occurred < 5 years after diagnosis (set I). To study associations with late recurrence, a second set was defined that contained only patients with a follow-up 5 years and no event < 5 years after diagnosis (set II). Results: Within all patients and the group that received 5 years of endocrine treatment only, higher expression of genes belonging to the Hallmark ‘estrogen response late’ was associated with longer DFS in set I and shorter DFS in set II. This Hallmark contains estrogen responsive genes identified in estradiol treated ERþ breast cancer cell lines. However, in patients who received no systemic treatment, higher expression of these genes was associated with shorter DFS in both set I and II. Conclusions: Higher expression of estrogen response genes is associated with a greater risk for late recurrence in patients with ERþ/HER2- breast cancer. Potentially, patients with ERþ tumors with high expression of these genes might benefit most from extended endocrine therapy. Legal entity responsible for the study: R.S.N. Fehrmann Funding: Dutch Cancer Society grants RUG 2010-4739 and RUG 2013-5960, NWOVENI grant (916-16025) and a Mandema Stipendium. Disclosure: All authors have declared no conflicts of interest.

197P

Understanding BRCA1 and BRCA2 mutated breast cancer cases in Romania: First report on founder mutations in Romanians

A. Eniu1, L. Pop2, A. Stoian1, E. Dronca2, R. Matei1, M. Ligtenberg3, H. Ouchene3, A. Onisim1, O. Rotaru1, R. Eniu4, N. Antone1 1 Department of Breast Tumors, Cancer Institute Ion Chiricuta, Cluj-Napoca, Romania, 2 Research Center for Functional Genomics, Biomedicine and Translational Medicine, Iuliu Hatieganu University of Medicine and Pharmacy, Cluj-Napoca, Romania, 3 Department of Human Genetics, Radboud University Medical Centre, Nijmegen, Netherlands, 4School of Medicine, Cardiff University, Cardiff, UK Background: First systematic analysis of BRCA 1(B1) or BRCA2(B2) mutations in high-risk Romanian breast cancer patients (pts) aiming at defining founder mutations.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Methods: This prospective study evaluated the germline B1/B2 mutations in 250 highrisk breast cancer pts tested between 02.2015-12.2016 at IOCN. Inclusion criteria selected pts diagnosed with triple negative breast cancer under the age of 50, or having conventional family history criteria. All pts signed an informed consent. B1/B2 testing was performed using an AmpliSeq-based sequencing analysis, on the Ion Torrent Personal Genome Machine at RCFG. Pathogenic mutations were validated using Sanger technology. MLPA was performed for all pts. Results: Of the 250 pts with breast cancer, 44 (17.6%) carried pathogenic mutations, 29 pts (11.6%) in B1 and 15 (6%)in B2, while 18 patients (7.2%) carried a Variant of Uncertain Significance (VUS). Patient features analysis confirmed the prevalence of younger age, higher grade, hormone receptor negative and Her2 negative status among mutated patients (data not shown). Out of the 16 distinct deleterious mutations identified, 7 (43.75%) occurred in B1 and 9 (56.25%) in B2. The founder mutations identified in B1 gene were: c.5329_5330insC (c.5266dupC) 11 pts (37.93%), c.3607C>T 9 pts (31.03%) and c.181T>G 4 pts (13.79%). Other B1 mutations where c.1687C>T 2 pts (6.89%), and c.4218delG (3.44%), c.212 þ 1G>T (3.44%), c.68_69delAG (3.44%) in one patient respectively. For B2 gene, c.9371A>T (46.66%) was identified as founder mutation (7 pts, 46.66%). Other mutations were found each in one patient (6.66%): c.1528G>T, c.4022C>G, c.7007G>A, c.8695C>T, c.9253delA, c.8680C>T, c.87551G>A, c.8695C>T. Of the founder mutations identified, two (c.3607C>T and c.9371A>T) have not been previously identified as founder mutations in any Eastern European country. Conclusions: This prospective study presents the first extensive results of germline B1/ B2 mutations in Romanian high-risk breast cancer pts. Our results indicate that at least four recurrent B1/B2 mutations qualify as founder mutations; two being newly identified as carrying a founder effect. ClinicalTrials.gov Identifier: NCT02317120. Clinical trial identification: NCT02317120 Legal entity responsible for the study: Alexandru Eniu Funding: AstraZeneca Disclosure: A. Eniu: Research support: AstraZeneca, Roche, Novartis, Celltrion. All other authors have declared no conflicts of interest.

198P

The prevalence of CD146 expression in breast cancer subtypes and its relation to outcome

I.E. de Kruijff1, A.M. Timmermans1, M.A. den Bakker2, A.M.A.C. Trapman1, R. Foekens1, E. Oomen-De Hoop1, M. Smid1, A. TjonA Fat - Hollestelle1, C.H.M. van Deurzen2, M.E. Meijer - van Gelder1, J.A. Foekens1, J.W.M. Martens1, S. Sleijfer1 1 Department of Medical Oncology and Cancer Genomics Netherlands, Erasmus MC Cancer Institute, Rotterdam, Netherlands, 2Department of Pathology, Erasmus MC Cancer Institute, Rotterdam, Netherlands Background: CD146 has several putative (patho)physiological roles in breast cancer. The most prominent is its involvement in the induction of epithelial-to-mesenchymal transition, which might have an effect on cancer phenotype and aggressiveness. Here, we investigated the prevalence of CD146 expression and its prognostic role in breast cancer subtypes. Methods: In total, 1,025 breast cancer patients were available for this retrospective study. From all patients, formalin-fixed paraffin-embedded primary breast cancer tissue was collected and embedded in tissue microarrays, which were stained for CD146. CD146 expression was defined as > 1% of the tumor cells showing CD146 membrane staining. Clinical data were available from all patients (median follow up 118 months, range 4-120). For subtype analysis the Pearson chi-square test was used and the Cox proportional hazards model for survival analyses. Only patients who were lymph node negative and did not receive (neo)adjuvant systemic treatment were included in the survival analyses (n ¼ 551). Results: 113 (11%) out of 1,025 tumors showed CD146 expression. Of these, 43% of the tumors had > 50% of the tumor cells showing CD146 membrane staining. From the molecular subtypes, CD146 positive tumors are often of the triple negative subtype (76 out of 119 (64%), p < 0.001) and histologically of the medullary type (11 out of 23 (48%), p < 0.001). In univariable analysis, CD146 was a prognostic factor for both poor metastasis-free survival (MFS) and overall survival (OS) (respectively HR 1.65, 95% CI 1.02-2.66, p ¼ 0.041 and HR 1.66, 95% CI 1.03-2.69, p ¼ 0.037). When correcting for the traditional prognostic factors (including age, tumor size and grade, ER, PR and HER2) in multivariable analysis, CD146 was not an independent prognostic factor for MFS and OS (respectively HR 1.63, 95% CI 0.93-2.87, p ¼ 0.088 and HR 1.46, 95% CI 0.82-2.61, p ¼ 0.197). Conclusions: CD146 protein expression is present in 11% of the primary breast cancer tumors and is most prevalent in the triple negative and medullary subtypes. CD146 is a prognostic factor for MFS and OS in breast cancer patients, but it is not independent of the traditional prognostic factors. Its potential impact on outcome to systemic treatment such as endocrine therapy, remains to be established. Legal entity responsible for the study: Erasmus University Medical Center Funding: None Disclosure: All authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

199P

Neutrophil-to-lymphocyte and lymphocyte-to-monocyte ratios in breast cancer

L. Gerratana1, D. Basile1, S. Zago2, M.G. Vitale1, G. Pelizzari1, M. Bonotto1, C. Bozza1, M. Bartoletti1, V. Fanotto1, C. Lisanti1, M. Cinausero1, S. Barban1, M. Lera1, I. Venuti1, M. Mansutti3, A.M. Minisini3, G. Fasola3, F. Curcio2 1 School of Medical Oncology, Department of Medicine, University of Udine, Udine, Italy, 2 Clinical Pathology Institute, University Hospital of Udine, Udine, Italy, 3Department of Oncology, University Hospital of Udine, Udine, Italy Background: Immunity plays a pivotal role in cancer progression and prognosis. A high neutrophil-to-lymphocite ratio (NLR) or a low lymphocyte-to-monocyte ratio (LMR) are respectively associated with systemic inflammation and immune suppression and have been associated with a poor outcome. Aim of this study is to further investigate the interaction between the immune system and breast cancer (BC) through the NLR and LMR. Methods: This retrospective study analyzed a consecutive cohort of 657 patients (pts) with a diagnosis of pT1 BC, without restrictions regarding lymph node status (T1BC), or metastatic BC (MBC) treated between 2004 and 2017 at the Department of Oncology of Udine (Italy). Differences in terms of NLR and LMR among the two cohorts and between clinico-pathological characteristics in the T1BC subgroup were explored through the Kruskal-Wallis test. The prognostic impact in terms of OS in the T1BC population was investigated through uni- and multivariate Cox regression. Results: Both NLR and LMR were significantly different between the T1BC and the MBC cohorts. In particular, pts with T1BC had a higher median LMR (3.9 vs 2.9; P ¼ 0.0001) and lower NLR (2 vs 2.7; P ¼ 0.0001). After stratification according to molecular profile, T1BC and MBC cohorts of Luminal B-like subtype were significantly different in terms of both LMR (4.2 vs 3; P ¼ 0.0001) and NLR (2 vs 2.5; P ¼ 0.0001). In triple negative subtype, the difference between TB1C and MBC was observed for NLR (1.9 vs 3.2; P ¼ 0.0272) only. On the other hand, no differences between TB1C and MBC were highlighted for the other subtypes. When focusing on the clinicopathological characteristics of the T1BC cohort, LMR was associated with progesterone receptor (PR) expression (P ¼ 0.0261) and marginally with the estrogen receptor (ER) expression, while NLR with tumor diameter (P ¼ 0.0240) and marginally with grading. Furthermore, among T1BC pts, NLR had no prognostic impact in terms of OS, while LMR was associated with a better outcome also when corrected for ER, PR and HER2 status (HR 0.44, 95%CI 0.28 - 0.71, P ¼ 0.001). Conclusions: These results suggest a role for systemic inflammation and immunesuppression in breast cancer, especially in the triple negative and luminal B-like subtypes. Legal entity responsible for the study: University of Udine Funding: None Disclosure: All authors have declared no conflicts of interest.

200P

Vitamin D as a prognostic factor in triple negative early breast cancer

S. Morales1, A. Gasol Cudos2, A. Novell Alvarez2, F. Vilardell Villellas2, M.J. Panades Siurana2, C. Canosa Morales2, J. Mele Olive2, E. Iglesias Martinez2 1 Breast Oncology, Hospital Universitari Arnau de Vilanova de Lleida, Alpicat, Spain, 2 Breast Oncology, Hospital Universitari Arnau de Vilanova de Lleida, Lleida, Spain Background: Triple negative breast cancer remains without a target therapy. Interventions that could improve pathological complete response (pCR) rates are required. Metabolits of vitamin D could be involved in chemotherapy response. Methods: A series of 147 patients with early or locally advanced triple negative breast cancers was retrospectively analyzed from 2007 to 2016. Patients from 2015 to 2016 period were supplemented with vitamin D and calcium (880UI/1000mg). Analysis of clinicopathological, immune variables and vitamin D pathway were correlated to pCR. Results: Median age was 53, median tumor size 30mm, 48% had nodal involvement, and median ki67 expression was of 70%. Androgen receptor was expressed in 28% of tumors analyzed, EGFR in 89%, CK5/6 in 63%. Mean stromal T lymphocytes infiltrates (sTILS) was of 28%, mean PDL1 expression of 128, mean 53BP1 expression of 125, and mean VDRnuc expression of 132. pCR rate was of 40%, and within patients with vitamin D supplementation was 64% (16/25). Only VDRnuc expression was associated with pCR (p ¼ 0.047) in the univariate and multivariate analysis. Patients with high expression of VDRnuc in tumor had no evidence of relapse (p ¼ 0.024), with similar curves than those who achieve pCR (p ¼ 0.000). Conclusions: VDRnuc expression is a strong predictive (p ¼ 0.047 with pCR) and prognostic (p ¼ 0.024 with relapse) in triple negative breast cancer. Role of supplementation needs to be tested if it could improve VDRnuc levels; whereas in our series patients with supplementation had better pCR rates. Legal entity responsible for the study: Hospital Universitari Arnau de Vilanova de Lleida Institut de Recerca Biome`dica Funding: None Disclosure: All authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx362 | 61

abstracts 201P

Annals of Oncology

CDK12: New breast and ovarian cancer predisposition gene in Tatar population?

E. Shagimardanova1, O. Brovkina2, M. Gordiev3, R. Enikeev4, L. Shigapova1, D. Khodyrev2, O. Gusev1, A. Kedrova2, V. Kosiy2, A. Nikitin2 1 Extreme Biology laboratory, Kazan Volga Region Federal University, Kazan, Russian Federation, 2Molecular genetic laboratory, Federal Research and Clinical Center, FMBA of Russia, Moscow, Russian Federation, 3Laboratory, Tatarstan Cancer Center, Kazan, Russian Federation, 4Chemotherapy, Kazan Clinical Oncology Center, Kazan, Russian Federation Background: The development of hereditary ovarian and breast cancer (OC/BC) is often caused by genetic defects in the DNA repair system. However, the diagnostics in most medical centers of Russia includes PCR-based identification only the eight common mutations in BRCA1/2 for the Slavic population. Previously we established that patients of the Tatar population with OC/BC did not possess most of Slavic mutation and a significant part of the predisposition is due to other mutations in the genes of the homologous recombination (HR) system. The aim of this work is the analysis of the germline mutations in the HR genes. Methods: The DNA from 175 blood samples from patients of the Volga District with hereditary OC/BC were analyzed by targeted NGS (Roche NimbleGen, Illumina MiSeq), the comparison groups included blood samples from patients of Slavic origin. Results: 62% of the detected pathogenic mutations were presented in the BRCA1/2 genes. The remaining mutations were found in other genes of the reparation system (HGMD Professional 2017.1 database). An unexpected finding was the detection of a germline splicing mutation c.1047-2A>G in CDK12 gene (Chr17(GRCh37):37627130A>G, NM_016507.3) in patients of Tatar origin (Table). Mutation c.1047-2A>G is more common in patients with OC/BC in comparison with healthy controls (7/224 vs 0/316, p ¼ 0.002, OR ¼ 21.49, CI 95% ¼ 1.22–377.25).

ER-positive, HER2-negative BC pts with up to 3 involved LNs. OncoMasTR, discovered via a novel network analysis methodology that determines upstream MTRs has been mechanistically verified and offers improved prognostic value compared to existing MGPS. OncoMasTR has been further trained to include LN-positive pts and CPI. Methods: Two independent sample sets: 225 pts from Malmo¨ University Hospital and 106 pts from Ska˚ne University Hospital were used for training, cross-validation and refinement of OncoMasTR. RNA extracted from 225 archived tissues was analysed by RT-qPCR and expression levels of the MTRs were determined by normalising against the expression levels of reference genes. The strongest prognostic combinations of MTRs were identified using statistical models of all possible combinations of MTRs. Clinical performance of the models with the best cross-validated performance in the training data were further evaluated in the 106 independent samples. Results: OncoMasTR classifies up to 72% of LN0 pts and 60% of LN0-3 pts as low risk, with only 4.9% and 5.5% recurrence rate within the respective groups. When incorporating selected CPI, its prognostic performance further improved to a concordance index of above 0.8. Results showed that the OncoMasTR Molecular score (mS) alone adds statistically significant information to the CPI, and the Combined score (cS) also adds statistically significant information to the mS. Conclusions: OncoMasTR offers significant prognostic information to the standard CPI and addresses the unmet clinical need of LN-positive pts. The binary output of OncoMasTR, giving no ambiguous intermediate group helps eliminate uncertainty in the formation of the final treatment decision. OncoMasTR is ready for large-scale clinical validation and, subsequently, clinical translation. Legal entity responsible for the study: OncoMark Ltd Funding: OncoMark Ltd. Disclosure: S. Barron, B. Moran, C-J.A. Wang, T. Loughman, B. Fender, P. Dynoodt, C. Lopez-Ruiz: Employee of OncoMark Ltd. W. Gallagher: Employee of OncoMark Ltd, has stock or ownership of OncoMark Ltd, is a co-inventor of the patent licensed to OncoMark Ltd and received travel, accommodation and expense from OncoMark Ltd. All other authors have declared no conflicts of interest.

Table: 201P CDK12 c.1047-2A>G frequency distribution Subjects origin

BC or OC

Healthy controls

Tatar from Volga District Non-Tatar from Volga District Slavic from Moscow ExAC NFE 1000G

6/94 (6.4%) 1/81 (1.2%) 0/49 (0%) – –

0/32 (0%) – 0/284 (0%) 29/64446 (0.045%) 2/5006 (0.039%)

Conclusions: Gene CDK12 is one of the most frequently altered genes in serous ovarian cancers, but significance of CDK12 germline mutations in hereditary cancers remains to be defined. Its role in carcinogenesis of OC was established recently and CDK12 was not included in most NGS panels of HR genes. Our study demonstrates that CDK12 may be novel candidate gene for OC/BC genetic predisposition. Notably, frequency of CDK12 c.1047-2A>G (6.4%) mutation is comparable with frequency of foundermutation BRCA1 5382insC (7.4%), that indicates its possible founder role in Tatar population. Legal entity responsible for the study: Tatarstan Cancer Center Funding: Kazan (Volga Region) Federal University, Tatarstan Cancer Center Disclosure: All authors have declared no conflicts of interest.

202P

Prognostic value of master transcriptional regulators (MTRs) in early stage breast cancer

S. Barron1, K. Jirstro¨m2, B. Nodin2, H. Jernstro¨m2, C. Ingvar3, B. Moran4, C-J.A. Wang1, T. Loughman1, B. Fender1, P. Dynoodt1, C. Lopez-Ruiz1, W. Gallagher4 1 Product Development, OncoMark Ltd, Dublin, Ireland, 2Department of Clinical Sciences Lund, Oncology and Pathology, Lund University, Lund, Sweden, 3Department of Clinical Sciences Lund, Division of Surgery, Lund University, Lund, Sweden, 4 Biomolecular and Biomedical Science, UCD Conway Institute, Dublin, Ireland Background: Multigene prognostic signatures (MGPS) enable identification of candidate patients (pts) for treatment de-escalation in early stage BC. However, currently available MGPS do not completely address clinical needs by adequately incorporating lymph node (LN)-positive pts and clinicopathological information (CPI). Here, we present OncoMasTR, a MGPS for determining the risk of distant recurrence (DR) in

62 | Breast cancer, early stage

203P

Survival outcomes of all patients treated with breast carcinosarcoma at a UK specialist cancer centre over a 10 year period (2004–2014)

M. Alameddine, J. Chan, F. Alam Clinical oncology, Clatterbridge Cancer Center, Wirral, UK Background: Carcinosarcoma of the breast is a rare and aggressive type of breast cancer presenting as a high grade tumour with lower rates of both lymph node metastasis and oestrogen and progesterone receptor (ER/PR) expression when compared to the more common types of breast cancer, carrying a less favourable prognosis. We present the clinical and pathological findings and outcomes of a series of patients diagnosed and treated for breast carcinosarcoma at a UK specialist cancer centre. Methods: We conducted a retrospective review of data for all patients diagnosed with breast carcinosarcoma between October 2004 and October 2014 at the Clatterbridge Cancer Centre NHS Foundation Trust. Results: Nine patients were diagnosed in the 10-year period, with a median age at diagnosis of 73 years (range 37-76 years). Seven patients (77.8%) were postmenopausal. Six patients (66.7%) presented with a palpable mass. T1, T2, and T3 were found in 1, 6 and 2 patients respectively. N0, N1, and N2 were found in 6, 2 and 1 patients respectively. All patients had G3 disease with a median diameter of 3cm (range 1.9-9.0 cm). Oestrogen receptor (ER) and progesterone receptor (PR) were both negative in 8 patients (88.9%). Whilst one patient had wide local excision, all the rest had mastectomy, of whom 4 had axillary nodal clearance and 4 had sentinel nodal biopsy. Five patients received adjuvant radiotherapy. Adjuvant chemotherapy was delivered to 5 patients (2 patients received neither treatments) and adjuvant hormone therapy was delivered to 2 patients (one of whom had a concurrent contralateral ER/PR positive tumour). Patients were followed up for a median period of 15 months (range 1-60 months). Median DFS is estimated to be 25 months and median OS is estimated to be 49 months (95% CI: 14-84 months). Two patients (22.2%) developed metastases with a DFS time of 13 and 14 months respectively, and both died within 5 months. Conclusions: Within a 10 year period during which our specialist cancer centre were referred 16,500 new breast cancer patients, only 9 patients had carcinosarcoma. Prognosis following recurrence is poor within our limited cohort, in agreement with the published literature. In order for more meaningful analysis of survival outcomes for such a rare form of breast cancer, a multicentre collaborative approach is required. Legal entity responsible for the study: Clatterbridge Cancer Centre Funding: None Disclosure: All authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 204P

ESR1, Ph-mTOR, CDK4/6 and PD-L1 expression as prognostic (and potentially druggable) drivers for pure invasive lobular breast carcinoma (ILC): Preliminary results of prognostic outliers according to a clinical-pathological model

L. Carbognin1, M. Brunelli2, E. Manfrin2, A. Cali o2, I. Sperduti3, M.V. Dieci4, G. Griguolo4, V. Guarneri4, R. Nortilli1, S. Pilotto1, E. Fiorio5, V. Parolin5, E. Orvieto6, F. Pellini7, G.P. Pollini7, P.F. Conte8, A. Scarpa2, G. Tortora1, E. Bria1 1 Medical Oncology, University of Verona, Verona, Italy, 2Department of Pathology and Diagnostics, University and Hospital Trust of Verona, Verona, Italy, 3Biostatistics, Regina Elena National Cancer Institute, Rome, Italy, 4Department of Surgery, Oncology and Gastroenterology, University of Padova, Padua, Italy, 5Oncology, AOU Integrata Verona, Verona, Italy, 6Department of Surgery, Oncology and Gastroenterology, Istituto Oncologico Veneto, Padua, Italy, 7Breast Surgery, AOU Integrata Verona, Verona, Italy, 8 Medical and Experimental Oncology Department, Medical Oncology 2, Istituto Oncologico Veneto IRCCS, Padua, Italy Background: The biological drivers of prognosis for pure ILC are not entirely clear. The aim of this analysis was to investigate the molecular and immune-related portrait of prognostic outliers to identify different patterns of expression associated with prognosis and potentially druggable. Methods: Clinical-pathological multi-center data of resected early-stage pure ILC patients (pts) were correlated to disease-free and overall-survival (DFS/OS). A continuous score was derived according to multivariate Hazard Ratios, in order to derive a 3-class model (Poor/Intermediate/Good Prognosis). IHC (for Ph-mTOR, CDK6, PD-L1), FISH (for ESR1, Ph-mTOR, CDK4, PD-L1) and H&E evaluation (for stromal Tumor Infiltrating Lymphocytes, sTILs) were performed upon pts at Poor and Good Prognosis. Odds Ratio (OR) with 95% CIs for the risk of association with prognostic class of biomarkers was determined. Results: Data from 457 pts were gathered (median age 57 years, median follow: 75 months). The 3-class cross-validated model significantly differentiated DFS and OS (p < 0.0001, prognostic accuracy: 0.65 and 0.71, respectively). Based on DFS, 154 and 20 pts with Good and Poor prognosis, respectively, were identified. The preliminary and exploratory analysis of the first 34 pts (Good/Poor 14/20) is reported (OR < 1: higher chance to be associated with Good prognosis; OR > 1 higher chance to be associated with Poor prognosis).

Table: 204P Biomarker [Method]

OR

95% CIs

High sTILs [H&E] Ph-mTOR deletion [FISH] ESR1 gain [FISH] PD-L1 positive [IHC] CDK4 deletion [FISH] PD-L1 gain [FISH] Score 3þ CDK6 [IHC] Score 3þ Ph-mTOR [IHC] ESR1 deletion [FISH] CDK4 gain [FISH]

0.22 0.33 0.53 0.67 1.22 1.25 2.29 2.48 2.75 3.18

0.01-5.8 0.07-1.49 0.09-3.18 0.08-5.4 0.1-15.1 0.19-8.2 0.21-24.68 0.61-10.05 0.27-23.04 0.15-66.36

Conclusions: Despite unpowered, these preliminary data suggest that Poor and Good prognosis are potentially associated to differential expression of a cluster of biomarkers: ESR1 deletion, CDK4 gain, CDK6 and ph-mTOR over-expression versus high sTILs, PD-L1 positive, ESR1 gain and ph-mTOR deletion, respectively. Legal entity responsible for the study: University of Verona Funding: None Disclosure: All authors have declared no conflicts of interest.

205P

The pregnancy and fertility (PREFER) study: A prospective cohort study on fertility-preserving (FP) strategies in young early breast cancer (EBC) patients (pts)

C. Dellepiane1, M. Lambertini2, V. Fontana3, F. Poggio1, E. Blondeaux1, B. Conte1, A. D’Alonzo1, M. Vaglica1, C. Bighin1, G. Iacono1, A. Abate1, S. Pastorino1, M.C. Pescio4, P. Anserini4, L. Del Mastro1 1 Medical Oncology, IRCCS AOU San Martino - IST-Istituto Nazionale per la Ricerca sul Cancro, Genoa, Italy, 2Breast Cancer Translational Research Laboratory, Institute Jules Bordet, Brussels, Belgium, 3Department of Clinical Epidemiology, IRCCS AOU San Martino - IST-Istituto Nazionale per la Ricerca sul Cancro, Genoa, Italy, 4Gynecology, IRCCS AOU San Martino - IST-Istituto Nazionale per la Ricerca sul Cancro, Genoa, Italy

procedures. To address the significant challenges related to fertility issues, the PREFER study was developed as a national comprehensive program aiming to optimize care and improve knowledge around this topic. Methods: This is a prospective cohort study ongoing across several Italian centers affiliated to the GIM (Gruppo Italiano Mammella) study group. Oncologists offer the available FP strategies to young EBC pts undergoing (neo)adjuvant CT: oocyte cryopreservation (OC), ovarian tissue cryopreservation (OTC) and LHRH analogue (LHRHa) during CT. Eligible pts are premenopausal, 45 years, no previously exposed to CT and/or radiotherapy. Primary objective is to obtain data about preferences and choices of young EBC pts on the FP strategies. Secondary objectives are to evaluate the success and safety of FP strategies, hormonal changes during CT and survival outcomes. The present analysis reports preliminary results of the study including pts enrolled at the coordinating center from November 2012 to May 2017. Results: A total of 131 EBC pts were enrolled; median age was 38.9 years (24.8-45.34). Nine pts (6.87%) refused all FP options. Reasons for refusal were no interest in fertility preservation (5 pts), previous pregnancy (3 pts), no interest in having children (1 pts). LHRHa was accepted by 120 pts (91.6%) and 27 pts (20.6%) accepted gynecologic counseling. Among these pts, 10 (7.6%) accepted OC or OTC. Main reason for refusal of cryopreservation procedures was fear of delaying cancer treatment (3 pts). No complications were observed among women who underwent OC or OTC. Median number of mature oocytes yielded and cryopreserved was 8.5 (4-13). A patient had a spontaneous pregnancy following adjuvant treatment. Conclusions: Despite the great importance of fertility issues in young EBC pts, a minority of them (7.6%) require to access cryopreservation procedures. This is crucial information from a public health perspective and for resource allocation. Clinical trial identification: NCT02895165 Legal entity responsible for the study: IRCCS AOU San Martino IST, Genoa Italy Funding: AIRC - Associazione Italiana per la Ricerca sul Cancro (IG 2013 No14272). Disclosure: All authors have declared no conflicts of interest.

206P

Incidence of permanent alopecia following adjuvant chemotherapy in women with early stage breast cancer

J. Crown1, J. Walshe1, D. Fennelly2, J-C. Long1, S. Cairney1, D. McDonnell1, J. Ballot1, D. Wildes1, E. Sills1, G. Gullo2 1 Medical Oncology, St Vincents University Hospital, Dublin, Ireland, 2Medical Oncology Department, St Vincents University Hospital, Dublin, Ireland Background: Alopecia is one of the most distressing toxicities of adjuvant chemotherapy for patients with breast cancer. Historically, oncologists have reassured patients (pts) that chemotherapy-induced alopecia is temporary, and followed by full hair recovery. More recently there have been troubling reports of permanent alopecia following adjuvant taxanes (Tax). We studied the incidence of long-term hair loss in patients treated on adjuvant trials in our institution. Patients who were enrolled on clinical trials involving Tax (D-Docetaxel, P-Paclitaxel) and/or Anthracyclines (A) were included. Methods: We conducted a telephone interview survey of pts who had completed adjuvant or neo adjuvant A and/or T chemotherapy on clinical trials more than one year before. Ongoing alopecia was graded as 0 (full hair recovery), 1 (mild hair loss) or 2(severe/total). The study was approved by the hospital audit committee. Results: We studied 295 pts who has been treated on 12 studies. Drug exposure: D-260 pts (D nonA-185, DþA-75); A-nonTax-12 pts; AþP 23 pts. The overall incidence of alopecia was 15% (11% grade 1 and 4% grade 2). For all D the incidence was 15% (12% Grade 1 and 3% Grade 2). For DþA-24% (19% Grade 1 and 5% Grade 2). For D non A the incidence was 13% (8% grade 1 and 5% grade 2). For A non T 8% (Grade 2-8%). For PA-13% (4% grade 1 and 9% grade 2). For patients receiving D non-A regimens, there were two levels of D exposure, 300mg/m2 (90 pts) or 450 mg/m 2 (95 pts). The incidence of alopecia was significantly D dose dependent: D300- 7% (all grade 1) and D 450-19% (14% grade 1, 5% grade 2) (p¼.02 chi2) . Among the higher dose D group, the companion drug choices carboplatin for HER2 positive, (55 pts) or cyclophosphamide (40 pts) were associated with similar incidences of permanent alopecia (22% v 16%). Conclusions: Permanent alopecia is a common complication of adjuvant chemotherapy. The risk appears to be highest in regimens which contain A and D, but it is also seen in D non-A, AP and in A non-Tax. For patients receiving D non A, the risk is dosedependent. Our set contains few P pts, and no pts undergoing low dose weekly P. Oncologists should warn all patients undergoing adjuvant therapy of the risk of permanent alopecia. Legal entity responsible for the study: John Crown Funding: None Disclosure: All authors have declared no conflicts of interest.

Background: Premature ovarian failure and subsequent infertility are possible longterm side effects of chemotherapy (CT) in young EBC pts. Limited data are available on the number of pts who consider FP strategies and on the reasons for refusal of these

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx362 | 63

abstracts 207P

Annals of Oncology

Significant changes in dietary intake and physical activity after breast cancer diagnosis in a Chinese breast cancer cohort study

Y. Lei1, W. Yeo1, S.C. Ho2, C-K.I. Lee1, K.L. Cheung1, R. Lee1, Y-Q. He1 Department of Clinical Oncology, The Chinese University of Hong Kong, Hong Kong, China, 2Division of Epidemiology, the Jockey Club School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong, China

1

Background: The diagnosis of cancer can motivate survivors to modify lifestyles, typically involving diet and physical activity. Few data have reported pre- and postdiagnostic lifestyle habits in Chinese population. Methods: In an on-going prospective cohort study which involved 1462 Chinese women with early-stage breast cancer, we evaluated dietary intake and lifestyle factors pre- and post- breast cancer diagnosis. Validated food frequency questionnaires were used to evaluate dietary intake. Leisure time physical activity was measured by a modified Chinese Baecke questionnaire. This report compared changes of dietary intake and physical activity between 12 months before and 18 months after diagnosis of breast cancer. Results: Intake of whole grains, refined grains, fruits, vegetables, eggs, and nuts increased significantly post-diagnosis (Range, 54%-72% increase; P < 0.001, each; Table). Conversely, after diagnosis consumption of red meat, processed meat, poultry, dairy products, soy, sugar drinks, and coffee significantly decreased (Range, 32%-63% decrease; P < 0.001, each). The level of physical activity (MET-hour per week) postdiagnosis was significantly increased (median, 0.75 vs. 5.25; P < 0.001), with 58% of patients became more physically active. However, the proportion of women that met the WCRF/AICR Cancer Prevention recommendations about physical activity (10 METhour per week) was still low, only 21% women met at pre-diagnosis, increasing to 34% at post-diagnosis.

Table: 207P Food group

Reﬁned grains Whole grains Fruits Vegetables Eggs Red meat Processed meat Poultry Dairy products Soy

Pre-diagnosis [Median (IQR), g/1000kcal/day]

Pre-diagnosis [Median (IQR), g/1000kcal/day]

269.9 (213.3-348.3) 3.6 (0-11.9) 95.1 (54.3-137.8) 194.1 (136.8-263.2) 8.3 (4.1-14.8) 48.3 (29.1-72.7) 1.7 (0.3-4.8) 20.4 (9.6-38.2) 15.4 (5.0-44.5) 28.1 (12.8-55.1)

319.6 (257.8-391.2) 7.4 (0.7-30.0) 140.4 (92.0-193.6) 271.5 (197.3-359.8) 10.1 (5.2-16.9) 38.2 (20.3-62.0) 0.4 (0-2.4) 1.1 (0-7.1) 5.7 (0.6-15.9) 23.3 (8.3-49.0)

*IQR, inter quartile range.

Conclusions: In this cohort study, Chinese breast cancer patients reported significant changes in dietary intake and increased physical activity level and a higher proportion met the WCRF/AICR Cancer Prevention recommendations after cancer diagnosis. These findings provided crucial information on lifestyle behaviors in Chinese breast cancer survivors, and provided information to healthcare professionals on survivors’ health and quality of life. Legal entity responsible for the study: The Chinese University of Hong Kong Funding: World Cancer Research Fund International (Grant Number WCRF 2010/249 and WCRF 2014/1197). Disclosure: All authors have declared no conflicts of interest.

208P

Prevention of cardiotoxicity in breast cancer patients: A randomized prospective study

A.M. Elkhateeb1, D.R. Diab1, H. Ibrahim1, N. Gado1, Y.A. Ali2, A.S. Ibrahim3 Clinical Oncology Department, Faculty of medicine Ain Shams University, Cairo, Egypt, 2 Cardiology department, Faculty of medicine Ain Shams University, Cairo, Egypt, 3 Radiodiagnosis, Faculty of medicine Ain Shams University, Cairo, Egypt

1

Background: Adjuvant anthracyclin chemotherapy (ANTC) and trastusumab are documented to prolong survival in breast cancer patients. However, these drugs are well known to induce Left Ventricular Systolic Dysfunction (LVSD). Multiple studies

64 | Breast cancer, early stage

showed that acetyl choline-esterase inhibitor (ACEIs) and beta Blockers (BBs) can prevent LVSD. Methods: One hundred and twenty six patients with M0 breast cancer patients to be treated with ANTC 6 trastuzumab were randomized to an intervention group (group 1) (n ¼ 63 patients) which received cardioprotective drugs; (ACEI;enalapril) and (BB;carvedilol) or to a control group which did not receive cardioprotective drugs (group 2) (n ¼ 63 patients). To evaluate systolic and diastolic functions conventional echocardiography (Simpson method and M- mode) and cardiac magnetic resonance imaging (CMR) were performed at baseline, after 3 cycles and 6 cycles of ANCT, and after 1 year of follow-up. Cardioprotective drugs received: Both enalapril and carvedilol were started at least 24 hours before the first chemotherapy cycle. Results: In the intervention group 58 patients had 3 cycles ANTC, 6 patients received 6 cycles ANTC, and 12 patients received trastuzumab. Whereas in the control group 47 patients had 3 cycles ANTC, 16 patients were given 6 cycles ANTC and 18 patients received trastusumab. After 3 ANTC cycles, LVEF did not change in group 1 (64.35% at baseline vs. 63.59%, p 0.2) but decreased by M- mode in the control group (64.84% at baseline vs. 63.42%, p 0.03) associated with statistically significant deterioration of diastolic function grades. At 1 year follow-up, while no change was observed in LVEF in group 1, there was decrease in LVEF by CMR in group 2 (65.78% at baseline, 61.48% at 1 year, p 0.048). No cases were detected with heart failure or with final EF < 45% in either group. Compared to controls, the intervention group had a statistically significant lower incidence of decrease EF 10% after finishing ANTC by CMR(1.9% vs. 12.5%, p 0.04). Conclusions: Combined prophylaxis with ACEI (enalapril) and BB (carvedilol) may prevent LVSD in patients with non-metastatic breast cancer treated with anthracyclines containing chemotherapy 6 trastuzumab. The clinical relevance of this strategy should be confirmed in larger randomized studies. Legal entity responsible for the study: Academic Group Funding: None Disclosure: All authors have declared no conflicts of interest.

209P

Quality-of-life results from a randomized, phase-II-study of the R ) in the preoperative therapeutic cancer vaccine L-BLP25 (StimuvaxV treatment of women with primary breast cancer (ABCSG-34)

V. Bjelic-Radisic1, C.F. Singer2, G. Pfeiler2, M. Hubalek3, R. Bartsch4, H. Sto¨ger5, A. Pichler6, E. Petru1, R. Greil7, V. Wette8, A. Petzer9, P. Sevelda10, D. Egle11, P. Dubsky12, F. Fitzal12, R. Jakesz13, M. Balic14, S. Frantal15, L. So¨lkner15, M. Gnant13 1 Gyneology and Obstet, Medical Univeristy Graz, Graz, Austria, 2Gyneology and Obstet, Medizinische Universitaet Wien (Medical University of Vienna), Vienna, Austria, 3 Gyneology and Obstet, Brustzentrum Schwaz, Schwaz, Austria, 4Oncology, Medizinische Universitaet Wien (Medical University of Vienna), Vienna, Austria, 5 Oncology, LKH-Univ. Klinikum Graz, Graz, Austria, 6Department of Hematology and Oncology, Landeskrankenhaus LKH Leoben, Leoben, Austria, 7Oncology, Landeskrankenhaus Salzburg - Universit€ atsklinikum der PMU, Salzburg, Austria, 8 Chirurgie, LKH Veit an der Glan, St Veit An Der Glan, Austria, 9Oncology, Krankenhaus der Barmherzigen Schwestern Linz (bhslinz), Linz, Austria, 10Gyneology and Obstet, Krankenhaus Hietzing, Vienna, Austria, 11Department of gynecology, Univ. hospital Innsbruck, Innsbruck, Austria, 12Chirurgie, Vienna General Hospital (AKH) - Medizinische Universit€ at Wien, Vienna, Austria, 13Surgery, Vienna General Hospital (AKH) Medizinische Universit€ at Wien, Vienna, Austria, 14Department of Hematology and Oncology, LKH-Univ. Klinikum Graz, Graz, Austria, 15ABCSG, Vienna, Austria Background: In ABCSG-34, patients with HER2-negative breast cancer were randomized to preoperative standard of care therapy (SoC) with or without L-BLP25 (Stimuvax). This report describes the quality-of-life (QoL) results of the trial. Methods: 400 patients were randomized to receive SoC with or without BLP25. Postmenopausal women with low-risk disease (ER þþþ/ þþ, Ki67 2 cm and ECOG PS 0-1 are eligible. Exclusion criteria include history of invasive BC, stage IV disease, and prior immunotherapy or autoimmune disease. Approximately 204 pts will be randomized 1:1 to receive atezo (840 mg q2w) or placebo with nab-pac (125 mg/m2 qw) for 12 weeks, followed by atezo (840 mg q2w) or placebo with doxorubicin (60 mg/m2 q2w) þ cyclophosphamide (600 mg/m2 q2w) for 4 cycles before surgery. Pts will be unblinded postsurgery and pts in the atezo arm will continue to receive atezo (1200 mg q3w 11 doses). Stratification factors include stage II vs III at diagnosis and PD-L1 expression (IC0 vs IC1/2/3). The primary endpoint is pathological CR (pCR); key secondary endpoints include pCR according to PD-L1 status, pt-reported outcomes, event-free survival and overall survival. Tumor samples will be taken at baseline, on treatment (optional), at surgery and post-recurrence and will be assessed for biomarkers associated with responses and immune escape. Clinical trial identification: NCT number available on poster Legal entity responsible for the study: F. Hoffmann-La Roche Ltd. Funding: F. Hoffmann-La Roche Ltd. Disclosure: E.A. Mittendorf: Financial support to the institution from the following to conduct clinical trials that I serve as the Principal Investigator for: AstraZeneca, EMD Serono, Galena BioPharma, Genentech. Does not personally receive any funding. C.H. Barrios: Research: Pfizer, Novartis, Amgen, AstraZeneca, Boehringer Ingelheim, Roche, Lilly, Sanofi, GSK, Taiho, Mylan, Merrimack, Merck, Astellas, Bristol-Myers Squibb. Consulting: Boehringer Ingelheim, GSK, Novartis, Pfizer, Roche-Genentech, Eisai. N. Harbeck: My COI is available for ESMO on their internal website. D. Miles: Honoraria for Advisory Boards from Roche-Genentech. S. Saji: Honoraria from AstraZeneca, Chugai Pharmaceutical, Eisai, Novartis Pharma, and research funds from AstraZeneca and Chugai Pharmaceutical. H. Zhang: Consultant for Genentech/Roche from 2015. A-N. Duc: Roche employee and stock. S. Rafii: Employed by Roche. C. Lai: Currently employed by Genentech/Roche and hold company stock. All other authors have declared no conflicts of interest.

.008

Conclusions: DbPET was superior to detect residual primary tumors, especially noninvasive carcinoma, after NAC than WBPET. TNR was expected as the better parameter of pathological evaluation than SUV. Legal entity responsible for the study: N/A Funding: None Disclosure: All authors have declared no conflicts of interest.

213TiP

PIONEER- Pre-operative wIndOw study of letrozole plus PR agonist megestrol acetate versus letrozole aloNE in post-menopausal patients with ER-positive breast cancer

S.S. Kumar1, G. Dougall2, A-L. Vallier2, L. Jones3, W. Qian2, E. Provenzano4, C. Caldas4, P. Pantziarka5, J. Carroll1, R. Baird4 1 Cambridge Institute, Cancer Research UK, University of Cambridge, Cambridge, UK, 2 Cambridge Cancer Trials Centre, Addenbrooke’s Hospital, Cambridge, UK, 3Cambridge Breast Cancer Research Unit, University of Cambridge, Cambridge, UK, 4Cambridge Biomedical Campus, Cambridge University Hospitals NHS Foundation Trust Addenbrooke’s Hospital, Cambridge, UK, 5Repurposing Drugs in Oncology, Anticancer Fund, Kingston Upon Thames, UK Background: Recent preclinical findings have been published which provide new insights into the functional cross-talk between the estrogen receptor (ER) and the progesterone receptor (PR) in breast cancer (Mohammed et al., Nature, 2015). Addition of a PR agonist to anti-estrogens directly modifies ERa chromatin binding and the transcriptional response in breast cancer cells, and is anti-proliferative in vitro and in vivo. Megestrol Acetate (MA), an off-patent semi-synthetic derivative of progesterone, has been licensed for many years as a treatment for ERþ metastatic breast cancer. There is

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx362 | 65

abstracts

Annals of Oncology

also good evidence for the effectiveness of MA as a supportive therapy to ameliorate endocrine therapy-related hot flushes. Trial design: PIONEER is a three-arm, open label, multi-centre randomised phase II pre-surgical window trial evaluating effects of 15 days of preoperative therapy with Letrozole (LET), or LET plus MA 40mg, or LET plus MA 160mg in postmenopausal women with newly diagnosed, ERþ HER2- invasive primary breast cancer. Patients are being recruited in Cambridge, with 5-6 other UK sites due to open in q3/4 2017.

anticancer activity. In addition, 560 BC-related gene signatures will also be analyzed. Enrollment started in July 2016 in 10 sites across Spain. To date, 47 patients have been included. We expect to report full study results by Spring 2018. Clinical trial identification: NCT02802748 Legal entity responsible for the study: SOLTI Breast Cancer Research Group Funding: Pierre Fabre Disclosure: All authors have declared no conflicts of interest.

Table: 213TiP 3-arm randomisation Arm A

LET 2.5mg daily

Arm B Arm C

LET 2.5mg daily þ MA 40mg daily LET 2.5mg daily þ MA 160mg daily

The primary endpoint is % change in proliferation between baseline and day 15 tumour biopsies, measured by Ki67 immunohistochemical (IHC) assessment. Secondary endpoints include: expression of Aurora Kinase A, Caspase 3 and Androgen receptor/PR/ EMT markers by IHC; and safety endpoints. Exploratory endpoints include: transcription factor mapping (ChIP-seq) on paired fresh-frozen tumour samples. Patients are randomised in a 1: 1.5: 1.5 ratio for arms A: B: C. Based on results from previous clinical trials, a mean 66% reduction in Ki67 is anticipated for LET alone (arm A), and a 77.5% reduction for combination arms B and C, based on preclinical data. A recruitment total of 189 patients is required. Pioneer will help determine if there is value in conducting a follow-on adjuvant study to investigate the longer term benefit of combining an aromatase inhibitor with MA, and if so, at what dose (40mg vs. 160mg). Clinical trial identification: EudraCT Number: 2016-003752-79 MHRA/REC number: v2.0 5th June 2017 Legal entity responsible for the study: Cambridge University Hospitals NHS Foundation Trust and the University of Cambridge Funding: Anticancer Fund Disclosure: All authors have declared no conflicts of interest.

214TiP

VENTANA (SOLTI-1501): Antiproliferative effect of the addition of oral metronomic vinorelbine to endocrine therapy in luminal/HER2negative early breast cancer: A window of opportunity trial

B. Adamo1, J.A. Perez Fidalgo2, E. Ciruelos3, M. Vidal1, S. Blanch4, A. Lopez5, P. Gomez nez9, X. Gonzalez10, J. Canes11, A. Prat1 Pardo6, L. Murillo7, K. Amillano8, N. Martınez Ja~ 1 Medical Oncology, Hospital Clınic de Barcelona, Barcelona, Spain, 2Medical Oncology, Hospital Clinico Universitario de Valencia, Valencia, Spain, 3Medical Oncology, Hospital on Instituto Universitario 12 de Octubre, Madrid, Spain, 4Medical Oncology, Fundaci on, Leon, Valenciano de Oncologıa, Valencia, Spain, 5Medical Oncology, Hospital de Le 6 Spain, Medical Oncology, Vall d’Hebron University Hospital, Barcelona, Spain, 7Medical 8 Oncology, Hospital Clınico Universitario Lozano Blesa, Zaragoza, Spain, Medical Oncology, Hospital Universitari Sant Joan de Reus, Reus, Spain, 9Medical Oncology, Hospital Universitario Ram on y Cajal, Madrid, Spain, 10Medical Oncology, Hospital Universitario Quir on Dexeus, Barcelona, Spain, 11Scientiﬁc, SOLTI Breast Cancer Research Group, Barcelona, Spain Background: The cornerstone of luminal breast tumors treatment both in early and advanced settings is endocrine therapy (ET). Although extended adjuvant ET has demonstrated benefit, a significant percentage of patients relapse. In previous studies, the combination of ET with CDK4/6 inhibitors has shown unprecedented efficacy suggesting that inhibition of the cell cycle in combination with ET is a strategy to keep exploring. The efficacy and safety of metronomic VNB have been confirmed in preclinical and clinical studies and it is now considered a multi-mechanisms of action therapy that could offer advantages when combined with other drugs. VENTANA study is a “window-of-opportunity” trial designed to explore whether, similarly to CDK4/6 inhibitors, oral metronomic VNB in combination with endocrine therapy induces a superior antiproliferative effect than ET alone. We hypothesize that the synergistic biological effect of the combined treatment could be an alternative to CDK4/6 inhibitors in the treatment of luminal breast cancer patients. Trial design: Pts are randomized (1:1:1) to receive LET 2.5mg daily, oral VNB 50mg 3 days a week, or the combination. After 3 weeks of treatment, pts undergo surgery. Preand post-treatment (surgical) samples will be analyzed for gene expression. The primary objective is to test if oral metronomic VNB and LET induce a superior antiproliferative effect than either drug alone in pts with early BC defined as Luminal by PAM50. This will be evaluated by the expression of 11 proliferative genes contained in the PAM50 subtype predictor (BIRC5, CCNB1, CDC20, CDCA1, CEP55, KNTC2, MKI67, PTTG1, RRM2, TYMS and UBE2C) as surrogate signature biomarker of its

66 | Breast cancer, early stage

215TiP

PALLAS: PALbociclib CoLlaborative adjuvant study: A randomized phase 3 trial of palbociclib with standard adjuvant endocrine therapy versus standard adjuvant endocrine therapy alone for HR1/HER2early breast cancer

E.L. Mayer1, A.M. Demichele2, G. Pfeiler3, W. Barry4, O. Metzger5, P. Rastogi6, F. Symmans7, H.J. Burstein4, K. Miller8, S. Loibl9, S. Schmatloch10, T. Goulioti11, D. Zardavas12, C. Fesl13, M. Koehler14, C. Huang Bartlett14, X. Huang15, M. Piccart16, E. Winer4, M. Gnant17 1 Dana-Farber Cancer Institute, Breast Oncology Center, Harvard Medical School/DanaFarber Cancer Institute, Boston, MA, USA, 2Perelman Center for Advanced Medicine, University of Pennsylvania-Perelman Center for Advanced Medicine, Philadelphia, PA, USA, 3Gyneology and Obstet, Medizinische Universitaet Wien (Medical University of Vienna), Vienna, Austria, 4Dana-Farber Cancer Institute, Harvard Medical School/DanaFarber Cancer Institute, Boston, MA, USA, 5Medical Oncology, Dana Farber Cancer Institute, Boston, MA, USA, 6Division of Hematology/Oncology, Univ. of Pittsburgh Schools of the Health Sciences, Pittsburgh, PA, USA, 7Department of Pathology, MD Anderson Cancer Center, Houston, TX, USA, 8Hematology/Oncology, Indiana University School of Medicine, Indianapolis, IN, USA, 9Medicine and Research, German Breast Group (GBG) Forschungs GmbH, Neu-Isenburg, Germany, 10Senology, Elisabeth Hospital, Kassel, Germany, 11Breast International Group, Brussels, Belgium, 12Breast International Group, Brussels, Belgium, 13Statistics, Austrian Breast and Colorectal Cancer Study Group (ABCSG), Vienna, Austria, 14Oncology, Pﬁzer, New York City, NY, USA, 15Oncology, Pﬁzer, San Diego, CA, USA, 16Chemotherapy Dept., Institute Jules Bordet, Brussels, Belgium, 17Department of Surgery and Comprehensive Cancer Center, Medical University of Vienna, Vienna, Austria Background: Cell cycle inhibition is a proven target for novel cancer therapeutics. Palbociclib (P) is an orally active inhibitor of CDK4/6, and arrests the cell cycle at the G1-S transition. P in combination with endocrine therapy (ET) has demonstrated efficacy in phase II and III randomized trials for patients with newly diagnosed and recurrent hormone receptor positive/HER2 negative (HRþ/HER2-) metastatic breast cancer (MBC), and is approved in these settings. Given confirmed benefits of P and ET for MBC, the PALLAS study was designed to determine if the addition of P to adjuvant ET improves outcomes over ET alone in HRþ/HER2- early breast cancer. Trial design: PALLAS is an international open-label phase III trial randomizing (1:1) patients (pts) to 2 years of P (125 mg daily, 21 days on 7 days off in a 28-day cycle) combined with at least 5 years of provider choice ET (AI, tamoxifen, þ/- LHRH agonist), versus ET alone. The primary objective of the study is to compare invasive disease-free survival (iDFS) for the combination of P and ET, versus ET alone. Secondary objectives include comparison of iDFS excluding cancer of non-breast origin, DRFS, LRRFS, OS, as well as safety. The principal objective of the translational investigations is to determine the predictive or prognostic utility of defined genomic subgroups with respect to iDFS and OS. Additional objectives include evaluation of cfDNA and tissue biomarkers predictive of benefit or resistance, pharmacogenomics, adherence, and patientreported QOL. Eligible pts are pre- or post-menopausal women or men with stage IIIII, HRþ/HER2- breast cancer. Patients may have already initiated ET, but must be randomized within 12 months of diagnosis and 6 months of initiation of adjuvant ET. Trial sample size is 4600 pts and stage IIA pts will be capped at a total accrual of 1000 pts. Interim analyses for safety, futility/efficacy and sample size re-estimation are planned. PALLAS opened in 9/2015 and accrual is ongoing. Contact information: [email protected]. Clinical trial identification: US: IND Nr.(FDA): 126003 clinicaltrials.gov NCT02513394 FDA approval 27 May 2015 Non-US (clinicaltrialsregister.eu dates): EudraCT Number: 2014-005181-30 Sponsor Protocol Number: AFT-05/ABCSG-42/ BIG_14-03 Start Date: 2015-07-09 Legal entity responsible for the study: Alliance Foundation Trials (AFT) LLC for US, ABCSG GmbH for participating countries outside of the US Funding: Pfizer Disclosure: E.L. Mayer, H.J. Burstein, E. Winer: Research funding to the institution from Pfizer. A.M. Demichele: Funding to the institution for clinical trials from Pfizer, Novartis, Genentech and Calithera. Advisory board activities for Pfizer, Novartis and Calithera. M. Koehler, C. Huang Bartlett, X. Huang: Employee and shareholder of Pfizer. M. Gnant: Grants and/or personal fees from Sanofi-Aventis, Novartis, Roche, GlaxoSmithKline, AstraZeneca, Nanostring Technologies, Accelsiors, Pfizer, Smith Medical, outside the submitted work. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

Annals of Oncology 216TiP

OlympiA: A randomized phase III trial of olaparib as adjuvant therapy in patients with high-risk HER2-negative breast cancer (BC) and a germline BRCA1/2 mutation (gBRCAm)

A. Tutt1, B. Kaufman2, J. Garber3, R. Gelber3, E. McFadden4, C. Goessl5, G. Viale6, C. Geyer7, D. Zardavas8, A. Arahmani9, D. Fumagalli9, E. De Azambuja9, N. Ponde9, P. Herbolsheimer10, W. Wu5, J. Constantino11, P. Rastogi12 1 Breast Cancer Now Research Unit, Division of Cancer Studies, King’s College London, London, UK, 2Breast Medical Oncology, Sheba Medical Center at Tel HaShomer, Ramat Gan, Israel, 3Center for Cancer Genetics and Prevention, Dana-Farber Cancer Institute, Boston, MA, USA, 4Statistics, Frontier Science, Inverness, UK, 5Global Medicines Development, AstraZeneca, Gaithersburg, MD, USA, 6Medical School Pathology, Istituto Europeo di Oncologia, Milan, Italy, 7Massey Cancer Center, Virginia Commonwealth University, Richmond, VA, USA, 8Breast International Group, Brussels, Belgium, 9Breast International Group, Brussels, Belgium, 10AstraZeneca, Gaithersburg, MD, USA, 11 Statistics, NRG Oncology, Philadelphia, PA, USA, 12Division of Hematology/Oncology, Univ. of Pittsburgh Schools of the Health Sciences, Pittsburgh, PA, USA Background: In a Phase II study (NCT00494234), treatment with olaparib, a potent, orally available PARP inhibitor, exerted antitumor activity in patients (pts) with advanced BC harboring a gBRCAm. In a randomized phase III trial, olaparib significantly improved PFS compared to chemotherapy (CT) for patients with HER2negative gBRCAm advanced BC (OlympiAD, NCT02000622, ASCO 2017). OlympiA (NCT02032823) is a phase III trial of olaparib as adjuvant therapy for pts with high risk gBRCAm HER2-negative BC who have completed local treatment and (neo)-adjuvant CT. Trial design: OlympiA is a double-blind trial in which high risk HER2-negative pts are randomized (1:1) to receive treatment with olaparib (300 mg tablets bid [2 x 150 mg]) or placebo for 12 months. Eligible pts must have completed local treatment and at least

Volume 28 | Supplement 5 | September 2017

abstracts 6 cycles of (neo)-adjuvant containing anthracyclines and/or taxanes. Pts with triple negative BC (TNBC) must have pT2 or pN1 in the adjuvant and non-pCR in the neoadjuvant setting. Pts with hormone receptor (HR) positive BC must have 4 positive lymph nodes in the adjuvant and non-pCR and CPS&EG score 3 in the neoadjuvant setting. Pts must also harbor a deleterious gBRCAm. Stratification factors include hormone receptor status, prior neoadjuvant versus adjuvant CT, and whether pts have received platinum therapy for current BC. The primary objective is invasive diseasefree survival (IDFS). Efficacy assessments will be made by mammograms/breast MRI scans annually for 10 years, beginning 6 months from randomization, and by medical history/physical examination from randomization every 3 months for 2 years, then every 6 months for a further 3 years and annually thereafter. Secondary objectives include overall survival, distant DFS, incidence of new non-BCs, HRQoL, safety and tolerability. The primary IDFS analysis will be performed after 330 IDFS events using a stratified log-rank test. Patient enrolment began in April 2014 and is currently ongoing. The target number for randomization is 1500 patients across 500 sites and 25 countries worldwide. Support: U10CA12027,-69651,-37377,-69974,-180868,-180822, -189867; AstraZeneca. Clinical trial identification: NCT00494234 Legal entity responsible for the study: AstraZeneca Funding: AstraZeneca and NRG Oncology Disclosure: A. Tutt: Grants/research support: AstraZeneca. C. Goessl: Employee, stock ownership: AstraZeneca. G. Viale: Advisory Board Member: AstraZeneca. D. Zardavas, A. Arahmani, D. Fumagalli: Research grants to BIG from AstraZeneca. P. Herbolsheimer, W. Wu: AstraZeneca employee. J. Constantino: Acts as the data center for the USA part of the OlympiA trial. While mainly funded by the US National Cancer Institute, the University receives funds from AstraZeneca for some aspects of OlympiA data center activity. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx362 | 67

Annals of Oncology 28 (Supplement 5): v68–v73, 2017 doi:10.1093/annonc/mdx364

BREAST CANCER, LOCALLY ADVANCED 217O

A gene signature of chemo-immunization to predict outcome in patients with triple negative breast cancer treated with anthracyclinebased neoadjuvant chemotherapy

C. Criscitiello1, M.A. Bayar2, G. Curigliano1, F. Symmans3, C. Desmedt4, H. Bonnefoi5, B. Sinn3, G. Pruneri6, C. Vicier7, J-Y. Pierga8, C. Denkert9, S. Loibl10, C. Sotiriou11, S. Michiels2, F. Andre´12 1 Division of Experimental Therapeutics, Istituto Europeo di Oncologia, Milan, Italy, 2 Service de Biostatistique et d’Epide´miologie, Gustave Roussy, Villejuif, France, 3 Department of Pathology, MD Anderson Cancer Center, Houston, TX, USA, 4 Translational Breast Cancer Laboratory, Institute Jules Bordet, Brussels, Belgium, 5 Medical Oncology, Institute Bergonie´, Bordeaux, France, 6Unita di Biobank for Translational Medicine, Istituto Europeo di Oncologia, Milan, Italy, 7Department of Medical Oncology, Institut Gustave Roussy, Villejuif, France, 8Department of Medical Oncology, Institut Curie, Paris, France, 9Institute of Pathology, Charite´ Hospital Berlin, Berlin, Germany, 10Medicine and Research, German Breast Group (GBG) Forschungs GmbH, Neu-Isenburg, Germany, 11Breast Cancer Translational Research Laboratory BCTL, Institut Jules Bordet, Brussels, Belgium, 12Breast Cancer Unit, Department of Medical Oncology, Institut Gustave Roussy, Villejuif, France Background: The extent of tumor-infiltrating lymphocytes (TILs) in the residual disease after anthracycline-based neoadjuvant chemotherapy (NACT) is associated with a better prognosis, in patients with triple-negative breast cancer (TNBC). We aimed to develop a genomic signature from pre-treatment samples to predict the extent of TILs after NACT, and then to test its prognostic value on survival. Methods: Using 99 pre-treatment samples (training set), we generated a four-gene signature that predicts post-NACT TILs using the LASSO technique. Prognostic value of the signature on survival was first assessed on the training set (n ¼ 99) and then evaluated on an independent validation set including 185 patients with TNBC treated with NACT. Results: A four-gene signature combining the expression levels of HLF, CXCL13, SULT1E1, and GBP1 predicted the extent of lymphocytic infiltration after NACT. In a multivariate analysis performed on the training set, a one-unit increase in the signature value was associated with distant-relapse free survival (DRFS) (HR: 0.28, 95%CI: 0.130.63, p ¼ 0.0018). For the validation dataset, the four-gene signature was significantly associated with DRFS in the entire set (HR: 0.26, 95%CI: 0.11-0.59, p ¼ 0.0012) and in the subset of patients with residual disease (HR: 0.23, 95%CI: 0.10-0.55, p ¼ 0.0008). Conclusions: We developed a four-gene signature of immune-activation, which predicts outcome in patients treated with NACT for TNBC. Legal entity responsible for the study: Carmen Criscitiello Funding: Transcan-2011, Operation Parrain Chercheurs, Odyssea, Fondation Dassault. Disclosure: All authors have declared no conflicts of interest.

218P

Prognostic estimates of Ki-67 percentage drop after neoadjuvant chemotherapy (NAC) in luminal B (lumB) and triple negative breast cancer (TNBC)

J. Pascual1, B. Rojas-Garcia1, V. Peg2, S. Diaz-Botero3, E. Zamora1, E. Mu~ noz Couselo1, M. Oliveira1, P. Gomez Pardo1, J. Perez Garcia1, F. Ruiz-Pace4, C. Viaplana4, S. Escriva1, L. Garrigos1, M. Arumi1, M. Espinosa-Bravo3, J. Corte´s1, I. Rubio3, C. Saura1, R. Dienstmann4, M. Bellet Ezquerra1 1 Medical Oncology Department, Vall d’Hebron University Hospital Institute of Oncology (VHIO), Barcelona, Spain, 2Anatomical Pathology, Vall d’Hebron University Hospital, Barcelona, Spain, 3Breast Pathology Unit, Vall d’Hebron University Hospital, Barcelona, Spain, 4Oncology Data Science Group (ODysSey), Vall d’Hebron Institute of Oncology (VHIO), Barcelona, Spain Background: Pathologic complete response (pCR) and residual cancer burden (RCB) after NAC are validated prognostic markers in BC. We assessed the impact of adding Ki-67% drop (baseline biopsy - surgery) to distant metastasis relapse-free survival (dRFS) models containing CP factors plus post-treatment stage (MDACC CPS score), estrogen receptor status and tumor grade (MDACC CPSþEG score). Methods: Records from 341 patients (pts) with lumB/HER2 neg (baseline Ki67 >20%) or TNBC who received NAC from 2008 to 2015 at our hospital were reviewed. Uni- and multivariate Cox models were constructed and concordance-index (c-index) calculated. Results: Pts: median age 47 years (24-83), 60% lumB and 40% TNBC, 62% stage 2, 38% stage 3. pCR: 12% lumB, 32% TNBC (p < 0.01). Median Ki-67% drop: 24% lumB, 5% TNBC (p < 0.01), without differences by NAC regimen. dRFS at 5 year-

C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

median follow-up was 75% in lumB (CI95% 67-83) vs 62% in TNBC (CI95% 53-74, p < 0.01); 90% in RCB 0/1 (CI95% 81-97) vs 74% in RCB 2/3 (CI95% 65-83, p < 0.01). As compared to pts with RCB 0/1, those with RCB 2/3 plus Ki-67% drop > ¼ 20% (best cut-off in univariate model) had similar dRFS at 5 years (90%, CI95% 81-100, p ¼ 0.48), irrespective of molecular group. Enrichment for Ki-67 > ¼20% drop in lumB (60%) vs TNBC (30%, p < 0.01) was observed. Both CPS and CPSþEG scores were validated as independent prognostic factors in univariate dRFS models (c-index of 0.70 and 0.78, respectively). The addition of Ki-67% drop (< 20% vs > ¼ 20%) to CPS and CPSþEG scores in multivariable models significantly improved their performance (c-index of 0.74 and 0.81, respectively). Ki-67 > ¼ 20% drop associates with 70-80% reduction in distant relapse risk (HR 0.27 and 0.16 in CPS and CPSþEG models, respectively, p < 0.05). Conclusions: Our data support the addition of Ki-67% drop after NAC in lumB and TNBC to existing dRFS online outcome calculators. In the context of RCB 2/3, Ki67 > ¼ 20% drop is mainly seen in lumB/HER neg tumors. Importantly, Ki-67 < 20% drop identifies a high-risk population that may be eligible to clinical trials with novel therapeutic interventions in the adjuvant setting. Legal entity responsible for the study: Vall d’Hebron University Hospital Funding: Vall d’Hebron Institute of Oncology Disclosure: M. Oliveira: Consulting or Advisory Role in Roche-Genentech and Puma Biotechnology. Research Funding in AstraZeneca and Genentech/Roche. R. Dienstmann: Consulting or Advisory Role in Roche-Genentech and in Astellas Pharma. Research Funding in Merck. All other authors have declared no conflicts of interest.

219P

Immune function and response to neoadjuvant chemotherapy in hormone receptor positive, HER2-negative breast cancer

A. Matikas1, J. Lo¨vrot1, A. Ramberg2, M. Eriksson2, T. Lindsten2, T. Lekberg1, I. Hedenfalk3, N. Loman3, J. Bergh1, A. Erlandsson4, T. Hatschek1, T. Foukakis1 1 Oncology/Pathology, Karolinska Institutet, Stockholm, Sweden, 2Clinical Pathology and Cytology, Central Hospital Karlstad, Karlstad, Sweden, 3Oncology/Pathology, Lund € € University, Orebro, Sweden University, Lund, Sweden, 4Urology, Orebro Background: Gene expression (GE) signatures and Tumor Infiltrating Lymphocyte (TILs) enumeration have shown promise as predictors of response to neoadjuvant chemotherapy in Hormone Receptor negative (HR-) and HER2þ, but not in HRþ/ HER2- breast cancer (BC). This study aimed to explore their predictive value in HRþ/ HER2- BC, based on previous work from our group on the association of immune function and chemosensitivity in advanced HRþ BC. Methods: The PROMIX phase 2 trial enrolled patients with locally advanced HER2BC to receive six cycles of epirubicin and docetaxel, plus bevacizumab during cycles 36. Patients underwent tumor biopsies at baseline and after cycle 2 for GE profiling using DNA microarrays and TIL enumeration according to standard guidelines. Since pathologic complete remission (pCR) is relatively rare in HRþ BC, we also associated an immune gene module score (IMS) and TIL counts with the non-dichotomous variable of decrease in tumor size. Results: Of the 150 enrolled patients, n ¼ 113 were HRþ. For n ¼ 71, both TIL and GE data were available at baseline, while for n ¼ 78 and n ¼ 49 patients longitudinal TIL and GE data at baseline and cycle 2 were available, respectively. At baseline, on both univariate (OR ¼ 2.29, P ¼ 0.037) and multivariate analysis (OR ¼ 2.35, P ¼ 0.044) IMS was associated with pCR, while its association with tumor shrinkage was only apparent on univariate (P ¼ 0.047) and not multivariate analysis (P ¼ 0.061). TIL infiltration >50% (n ¼ 9) was associated with neither pCR (OR ¼ 1.812, P ¼ 0.61) nor tumor shrinkage (P ¼ 0.99). However, decreases in TIL counts in cycle 2 compared with baseline were associated with lesser decreases in tumor size (P ¼ 0.043 for univariate and P ¼ 0.044 for multivariate analysis). Conclusions: Baseline immune function as assessed by GE analysis, but not TIL enumeration, and a preserved abundance of TILs after chemotherapy were predictive for chemosensitivity at the neoadjuvant setting in patients with HRþ, HER2- BC. Clinical trial identification: NCT00957125 Legal entity responsible for the study: Karolinska University Hospital Funding: Swedish Cancer Society, BioCARE, Governmental Funding of Clinical Research within the National Health 1Service and unrestricted grants from Roche Sweden AB Disclosure: J. Bergh: Personal fees: UpToDate. Institutional grants: AstraZeneca, Amgen, Bayer, Merck, Pfizer, Roche, Sanofi-Aventis. T. Foukakis: Personal fees: Novartis, Pfizer, Roche, UpToDate. Research institutional: Pfizer, Roche. All other authors have declared no conflicts of interest.

abstracts

Annals of Oncology 220P

Safe resection margins in breast-conserving surgery 1

1

1

2

3

Y. Kostiuchenko , I. Motuzyuk , O. Sydorchuk , N. Kovtun , M. Krotevich Oncology Department, Bogomolets National Medical University, Kyiv, Ukraine, 2 Department of Statistics and Demography, Taras Shevchenko National Univeristy of Kyiv, Kyiv, Ukraine, 3Pathology, National Cancer Institute, Kyiv, Ukraine

1

Background: Our first research concerning resection margins in breast-conserving surgery (BCS) for breast cancer patients (BCP) was performed in 2003-2006. In the group of BCP where after positive margins we performed mastectomy (0%) and in the group where the negative margins were achieved immediately (1.27%) the incidence of local recurrences was lower comparing to the group that had re-resections (4.76%) or the group without microscopic control of margins (8.16%). We tried to understand the causes of local recurrences after BCS, and develop recommendations on achieving safe resection margins. Methods: To clarify the possibility of cancer spreading to the side of regional lymph nodes we were injecting a solution of aqueous methylene in four points around a tumor in 30 minutes before surgery and subsequently we were studying cuts made at the distance up to 5 cm from the tumor margin. In 12% of cases we identified tumor elements in cuts located at a distance of more than 3 cm. In this study we included 996 BCP of IIII stages who had BCS at the National Cancer Institute in 2008-2015. The main group consisted of 379 BCP who had an additional removal of tissues towards the axillary area as a monoblock during the BCS (with additional histological control as mentioned above). 617 BCP in the control group had standard BCS. The average age of BCP was 50.661,1 in the main group, and 49,860,9 in the control group (p > 0.05). Median follow-up was 142 weeks. Results: Achieving negative margins at the area of possible metastasis significantly reduces local recurrence rates, especially in a stage II (5.34%) and III (3.35%) BCP. The incidence of distant metastases (diagnosed in 93 (9.34%) cases) was significantly lower in the main group – 21 (5.54%) comparing to the control group – 72 (11.67%), both in general and in each stage (p < 0.05). Time of distant metastases occurrence was slightly lower in the control group – 84,768,9 weeks, comparing to the main group – 101,369,5 weeks (p > 0.05). Conclusions: BCS with the additional removal of tissues towards the axillary area and lymph nodes dissection as a monoblock is reasonable and significantly reduces local and distant metastases rates. However, more long-term research in this area is urgently needed, especially in terms of benefit in survival. Clinical trial identification: The study is approved by the Commission on issues of ethics of the National Cancer Institute (Protocol No. 7 of 08.04.2010) and the Commission on issues of ethics of the Bogomolets National Medical University (Protocol No. 71 of 10.04.2013). Legal entity responsible for the study: Bogomolets National Medical University, National Cancer Institute Funding: None Disclosure: All authors have declared no conflicts of interest. 221P

A clinically applicable model to predict risk of relapse in patients treated for locally advanced breast cancer: Potential utilization in future clinical trials

Legal entity responsible for the study: Olexiy Aseyev Funding: None Disclosure: All authors have declared no conflicts of interest. 222P

Role of radiotherapy and its impact on survival of male breast cancer: Experience from a tertiary cancer center

R. Upadhyay1, P. Julka2, G.K. Rath3 Radiation Oncology, All India Institute of Medical Sciences, New Delhi, India, 2Medical Oncology, B.R. Ambedkar Institute Rotary Cancer Hospital (AIMS), New Delhi, India, 3 Radiation Oncology, B.R. Ambedkar Institute Rotary Cancer Hospital (AIMS), New Delhi, India 1

Background: Male breast Cancer (MBC) is a rare disease accounting for about 1% of all malignancies in men and 1% of all breast cancers. These patients(pts) are managed like female breast cancers. There is limited literature available defining the role of radiotherapy(RT) in management of MBC. We conducted a retrospective analysis to study the impact of adjuvant RT on outcome of MBC pts treated at our centre. Methods: Review of MBC pts presenting to our centre from 2005 to 2015 was done. All underwent pre-treatment evaluation in a combined tumor clinic comprising radiation, surgical and medical oncologist. Most pts were treated with surgery followed by adjuvant therapy and kept on regular follow up. Overall survival (OS) was defined as time from pathologic diagnosis to last follow up or death. Disease free survival (DFS) was defined as time from diagnosis to first relapse. Results: 96 pts of MBC were identified. Median age was 58 years (range 28-83). Clinical stage I, II, III and IV were 8, 27, 39 and 22 respectively. Of 66 pts with known receptor status, 83% were ER positive, 82% PR positive, 20% Her-2/neu positive and 7.5% triple negative. 69 pts underwent modified radical mastectomy or wide local excision. 54% were pathologically node positive. Adjuvant RT was delivered to 34% pts at 1.8-2 Gy per fraction to a median dose of 50 Gy (range 45-60 Gy). Radiation field comprised of chest wall with (75%) or without regional nodes (25%). Median follow up was 24 months (range 9-132). 16 pts had relapse out of which 4 had local and 13 had distant (most common site bone) metastases after a median duration of 19 mnths. 2 yr estimated DFS for the entire cohort was 79.2% and 2 yr OS was 85.7%. The 2 yr DFS in pts undergoing surgery was 86.4% vs 29.2% in those who did not(p ¼ 0.001). Pts who received adjuvant RT had better 2 yr DFS (92.4% vs 52.9%, p ¼ 0.002). Adjuvant chemotherapy did not significantly affect the 2 yr DFS (93.1% vs 73%, p ¼ 0.123). Conclusions: MBC mostly present in advance stages at our centre and harbor HR positive disease with low HER-2 overexpression. Adjuvant RT provided a statistically significant improvement in outcome. Longer follow up of these cohort of pts is required for accurate evaluation of role of RT in MBC. Legal entity responsible for the study: All India Institute of Medical Sciences Funding: None Disclosure: All authors have declared no conflicts of interest.

223P

Early characteristics of breast neoplasms on contrast-enhanced ultrasonography and its clinical value

O. Aseyev, L. Simmonds, M. Gertler, S. Verma Medical Oncology, The Ottawa Hospital Regional Cancer Centre, Ottawa, ON, Canada

S. Qin, Z.W. Jie Ultrasound Medicine, Nanjing Jingdu hospital, Nanjing, China

Background: Despite advances in cancer treatment, over 25% of patients (pts) with locally advanced breast cancer (LABC) relapse during first 5 years after treatment. The primary objective was to construct a prediction tool for risk of relapse (RR) in pts with LABC after neoadjuvant therapy (NAT). Methods: This was single center, retrospective study of 546 pts with LABC who received NAT at the Ottawa Hospital Cancer Center between 2005 and 2015. Median follow-up was 49 months. The following data collected: demographics, tumor size, nodal and receptor status, grade, HER-2, stage, treatment and clinical outcomes. Primary endpoints were local and/or distant recurrence rate and time to relapse during the first 5 years. A prediction tool was devised based on the Cox regression model. Results: Over 60 variables were included in primary analysis. Cox regression proportional hazards model analysis resulted in only 5 factors with significant influence on risk of relapse during first 5 years of follow up. Risk factors and their risk prediction value are: 1) residual disease (yes- 4; no-0), (HR ¼ 4.25; p-value ¼G3) were leukocytopenia (87.5%, 83.3%), neutropenia (87.5%, 83.3%), febrile neutropenia (20.8%, 33.3%), anemia (16.7%, 41.7%) and thrombosis (0%, 8.3%), respectively. Conclusions: The combination of olaparib and eribulin was well tolerated and showed a promising efficacy and safety for metastatic TNBC. Clinical trial identification: release date: 31/March/2017 (in Japanese) Legal entity responsible for the study: National Cancer Hospital () National Cancer Center National Cancer Center Funding: Exploratory Oncology Research & Clinical Trial Center in National Cancer Center Disclosure: K. Aogi: Personal fees as honoraria from Eisai, AstraZeneca, Taiho Pharmaceutical, Daiichi Sankyo, Mochida Pharmaceutical, Ono Pharmaceutical, and Eli Lilly Japan. Research funds from Chugai Pharmaceutical, Eisai and Sanofi. K. Yonemori: Personal fee as honorarium from Eisai. M. Takahashi: Personal fees as honoraria from Chugai Pharmaceutical, Eisai, and AstraZeneca. N. Masuda: Personal fees as honoraria from Chugai Pharmaceutical and AstraZeneca, and his institution received research funds from Chugai Pharmaceutical and Eisai. Y. Naito: Honoraria from Eisai, Chugai Pharmaceutical, Taiho Phamaceutical, Novartis Pharma, Eli Lilly Japan, Meiji Seika Pharma, and Bayer Yakuhin. He received research funds from Merck Serono, AstraZeneca, Eli Lilly Japan, and Nippon Kayaku. Y. Fujiwara: Honoraria from AstraZeneca, Eisai, Daiichi Sankyo, Taiho Pharmaceutical, Chugai Pharmaceutical, and Eli Lilly Japan. He received grants from Taiho, Takeda, Chugai, Eli Lilly Japan, and Nippon Kayaku. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

252P

Phase II study of gemcitabine, trastuzumab, and pertuzumab for HER2positive metastatic breast cancer after prior pertuzumab-based therapy

N.M. Iyengar, L.M. Smyth, D. Lake, A. Gucalp, J.C. Singh, T.A. Traina, P. Defusco, M.N. Dickler, M.N. Fornier, S.B. Goldfarb, K. Jhaveri, S. Modi, T. Troso-Sandoval, K. Jack, G. Ulaner, M. Jochelson, J. Baselga, L. Norton, C.A. Hudis, C.T. Dang Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, USA Background: The combination of taxanes with trastuzumab (H) and pertuzumab (P) for first line treatment of HER2-positive metastatic breast cancer (MBC) is associated with improved progression-free (PFS) and overall survival (OS). Treatment per physician’s choice with anti-HER2 therapy after second line therapy is associated with a median PFS of 3 months. While continued use of H in therapeutic combinations after progression on H-based therapy is standard, the efficacy of continuing HP-based treatment after progression on P-based therapy is unknown. Methods: This is a single arm phase II trial of gemcitabine (G) with HP. Eligible patients (pts) had HER2-positive (IHC 3þ or FISH 2.0) MBC with prior HP-based treatment and 3 prior chemotherapies. Pts received G (1200 mg/m2) on days 1 and 8 of a q 3 week (w) cycle, and H (8 mg/kg load ﬁ 6 mg/kg) and P (840 mg load ﬁ 420 mg) q3w. The primary endpoint is PFS at 3 months. Secondary endpoints include OS, safety and tolerability. An exploratory endpoint is to compare PFS by RECIST criteria versus 18-F FDG-PET response criteria. The study therapy will be considered successful if at least 27/45 (60%) patients are progression free at 3 months. Results: As of 27 April 2017, 45 of 45 pts are enrolled; 39 are evaluable at 3 months and 6 have not had 3-month evaluation. At 3 months, 30/39 (77%) are progression free (1 CR, 8 PR, 21 SD); 9 pts progressed. There are no cardiac or febrile neutropenic events to date. 4 pts required G dose reduction (3 grade 3 neutropenia and 1 grade 3 vomiting) and the study was amended to lower initial G dose to 1000 mg/m2. Conclusions: The preliminary 3 month-PFS is 77% in evaluable pts (95% CI 62% to 87%). The updated 3 month-PFS results will be presented. Continuation of P beyond progression is associated with apparent clinical benefit. A randomized trial is justified to confirm this clinically important observation. Clinical trial identification: NCT02252887 Legal entity responsible for the study: Memorial Sloan Kettering Cancer Center Funding: Genentech/Roche Disclosure: All authors have declared no conflicts of interest.

253P

Abemaciclib plus fulvestrant in patients (pts) with HR1/HER2endocrine therapy naı¨ve (EN) advanced breast cancer - an exploratory analysis of MONARCH 2

P.A. Kaufman1, H. Iwata2, P. Nikolinakos3, J.T. Beck4, H. Koh5, J.L. Gonzalez-Trujillo6, Y. Lin7, S. Barriga8, D. Headley9, N. Bourayou10, A. Llombart Cussac11, G.W. Sledge Jr.12 1 Medicine, Section of Oncology, Norris Cotton Cancer Center Dartmouth-Hitchcock Medical Center, Lebanon, NH, USA, 2Breast Oncology, Aichi Cancer Center Hospital, Nagoya, Japan, 3Hematology and Medical Oncology, University Cancer & Blood Center, LLC, Athens, GA, USA, 4Department of Medicine, Highland Oncology Group, Fayetteville, AR, USA, 5Department of Medical Oncology, Kaiser Permanente Medical Group, Bellﬂower, CA, USA, 6Oncology, Fundacion Rodolfo Padilla AC, Leon Guanajuato, Mexico, 7Global Statistical Sciences, Eli Lilly and Company, Indianapolis, IN, USA, 8 Oncology Clinical Development, Eli Lilly and Company, Madrid, Spain, 9Oncology Clinical Development, Eli Lilly and Company, Indianapolis, IN, USA, 10Oncology Clinical Development, Eli Lilly and Company, Paris, France, 11Medical Oncology Service, Hospital Arnau Vilanova, Valencia, Spain, 12Department of Medicine, Stanford University, Stanford, CA, USA Background: MONARCH 2 demonstrated that the addition of abemaciclib, dosed on a continuous schedule at 150 mg twice daily, to fulvestrant (F) significantly improved progression-free survival (PFS) and objective response rate (ORR) compared to placebo (P) plus F (PFS hazard ratio [HR], 0.553, P2 cycles in 15 mg pt, with T, P, T-DM1, L); and 3 PD. BoR in 5 GE pts (1 too early): 1 SD > 4 cycles (10 mg, prior T, IHC 3þ) and 4 PD. BoR in adnexal (10 mg) and CRC (15 mg) pts was PD. Conclusions: ZW25 was well tolerated with promising single-agent activity including SD > 6 mo and PR in pts with HER2 expressing ca who have progressed after standard tx, including multiple lines of prior HER2 targeted agents. These early signs of activity support the therapeutic potential of ZW25. Clinical trial identification: NCT02892123 Legal entity responsible for the study: Zymeworks Inc Funding: Zymeworks Inc Disclosure: M. Beeram, D. Rasco: Employee of START. D. Hausman: Stockholder and employee Zymeworks Biopharmaceuticals Inc. R. Korn: Employee and leadership role at Imaging Endpoints. G. Rowse: Employee of Zymeworks Inc. J. Thimmarayappa: Employee of Zymeworks Biopharmaceuticals Inc. A. Tolcher: Employee of START, leadership role at Symphogen, member of Zymeworks Clinical Advisory Board. F. Meric-Bernstam: Honoraria: Genentech, Roche Diagnostics. Consulting or Advisory Role: Celgene, Genentech, Inflection Biosciences, Novartis, Roche. All other authors have declared no conflicts of interest. 256P

Efficacy of palbociclib plus fulvestrant in advanced Hormone Receptorpositive (HR1) metastatic breast cancer (MBC) pretreated with everolimus: Real-life data from the french temporary authorization for use (TAU) at the Institut de Cance´rologie de l’Ouest

P. Du Rusquec1, C. Palpacuer2, L. Campion2, A. Patsouris3, P. Augereau3, C. Gourmelon1, M. Robert1, L. Dumas4, C. Folliard5, M. Campone1, J-S. Frenel1 1 Medical Oncology, Centre Rene Gauducheau, Nantes, France, 2Biostatistics, Centre Rene Gauducheau, Nantes, France, 3Medical Oncology, Centre Paul Papin, Angers, France, 4Pharmacy, Centre Rene Gauducheau, Nantes, France, 5Pharmacy, Centre Paul Papin, Angers, France Background: The CDK4-6 inhibitor palbociclib, combined with hormonal therapy is a new standard of treatment for HRþ Metastatic Breast Cancer. Before the European Medicines Agency approval, a Temporary Authorization for Use (TAU) has been set up in France restricted to patients pretreated with everolimus. We present the efficacy data of this combination in this population. Methods: Between November 2015 and November 2016, all the patients treated with palbociclib þ fulvestrant according to the TAU in our institution were prospectively included. Datas from their medical records and adverse events (AE) were collected. Results: 60 patients received at least one dose of palbociclib in combination with fulvestrant with a median age of 61 years. 50 patients (83.3%) had visceral metastasis and 10 (16.7%) had bone only disease. Patients had received an average of 5.3 lines of treatment before palbociclib initiation, including hormonal therapy (mean ¼ 3.0) and chemotherapy (mean ¼ 2.3). Of note, 28 patients (46.7%) had received fulvestrant previously and all had been pretreated with everolimus. With a median follow-up of 8.1 months, median progression free survival (PFS) was 6.1 months (95% CI, 4.2 to 7.4) and median overall survival was not reached. PFS was the same according to the presence of visceral metastasis or no (HR 1.46 (95% CI, 0.57 to 3.74), p ¼ 0.42). Interestingly, patients treated previously with fulvestrant and subsequently rechallenged with fulvestrant and palbociclib had a PFS of 6.4 months, which was similar to patients who didn’t receive fulvestrant previously (HR ¼ 1.00 (95% CI 0.55 to 1.83), p ¼ 1.00). The most common AE were neutropenia (n ¼ 53), thrombocytopenia (n ¼ 25) and anemia (n ¼ 20). At the time of this analysis (April 2017), 36 patients received a further line of treatment after progression. Conclusions: In this heavily pretreated population, the association of fulvestrant plus palbociclib provides an interesting median PFS of 6.1 months. Patients previously treated with fulvestrant seem to derive the same magnitude of benefit compared to fulvestrant naive patients. Legal entity responsible for the study: Institut de Cance´rologie de l’Ouest Funding: None Disclosure: All authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology

257P

A phase 1 study of BYL719, an a-isoform selective PI3K inhibitor, in Japanese patients with advanced solid malignancies

J. Tomomatsu1, S. Iwasa2, H. Saka3, S. Takahashi1, K. Nakano1, S. Morita4, M. Inoue4, H. Nakahama2, Y. Kogure3, T. Kakizume5, K. Natsume5, T. Aoki5, C. Quadt6, Y. Yamada2, Y. Ando4 1 Department of Medical Oncology, The Cancer Institute Hospital of Japanese Foundation for Cancer Research, Tokyo, Japan, 2Department of GI Medical Oncology, National Cancer Center Hospital, Tokyo, Japan, 3Department of Respiratory Medicine, Nagoya Medical Center, Nagoya, Japan, 4Department of Clinical Oncology and Chemotherapy, Nagoya University Hospital, Nagoya, Japan, 5Oncology, Novartis Pharmaceuticals KK, Tokyo, Japan, 6Oncology, Novartis Pharmaceuticals AG, Basel, Switzerland Background: BYL719 is an oral inhibitor that selectively targets the a-isoform of class l PI3K. In a first-in-human ph1 study in mostly Western patients (pts) with PIK3CA alteration, the maximum tolerated dose for once-daily (qd) BYL719 was declared as 400 mg and preliminary antitumor activity was observed. Previous preliminary findings showed tolerability, safety, and pharmacokinetic (PK) results in dose escalation of the first-in-Japanese ph1 study. Here, we report results of the food effect on the PK profile of BYL719 at steady state, additional safety, and preliminary efficacy in Japanese pts. Methods: Pts were aged 18 years with histologically confirmed, advanced solid tumors. Pts received BYL719 qd in 28-day cycles until disease progression, unacceptable toxicity, or investigator/pts decision. The objectives in the expansion part were to assess food effect on PK profile, preliminary antitumor activity, and safety. Pts with PIK3CA alteration were selected in the expansion part and were randomized to receive BYL719 at the recommended dose 350 mg qd in fasted or fed condition on cycle 1 day 22 and cycle 2 day 1 in a crossover fashion. Pts received BYL719 1 hr following a light breakfast and continued to be fasted for 1 hr after each dose. Results: Thirty-three pts were enrolled and all pts discontinued treatment. The median duration of exposure was 71.0 days (range, 6-462). The common BYL719-related allgrade (Gr) AEs (>30%) were hyperglycemia, rash maculopapular (48.5%, each), diarrhea (45.5%), and decreased appetite (33.3%). BYL719-related Gr 3 or 4 AEs (20%) were rash maculopapular (24.2%) and hyperglycemia (21.2%). Eight pts were enrolled in the expansion part; six of them were eligible for PK analysis. The geometric mean values of dose-normalized Cmax and AUC0-24 in fed state were 78% and 56% higher than those in fasted state, respectively. One pt experienced Gr 4-infected neoplasm meeting DLT criteria, 1 pt with uterus cancer and PIK3CA mutation had an unconfirmed partial response, and 18 pts had a stable disease. Conclusions: In this ph1 study of BYL719 in Japanese pts, a positive food effect and preliminary antitumor activity were observed at steady state in pts with PIK3CA mutation status. Clinical trial identification: NCT01387321 Legal entity responsible for the study: Novartis Pharmaceuticals KK, Tokyo, Japan Funding: Novartis Pharmaceuticals KK, Tokyo, Japan Disclosure: H. Saka: Grants from Novartis, KHK, Daiichi Sankyo, Merck, Esai, BristolMyers Squibb, Taiho, Ono, Chugai, Eli Lilly, Beyer, MSD, Quintiles, West JCOG. Personal fees Chugai, Kyorin, NU, BI, Eli Lilly, Astellas, Nobelpharma, JSRE, Ono, Chunichi Shimbun, Taiho. S. Takahashi: Grants from Novartis, during the conduct of the study; grants from Chugai, grants from Astrazeneka, grants from Daiichisankyo, grants from Bayer, grants from Parexel, outside the submitted work. K. Natsume, T. Kakizume: Personal fees from Novartis Pharmaceuticals KK, during the conduct of the study; Y. Ando: Grants and personal fee: Chungai, Takeda, KHK, Eisai, Taiho, Nippon, YakultHonsya, Mochida, Merck, Ono, Eli Lilly, Novartis, Janssen, Hisamitsu, GSK, Terumo, Bayel, Meiji, BSC, Boehringer Ingelheim, Bristol-Myers Squibb, Sawai, Otsuka, Shionogi, outside the submitted work. All other authors have declared no conflicts of interest.

258P

Phase 1 study of RX-5902, a novel orally bioavailable inhibitor of phosphorylated P68, which prevents b-catenin translocation in advanced solid tumors

J. Diamond1, G. Eckhardt2, L. Gluck3, M. Gutierrez4, C. Peterson5, R. Pila6, E. Benaim7 Department of Medicine, University of Colorado Denver School of Medicine, Aurora, CO, USA, 2Anschutz Cancer Center, University of Colorado Denver School of Medicine, Aurora, CO, USA, 3Oncology, Clinical Research Unit ITOR, Greenville, SC, USA, 4John Theurer Cancer Center, Hackensack University Medical Center, Hackensack, NJ, USA, 5 Clinical Operations, Rexahn Pharmaceuticals, Inc, Rockville, MD, USA, 6Clinical Operations, Rexahn Pharmaceuticals, Inc, Rockville, MD, USA, 7Oncology, Rexahn Pharmaceuticals, Inc, Rockville, MD, USA

1

Background: RX-5902 is a novel compound that targets phosphorylated p68 RNA helicase (also known as DDX5), a member of the DEAD box family of RNA helicases. Phosphorylated p68 may play a vital role in cell proliferation and tumor/cancer

Volume 28 | Supplement 5 | September 2017

progression through inhibition of b-catenin translocation. We report preliminary results of the Phase 1 study of RX-5902 as a single agent to treat advanced solid tumors. Methods: The dose finding portion of the Phase 1 study (NCT02003092) was designed to evaluate safety, tolerability and dose limiting toxicities, to identify the maximum tolerated dose and a recommended phase 2 dose and schedule (RP2D). Secondary objectives were pharmacokinetics (PK) and antitumor activity (RECIST v1.1). Eligible subjects (aged 18 years), with relapsed/refractory solid tumors that had been heavily pretreated, received oral RX-5902 ranging from 25 mg to 350 mg and administered at 1, 3, 5 or 7 times per week for 3 weeks followed by 1 week of rest or for 4 weeks without rest. Results: As of May 2017, 35 subjects (23 Females, 12 males) were treated with oral RX5902. The dose limiting toxicities were G4 hyponatremia (n ¼ 1) and G3 fatigue (n ¼ 1) at 300 mg administered daily for 4 weeks. The maximum tolerated dose of 250 mg, will be studied further in the Phase 2a portion. Of the 35 subjects treated, 15 subjects (breast ERþ/PRþ/Her2-, triple negative breast (n ¼ 2), cervical (n ¼ 2), neuroendocrine (n ¼ 3), paraganglioma, colorectal (N ¼ 3), pancreatic, ovarian, head and neck cancers) experienced stable disease; 2 subjects have received treatment for > 2.5 years. The most common related adverse events were G1/G2 anorexia, G1/G2 nausea, G1/G2 vomiting, G1/G2 diarrhea, G1 weight loss and G1/G2 fatigue. Oral RX-5902 was bioavailable with median Tmax of 2 hours and median elimination half-life of 12 hours. Conclusions: Data from this study support that RX-5902 is safe and well-tolerated at the doses and schedules tested. The RP2D of 250 mg of RX-5902 administered daily for 5 consecutive days for 4 weeks is being studied further in metastatic triple negative breast cancer in the Phase 2 portion of this study. Clinical trial identification: NCT02003092 Legal entity responsible for the study: Rexahn Pharmaceuticals, Inc. Funding: Rexahn Pharmaceuticals, Inc Disclosure: C. Peterson, R. Pila, E. Benaim: Employee of Rexahn Pharmaceuticals, Inc. All other authors have declared no conflicts of interest.

259P

Adherence to International ESO-ESMO (ABC) guide-lines in HER2-ve metastatic breast cancer (MBC) patients (pts): Preliminary results of the GIM 13 - AMBRA Study

M. Cazzaniga1, G. Mustacchi2, M. Giordano3, O. Garrone4, M. Donadio5, L. del mastro6, L. Livi7, C. Natoli8, A. Michelotti9, A. Turletti5, R. Ferdinando10, M. De Laurentiis11, P. Marchetti12, M. Montemurro13, E. Romagnoli14, S. De Placido15, P. Pronzato16, L. Biganzoli17, G. Bisagni18, E. Bria19 1 Medical Oncology, Osp San Gerardo, Monza, Italy, 2Oncology, University of TriesteOspedale di Cattinara, Trieste, Italy, 3Oncology, Presidio Ospedaliero Ospedale Sant’Anna, San Fermo Della Battaglia, Italy, 4Medical Oncology, Azienda Ospedaliera St. Croce e Carle, Cuneo, Italy, 5Medical Oncology, AOU S. Giovanni Battista - Molinette, Turin, Italy, 6Medical Oncology, AO San Martino IST, Genoa, Italy, 7Radiotherapy, Azienda Ospedaliera Universitaria Careggi, Florence, Italy, 8Medical Oncology, P.O. Clinicizzato Ss. Annunziata Universita’ Degli Studi, Chieti, Italy, 9Medical Oncology, Azienda Ospedaliera Universitaria S.Chiara, Pisa, Italy, 10Medical Oncology, AO Cardarelli, Naples, Italy, 11Medical Oncology, IRCCS Fondazione Pascale, Naples, Italy, 12 Medical Oncology, Azienda Ospedaliera St. Andrea, Rome, Italy, 13Medical Oncology, Ist Candiolo, Turin, Italy, 14Medical Oncology, Ospedale Generale Provinciale Macerata, a degli studi di Napoli “Federico II”Macerata, Italy, 15Medical Oncology, Universit Dipartimento di Medicina Clinica e Chirurgia – Oncologia, Naples, Italy, 16Medical Oncology, IRCCS AOU San Martino - IST-Istituto Nazionale per la Ricerca sul Cancro, Genoa, Italy, 17Medical Oncology, Ospedale di Prato Sandro Pitigliani Med. Oncology Unit, Prato, Italy, 18Medical Oncology, Azienda Ospedaliera Arcispedale Santa Maria Nuova - IRCCS, Reggio Emilia, Italy, 19Medical Oncology, Ospedale Universitario Verona, Verona, Italy Background: ESO/ESMO develops consensus guidelines for MBC treatment every 2 years, potentially applicable worldwide. Aim of the present analysis is to verify the adherence to ABC recommendations for HER2-ve MBC in the context of the AMBRA study. Methods: AMBRA is a longitudinal cohort study, aiming to describe the choice of first and subsequent lines of treatment in HER2-ve MBC pts receiving at least one Chemotherapy (CHT) (SABCS 2016, P5-15-07 & P5-14-09). For the present analysis, we selected 4 statements from the ABC1 & ABC2 Conferences, comparing them with the clinical choices of 1st-line therapy in all evaluable cases. Results: So far, 791 pts have been enrolled, of whom 586 (74.1%) have received CHT as 1st-line treatment. 479 pts (81.7%) had Luminal tumours at diagnosis. First-line CHT was monotherapy in 89 pts (25.2%), anthra-based CHT in 45 (7.7%) and platinumbased in 38 (6.5%), mainly TNBC (20/38, 52.6%). Selected recommendations and percentages of adherence are reported in Table. Conclusions: The adherence to clinical recommendations is very low, being only partially applied in all the clinical situations analyzed. Educational interventions are urgently needed and a confrontation with the clinical community is strongly recommended.

doi:10.1093/annonc/mdx365 | 83

abstracts

Annals of Oncology

Table: 259P Adherence to ABC recommendations for HER2-ve pts

Table: 260P Three-year cancer-speci˚c survival according to Mbioscore

n/N (%of adherence)

M-bioscore 1 - Anthracycline (A)- or Taxane (T)based regimens, preferably as a single agent, would usually be considered as ﬁrst-line CHT, in those pts who have not received these regimens as adjuvant treatment 2 - In pts with taxane-naive and anthracycline-resistant MBC, taxane-based therapy, preferably as a single agent, would usually be considered as the treatment of choice 3 - Other options are Capecitabine and Vinorelbine 4 - If given in the adjuvant setting, a taxane can be re-used in the metastatic setting, particularly if there has been at least 1 year of disease-free survival

A/T mono 59/ 270 (21.8%) A/T poly 101/ 270 37.4%)

160/270 (59.2%)

T mono 89/ 431 (20.6%) T poly 119/ 431 (27.6%)

208/431 (48.2%)

135/586 (23.0%) DFI > 12 months 74/79 (93.7%) DFI 30% of pts and are associated w significant neurologic morbidity and mortality. Current treatment strategies for BM primarily utilize radiotherapy (RT) and in selected instances surgical resection. Tucatinib is a highly selective oral HER2þ tyrosine kinase inhibitor that has shown promising results in HER2þ MBC both in pts w and w/o BM. Methods: Two Phase 1b studies of tucatinib were pooled to compare PFS and sites of relapse in pts w or w/o BM. Results: 77 pts were analyzed, all treated at the recommended Phase 2 dose of 300mg PO BID of tucatinib, 50 in the 004 trial (tucatinib þ T-DM1) and 27 in the 005 trial (tucatinib þ trastuzumab þ capecitabine). All pts were heavily pretreated w a median of 3 prior therapies including a taxane, trastuzumab, pertuzumab, T-DM1 and lapatinib. Four cohorts of pts were identified: 46% (35) had systemic metastases only, 17% (13) had previously treated (RT w or w/o surgery) and stable BM, 19% (15) had previously treated and progressive BM and 17% (13) had asymptomatic untreated BM demonstrated by screening MRI. Median PFS across cohorts was 8.5, 6.1, 9.0 and 7.1 months, respectively. No statistically significant difference in PFS was seen when comparing the non-BM cohort to all BM cohorts (median of 8.5 vs. 6.7 months; p ¼ 0.65). The risk of progression in brain in pts w baseline BM was 48.8% overall (41.5% in brain only; 7.3% in brain and body) compared to an 11.1% overall (8.3% in brain only; 2.8% in brain and body) in pts w/o baseline BM. Conclusions: 54% of pts entered tucatinib studies w baseline BM, either previously treated (stable or progressive) or untreated. The cohorts of pts analyzed appear to differ only in the site of disease progression. Although pts w/o baseline BM primarily have progression in extraneural sites and pts w baseline BM primarily have progression in the CNS, PFS is comparable across cohorts. Furthermore, pts both w and w/o BM have durable responses w these combination therapies following multiple lines of prior HER2 targeted therapy. These data support the use of tucatinib in both pts w and w/o BM in the accruing HER2CLIMB trial. Clinical trial identification: ONT-380-004 and ONT-380-005 Legal entity responsible for the study: Cascadian Therapeutics, Inc. Funding: Cascadian Therapeutics, Inc. Disclosure: All authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx365 | 85

abstracts 265P

Survival of patients with aromatase inhibitors sensitive, HR1/HER2metastatic breast cancer treated with a first-line endocrine therapy or chemotherapy in a multicenter national observational study

E. Jacquet1, A. Lardy-Cleaud2, B. Pistilli3, P. Cottu4, S. Delaloge5, M. Debled6, L. Vanlemmens7, G. Anne-Vale´rie8, M. Leheurteur9, L. Laborde10, W. Jacot11, D. Berchery12, B. Coudert13, J-M. Ferrero14, D. Parent15, V. Die´ras16, M. Velten17, C. Courtinard18, M. Robain19, T. Bachelot20 1 Oncology, Centre Le´on Be´rard, Lyon, France, 2Statistics, Centre Le´on Be´rard, Lyon, France, 3Oncology, Gustave Roussy, Villejuif, France, 4Medical Oncology, Institut Curie, Paris, France, 5Breast Cancer Group, Institut Gustave Roussy, Villejuif, France, 6Medical Oncology, Institute Bergonie´, Bordeaux, France, 7Medical Oncology, Centre Oscar Lambret, Lille, France, 8Medical Oncology, Centre Franc¸ois Baclesse, Caen, France, 9 Oncologie Me´dicale, Centre Henri Becquerel, Rouen, France, 10Unite´ de gestion des donne´es, Institut Paoli Calmettes, Marseille, France, 11Department of Medical Oncology, ICM Regional Cancer Institute of Montpellier, Montpellier, France, 12Epidemiology, Institut Claudius Regaud, Toulouse, France, 13Oncology, Centre Georges-Franc¸ois Leclerc (Dijon), Dijon, France, 14Department of Medical Oncology, Centre Antoine Lacassagne, Nice, France, 15Medical Oncology, Institut Jean Godinot, Reims, France, 16 Department of Medical Oncology, Center Euge`ne Marquis, Rennes, France, 17Medical Oncology, Centre Paul Strauss, Strasbourg, France, 18Research and Development, UNICANCER, Paris, France, 19ESME Program, UNICANCER, Paris, France, 20De´partement d’Oncologie me´dicale adulte, Centre Le´on Be´rard, Lyon, France Background: For HRþ/HER2– metastatic breast cancer (mBC), international guidelines recommend the use of endocrine therapy (ET) as first-line (L1) treatment except in case of “visceral crisis” for which chemotherapy (CT) is advised. Few studies directly compare these two treatment options. In 2014, UNICANCER launched the Epidemiological Strategy and Medical Economics (ESME) Research program to centralize real-world data in oncology. We sought to use this database to study this question. Methods: All patients (pts) who initiated treatment for a newly diagnosed mBC between January 2008 and December 2014 in all 18 French Comprehensive Cancer Centers were included in the ESME mBC database. ESME Research program centralized all existing data using retrospective data collection. Primary endpoint of the present study was progression free survival (PFS1) and overall survival (OS) according to L1 treatment for aromatase inhibitors sensitive (AIS) HRþ/HER2mBC pts. Results: 6265 pts out of 16703 in ESME, had AIS HRþ/HER2- mBC. As L1 therapy, 2733 pts (43.6%) received ET alone, while 3532 received CT (56.4%). Among these 3532 pts, 2073 (58.7%) received ET as maintenance treatment after CT. A Cox multivariate analysis with significant prognostic variables identified a lower risk of death in the patients with L1 ET (HR ¼ 0.839, 95% IC [0.772-0.911], p < 0.0001). Patients receiving CT were younger (median age 56.0 vs 66.0, p < 0.001), more likely to have visceral metastasis (61.6% vs 40.1%, p < 0.001) and SBR III primary tumors (31.3% vs 18.8%, p < 0.001). Median PFS1 was 15.18 months for L1 ET (95% CI, 14.45-16.20) vs 12.58 months for L1 CT þ/- hormone maintenance (95% CI, 11.89-13.14), p < 0.0001. Median OS was 60.78 months for L1 ET (95% CI, 57.16-64.09) vs 49.64 months for L1 CT (95% CI, 47.31-51.64), p < 0.0001. Conclusions: The results show that despite guidelines, a majority of AIS HRþ/HER2mBC pts still received CT as first-line treatment in the past years. PFS1 and OS data do not suggest any advantage of this aggressive strategy over ET alone. Advanced statistical methods using the propensity score will be presented in order to control for potential selection bias. Legal entity responsible for the study: UNICANCER Funding: UNICANCER Disclosure: All authors have declared no conflicts of interest.

266P

Use of everolimus in advanced hormone receptor positive metastatic breast cancer in a multicenter national observational study

A. Lardy-Cleaud1, P. Cottu2, S. Frank2, O. Le Saux3, S. Chabaud1, D. Parent4, B. Pistilli5, M. Debled6, A. Mailliez7, C. Veyret8, T. Petit9, L. Uwer10, S. Guiu11, M. Ung12, E. Chamorey13, P. Arveux14, T. Guesmia15, P. Augereau16, G. Simon15, T. Bachelot17 1 Biostatistics, Centre Le´on Be´rard, Lyon, France, 2Medical Oncology, Institut Curie, Paris, France, 3Medical Oncology, Centre Le´on Be´rard, Lyon, France, 4Medical Oncology, Institut Jean Godinot, Reims, France, 5Medical Oncology, Gustave Roussy, Villejuif, France, 6Medical Oncology, Institute Bergonie´, Bordeaux, France, 7Medical Oncology, Centre Oscar Lambret, Lille, France, 8Medical Oncology, Centre Henri Becquerel, Rouen, France, 9Medical Oncology, Centre Paul Strauss Centre de Lutte contre le Cancer, Strasbourg, France, 10Medical Oncology, Institut de Cance´rologie de Lorraine - Alexis Vautrin, Vandoeuvre Les Nancy, France, 11Medical Oncology, Institut du Cancer de Montpellier, Mont Pellier, France, 12Medical Oncology, ’Institut Universitaire du Cancer de Toulouse, Toulouse, France, 13Pharmacy, Centre Antoine Lacassagne, Nice, France, 14 Biostatistics, Centre Georges Franc¸ois Leclerc, Dijon, France, 15R&D, UNICANCER, Paris, France, 16Medical Oncology, Centre Paul Papin, Angers, France, 17De´partement d’Oncologie me´dicale adulte, Centre Le´on Be´rard, Lyon, France Background: The everolimus-exemestane combination has been included in the International guidelines for advanced HRþ breast cancer (mBC) since the results of the Bolero-2 trial. Marketing authorization has been granted in France in July 2012. We

86 | Breast cancer, metastatic

Annals of Oncology evaluated the incidence and indication of everolimus (EVE) use before and after marketing authorization and reimbursement. Methods: All patients who initiated treatment for a newly diagnosed mBC between 01/ 2008 and 12/2015 in all 18 French Comprehensive Cancer Centers have been included in the real life ESME database, which collects retrospective data using a clinical triallike methodology. Results: The ESME program included a total of 16,703 patients of which 9,921 had HRþ/HER2- mBC. Median age at metastatic diagnosis was 62.0 year (range 23-96). Visceral metastases were present in 60.3% of cases. Only 4123 patients (41.6%) received endocrine therapy alone as first-line therapy, and 60% were deemed endocrine resistant. Overall, 1,217 (12.3%) pts have received EVE during therapy as of Dec. 2015 (all lines). EVE was given as first line therapy in 117 pts (10% of all EVE pts and 1.2% of pts receiving a first line therapy). In 99/117 pts (85%) EVE was combined with exemestane. Before 2012, EVE was used within clinical trials. After 2012, use of EVE increased steadily. Percentages in the Table refer to the total of pts who received any kind of treatment during a given year of observation (eg 506 pts took EVE in 2015 out of 4435). Median duration of EVE use was 6.0 months (range 0-65) as first line treatment and 3.9 months (range 0-65) in pretreated patients. Patient population and causes of EVE cessation will be detailed at the meeting.

Table: 266P Year

2008

2009

2010

2011

2012

2013

2014

2015

N %

4 0.20

7 0.22

13 0.30

11 0.21

133 2.3

391 6.6

493 8.6

506 11.41

Conclusions: In this very large French national and representative cohort of HRþ HER2- mBC, EVE use rose quickly as soon as marketed. EVE was mostly used in pretreated mBC albeit in probably too advanced pts. These data underline the need for physician and patient education for oral therapies. Legal entity responsible for the study: UNICANCER R&D Funding: UNICANCER Disclosure: All authors have declared no conflicts of interest.

267P

Can contemporary trials in HER2-negative metastatic breast cancer (mBC) detect overall survival (OS) benefit?

S. Ku¨mmel1, C. Jackisch2, V. Mu¨ller3, A. Schneeweiss4, S. Klawitter5, M.P. Lux6 Cancer Center Essen, Klinikum Essen Mitte, Essen, Germany, 2Obstetrics and Gynecology, Sana Clinic Offenbach, Offenbach, Germany, 3Department of Gynecology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, 4Division Gynecologic Oncology, National Center for Tumor Diseases (NTC) University Hospital, Heidelberg, Germany, 5Biostatistics and Epidemiology, Roche Pharma AG, Grenzach Wyhlen, Germany, 6OBGYN-Department, University Breast Center Franconia, University Hospital Erlangen, Erlangen, Germany

1

Background: Although several recent trials have demonstrated improved progressionfree survival (PFS) or time to progression (TTP) with first-line regimens for HER2negative mBC, OS benefit is often elusive. We calculated required sample sizes to power for OS based on published data in recent trials. Methods: Randomised superiority trials of first-line chemotherapy or targeted therapy for HER2-negative mBC including >150 patients, meeting the primary efficacy objective and published between 2000 and 2014 were identified. The sample sizes required to power for PFS (or TTP) and OS were calculated retrospectively for each trial using the observed results and study/recruitment follow-up durations (alpha¼0.05, 2-sided logrank test, 80% power), and summarised as a factor relative to the actual sample size (x < 1 required x-fold fewer cases to show the same gain; x > 1 required x-fold more cases). Results: Only 8 of the 14 identified trials reported all information required for retrospective sample size calculation (Table). Most would have required a far larger sample size to demonstrate an OS gain (x: 0.5–2479) with the observed results. In 10 of 13 trials, the sample size required to power for OS was at least 5-fold larger than that needed to power for PFS. Conclusions: Designing trials to test potential new treatments for HER2-negative mBC is challenging and requires a balance of regulatory acceptability, feasibility and realistic medical assumptions to calculate sample sizes, which can be particularly difficult in heterogeneous study populations with long post-progression survival and heterogeneous subsequent therapies. However, ongoing and future trials of cancer immunotherapy (new mode of action) focusing on triple-negative mBC (a more homogeneous population with shorter OS and post-progression survival, and fewer treatment options) may show a new pattern. Legal entity responsible for the study: Roche Pharma AG Funding: Roche Pharma AG Disclosure: S. Ku¨mmel, C. Jackisch: Membership on advisory board: Roche Pharma AG. V. Mu¨ller: Membership on advisory board or board of directors: Amgen,

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology

Table: 267P Summary of results Study

Total No. of patients in trial (randomisation if not 1:1)

Ackland 2001 Jassem 2001 Ejlertsen 2004 Bontenbal 2005 Feher 2005 von Minckwitz 2005 Paridaens 2005 Albain 2008 Gray 2009 Sparano 2009 Miles 2010 Robert 2011 Gligorov 2014 Lorusso 2014

460a 267a 387a 216 397a 364a 331a 529 722 751 488 (1:1:1) 615 (1:2) 185 233

Observed median, months (arm A v B)

Retrospectively calculated sample size

Factor (x)

PFS

TTP

OS

PFS/TTP

OS

OS sample size/N

OS/PFS sample size

– – 8.2 v 10.1b – – – 3.9 v 7.5b – 5.8 v 11.3b – 8.2 v 10.1b 5.7 v 8.6b 4.3 v 11.9b –

6.3 v 8.7b 6.2 v 8.3b – 6.6 v 8.0b 3.4 v 6.1b 6.7 v 8.2b – 4.0 v 6.1b – 7 v 9.8b – – – 7.8 v 9.4b

18.2 v 20.1 18.3 v 23.3b 18.0 v 19.1 16.2 v 22.6b 11.8 v 19.1b – 15.6 v 18.3 15.8 v 18.6b 24.8 v 26.5 20.6 v 20.5 31.9 v 30.2 22.8 v 25.7 23.7 v 39.0b 28.0 v 30.1

360 592 788 1022 98 820 80 168 72 294 802 234 34 972

5906 1792 11 988 476 212 – 1882 1404 8262 1 862 010 20 802 3840 242 9654

12.8 6.7 31.0 2.2 0.5 – 5.7 2.7 11.4 2479.4 42.6 6.2 1.3 41.4

16.4 3.0 15.2 0.5 2.2 – 23.5 8.4 114.8 6333.4 25.9 16.4 7.1 9.9

a

Duration of accrual and/or follow-up not reported; accrual period assumed to be 1/3 of study duration. Statistically signiﬁcant.

b

AstraZeneca, Celgene, DaiichiSankyo, Eisai, Pfizer, Pierre-Fabre, Novartis, Roche, Teva, Janssen-Cilag, Genomic Health, Roche, Pierre Fabre, Amgen, Daiichi-Sankyo and Eisai. S. Klawitter: Roche Pharma AG (Employment). M.P. Lux: Membership on advisory board: Roche, Novartis, Pfizer, MSD, AstraZeneca; Corporate-sponsored research: Roche, Novartis, MSD, Celgene, Amgen. All other authors have declared no conflicts of interest.

268P

What are the treatment patterns and overall survival (OS) in patients with metastatic triple-negative breast cancer (mTNBC) in US clinical practice?

P. Bajaj1, D. Latremouille-Viau2, A. Guerin2, C. Reyes1, A. Stein1, A. Kurian3, P. Cortazar1 US Medical Affairs, Genentech Inc. - Roche - USA, South San Francisco, CA, USA, 2 Analysis Group, Montreal, QC, Canada, 3Division of Oncology, Stanford University School of Medicine, Stanford, CA, USA

1

Background: mTNBC is a condition with significant unmet medical need due to limited treatment (tx) options and short survival after chemotherapy (chemo) failure. Tx patterns and OS in patients (pts) with mTNBC in US clinical practice were assessed. Methods: This retrospective chart review study used data collected in October and November 2016 from US oncologists via an online case report form. Data of adult pts who were initiated on a pharmacological tx after a diagnosis of recurrent or de novo mTNBC between January 2012 and June 2015 were reviewed. Tx regimens used in 1stand 2nd-lines for mTNBC were summarized. Median OS from 1st- and 2nd-line tx initiation were estimated using Kaplan-Meier analyses. Results: 125 oncologists provided data on 411 mTNBC pts; mean age was 57 years; 298 (73%) had 2 lines of tx; 256 different tx sequences were identified. Mean duration of tx was 7.5 months (mos) in 1st-line and 7.3 mos in 2nd-line. The most prevalent tx regimens were single agent chemo with taxane (22%) in 1st-line and capecitabine (Cap; 20%) in 2nd-line. Median OS was 16.7 mos in 1st-line and 14.2 mos in 2nd-line. Conclusions: Findings suggest that, in US clinical practice, there is substantial heterogeneity in the tx of pts with mTNBC. Pts had a median OS < 18 mos after tx initiation. A better understanding of the mTNBC pt subpopulation and optimal tx sequencing is warranted to improve tx strategies and prolong survival. Legal entity responsible for the study: Genentech, Inc. Funding: Genentech, Inc. Disclosure: P. Bajaj, C. Reyes, A. Stein, P. Cortazar: Employee of Genentech, Inc. and owner of Roche stock. D. Latremouille-Viau, A. Guerin: Employee of Analysis Group, Inc., which has received consulting fees and research funding from Genentech, Inc. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

Table: 268P st

1 -line, N 5 411

%

Median OS [95% CI], mos

All Single agent Combination Tx regimen Single agent taxane (docetaxel/paclitaxel (pac)/ nab-pac) Anthracycline (ATC; doxorubicin [Dox]/epirubicin/liposomal Dox) þ cyclophosphamide þ/taxane Platinum (carboplatin/cisplatin/oxaliplatin) þ taxane Bevacizumab-containing Gemcitabine (Gem) þ platinum Cap Other 2nd-line, N 5 298 All Single agent Combination Tx regimen Cap Platinum-containing (w/o taxane) Single agent taxane Taxane-containing combination Eribulin ATC-containing (w/o taxane or platinum) Gem Other

45 55

16.7 [15.2; 18.0] 15.6 [13.2; 18.6] 17.0 [15.1; 19.1]

22

–

14

–

13

–

10 10 9 21

– – – –

72 28

14.2 [10.5; 22.3] 12.9 [10.5; 22.3] 16.2 [8.1; -]

20 13 13 12 9 8 7 16

– – – – – – – –

doi:10.1093/annonc/mdx365 | 87

abstracts 269P

Annals of Oncology

Outcomes of systemic therapy for advanced triple-negative breast cancer: A single centre experience

N. Battisti1, D. Okonji1, T. Manickavasagar1, K. Mohammed2, M. Allen1, A. Ring1 Breast Unit, The Royal Marsden NHS Foundation Trust, London, UK, 2Research and Development Department, The Royal Marsden NHS Foundation Trust, London, UK

1

Background: Systemic treatment outcomes for advanced triple-negative breast cancer (aTNBC) are worse compared to other disease subtypes, due to aggressive behaviour, heterogeneity and lack of molecular targets. Many options are under investigation although most patients receive standard cytotoxic chemotherapy. We aimed to provide better insight into the efficacy of different lines of therapy for aTNBC (overall response rate [ORR], median progression-free survival [mPFS] and median overall survival [mOS]) to better inform discussion with patients, decision-making and referral for clinical trials. Methods: We retrospectively identified 268 patients diagnosed with aTNBC from 01/ 12/2011 to 30/11/2016 from our electronic records. Patients’ and tumour characteristics were recorded, along with systemic treatment outcomes. Chi-squared/Fishers exact test and Kaplan-Meier statistical methods were utilised. Results: 186 patients treated with 1 line of palliative systemic treatment were eligible for the analysis with a median age at 55 (range 26-91). 53.8% had ECOG Performance Status 0 and 69.9% visceral involvement. 38.6% had a disease-free interval (DFI) 3 BM. Number of BM (1,2,3,>3) was significantly associated with OS (p < 0.001). Both breast-GPA and modified breast-GPA predicted OS (Table). Concordance indices were 0.641 (95% CI, 0.6405 to 0.6422) and 0.667 (95% CI, 0.6662 to 0.6678) for breast-GPA and modified breast-GPA, respectively (p < 0.001). Conclusions: Number of BM is a significant prognostic factor in BC patients with BM and modified breast-GPA performs better than breast-GPA in predicting prognosis of these patients. Legal entity responsible for the study: Department of Surgery, Oncology and Gastroenterology, University of Padova, Padua, Italy Funding: None Disclosure: All authors have declared no conflicts of interest.

Table: 273P Breast GPA category

Number of patients (%)

Median OS, months (95% CI)

Hazard Ratio (95% CI)

p

3.5-4 2.5-3 1.5-2 0-1.0 Modiﬁed breast GPA category 3.5-4 2.5-3 1.5-2 0-1.0

86 (13.5%) 248 (38.8%) 194 (30.4%) 111 (17.4%) Number of patients (%) 37 (5.8%) 209 (32.8%) 257 (40.3%) 135 (21.2%)

18.8 (14.5-22.6) 10.3 (8.8-11.8) 6.2 (4.9-7.6) 2.5 (1.8-3.2) Median OS, months (95% CI) 18.9 (17.2-20.5) 15.2 (12.1-18.3) 7.9 (5.9-9.9) 2.3 (1.9-2.8)

ref 1.45 (1.12-1.88) 2.04 (1.56-2.66) 4.97 (3.67-6.71) Hazard Ratio (95% CI) ref 1.43 (0.98-2.09) 2.30 (1.59-3.33) 7.03 (4.72-10.46)

120 cell lines of various malignancies CB-103 was active on a

Volume 28 | Supplement 5 | September 2017

subset of 24 cancer cell lines, including different solid tumours (breast, lung, sarcomas), lymphomas and leukaemias. Results: Moreover, CB-103 demonstrated anti-NOTCH activity in the Triple-Negative Breast Cancer HCC1187 cell line, being resistant to GSIs due to a NOTCH2 chromosomal translocation. In addition, CB-103 exhibited anti-tumour efficacy in multiple in vivo models and patients derived xenograft models. Conclusions: Safety pharmacology and toxicology studies have been completed and revealed an excellent non-clinical safety profile of CB-103. A first-in-human Phase I/IIA clinical study in advanced solid tumours and haematological malignancies is under preparation. Legal entity responsible for the study: Cellestia Biotech AG Funding: Cellestia Biotech AG Disclosure: D. Weber: Chief Medical Officer of Cellestia, co-founder, stock ownership. R. Lehal: Chief Scientific Officer, co-founder, stock ownership. J-P. Bourquin: Medical advisory board Cellestia. M. Bauer: Chief Executive Officer, co-founder, stock ownership. M. Murone: Chief Operating Officer, co-founder, stock ownership. F. Radtke: Chairman of the Board, co-founder, stock ownership. All other authors have declared no conflicts of interest.

412P

Design and development of potent E1 ubiquitin activating enzyme inhibitor, CPL-410-005, as novel anticancer therapy

A. Stanczak, A. G ornicka, B. Stypik, M. Mroczkiewicz, J. Pieczykolan, K. Dubiel, M. Wieczorek Preclinical Development Department, Celon Pharma S.A., Lomianki, Poland Background: The ubiquitin-proteasome system is crucial in tumorigenesis. The division rate of cancer cells, thus protein synthesis, is increased in comparison to normal ones, what sensitizes tumors for any protein changes. Proteasome inhibitor - bortezomib was the first inhibitor registered in treatment of blood cancers. However, drugs beneficial for solid tumors are still missing. Therefore targeting of E1 enzyme as a start of UPS pathway may serve as a promising anticancer therapy. Methods: We have designed a novel E1 small molecule inhibitor, CPL-410-005. The inhibitory potency of compound was assessed on purified E1 enzyme, using biochemical assay. Assays to measure cellular poliubiquitylation or ubiquitin-like modifications level were developed. The compound’s biological activity and selectivity was evaluated in a number of cancer cells using cell viability assays, Western Blot and flow cytometry, analyzing programmed cell death, unfolded protein response or cell cycle inhibition. Results: CPL-410-005 inhibits E1 enzyme with greater potency than MLN7243. This results in cellular polyubiquitylation inhibition, while the impact on other ubiquitinlike modifications (neddylation, sumoilation) is minor. Tumor proliferation rate inhibition was pronounced in reference to the non-malignant cells [IC50 values of 20 nM for HCT-116 cells vs IC50 values of 400 nM for HEK293 cells]. Moreover, a higher level of unfolded protein response or programmed cell death was observed in cells treated with CPL-410-005 in reference to MLN7243 (5% apoptotic cells vs 30% for MLN7243 vs CPL-410-005, respectively). In case of CPL-410-005, apoptosis rate was higher in tumor than in normal cells (80% apoptotic cells vs 20%, respectively). A high throughput study was performed, to determine the activity of CPL-410-005 on 120 human tumor cell lines, showing that >85% tested cell lines responded with IC50 value below 100nM. The initial in vivo studies on tumor human xenografts are ongoing. Conclusions: We have designed and evaluated in vitro a potent E1 inhibitor - CPL410-005, which shows promising in vitro activity. Further preclinical studies are necessary to develop this compound as a novel anticancer therapy. Legal entity responsible for the study: Celon Pharma Funding: Celon Pharma Disclosure: A. Stanczak, A. G ornicka, B. Stypik, M. Mroczkiewicz, J. Pieczykolan, K. Dubiel: Full-time employee of Celon Pharma S.A., Lomianki, Poland. M. Wieczorek: Chief Executive Officer of Celon Pharma S.A., Lomianki, Poland.

413P

RX-3117, a novel hypomethylating agent, shows promising therapeutic activity in combination with nab-paclitaxel and checkpoint inhibitors in preclinical models

J. Frank1, Y.B. Lee1, D.J. Kim1, E. Benaim2 Research, Rexahn Pharmaceuticals, Rockville, MD, USA, 2Clinical Trial, Rexahn Pharmaceuticals, Rockville, MD, USA

1

Background: A novel nucleoside analogue, RX-3117, is being evaluated in a Phase IIa study in patients with advanced pancreatic and bladder cancer. RX-3117 shows promising antitumor activity in xenografts including patient-derived xenografts resistant to gemcitabine. Here we demonstrate the preclinical effects of combination therapy with RX-3117 þ Abraxane or anti-PD1 immunotherapy. Methods: One colorectal (MC38), pancreatic (Pan02) syngeneic xenograft and patientderived pancreatic (CTG-0723) xenograft model were exposed to 60 mg/kg RX-3117 po, 5 days on, 2 days off for three weeks. Pan02 and MC38 received RX-3117 alone or in combination with 100ug anti-PD1, ip. PDx CTG-0723 received one cycle of RX-3117, followed by a second cycle of RX-3117 þ 10 mg/kg Abraxane, iv. MC38 tumorinfiltrating lymphocytes were measured at days 5 and 12 with RX-3117. Results: In MC38 at day 28, RX-3117 or anti-PD1 showed TGIs of 90% and 93%, whereas the combination showed 99% TGI. Differences were also observed in TILs.

doi:10.1093/annonc/mdx367 | 137

abstracts Relative to vehicle (CD4þ:10.6þ/-1.6, CD8þ: 8.6þ/-1.1), %CD4 þ (17.4þ/-1.4) and CD8þ cells (12.3þ/-1) increased. %MDSCs decreased on Day 5 in blood (42þ/-7.7 vs 29þ/-6). %CD8þ increased (9.6þ/-3.3 vs 12.3þ/-3.2) and %MDSC decreased (15.4 þ/- 3.7 vs 10.6 þ/- 3.3) in tumor on Day 12. In Pan02, RX-3117 þ anti-PD1 resulted in a day 32 TGI of 60%. Anti-PD1 alone had a day 32 28% TGI. In CTG-0723, the first cycle of RX-3117 at 10, 30 and 60 mg/kg produced TGIs of 33%, 46% and 77%. The second cycle, RX-3117 þ Abraxane, day 46 TV showed TGIs of 55%, 58% and 83%. Conclusions: We demonstrate the antitumor effect of RX-3117 as a single agent and in combination with Abraxane or anti-PD-1. The combination of RX-3117/anti-PD1 in MC38 produced 7 tumor-free survivors out of 10 compared to 2 of 10 by anti-PD1 alone, indicating RX-3117 may mobilize the right population of lymphocytes to enable anti-PD-1 to work more effectively. In Pan02, RX-3117 exhibited better TGI than antiPD-1. In CTG-0723, the combination of RX-3117 and Abraxane showed additive TGI. These studies demonstrate the therapeutic potential of RX-3117 in multiple cancers and validate the combination of RX-3117 with anti-PD1 in several cancer types. Legal entity responsible for the study: Rexahn Pharmaceuticals, Inc. Funding: Rexahn Pharmaceuticals, Inc Disclosure: J. Frank, Y.B. Lee, D.J. Kim: Employee of Rexahn Pharmaceuticals. E. Benaim: Officer at Rexahn Pharmaceuticals, employee, stock-holder.

414P

A phase Ib trial of JX-594 (Pexa-Vec), a targeted multimechanistic oncolytic vaccinia virus, in combination with low-dose cyclophosphamide in patients with advanced solid tumors

M. Toulmonde1, S. Cousin1, A. Bessede2, M. Homerin3, N. Stojkowitz3, M. Lusky4, M. Pulido5, A. Italiano1 1 Medical Oncology, Institute Bergonie´, Bordeaux, France, 2IMMUSMOL, Immusmol, Bordeaux, France, 3Clinical Research, Transgene SA, Illkirch-Graffenstaden, France, 4 Project Management, Transgene SA, Illkirch-Graffenstaden, France, 5Biostatistics, Institute Bergonie´, Bordeaux, France Background: JX-594 (Pexa-Vec) is a targeted oncolytic vaccinia virus designed to selectively replicate in and destroy cancer cells with epidermal growth factor receptor (EGFR)/ras pathway activation. Direct oncolysis plus GM-CSF expression is accompanied by tumor vascular disruption and anti-tumoral immunity. JX-594 was welltolerated intravenously (IV) and intratumorally (IT). Given the immunomodulatory effects of low-dose cyclophosphamide (CP), anti-tumor synergy is predicted with JX594. Methods: CP was delivered orally at the dose of 50 mg BID one week on one week off. JX-594 was delivered IV at day 8 of each day 28-cycle. 2 dose levels of JX-594 were explored: 3.108 and 1.109 plaque forming units (pfu). The primary objective of the study was to determine the safety of JX-594 in combination with low-dose CP in patients with advanced solid tumors. Secondary objectives include response rates, PFS, pharmacokinetics and pharmacodynamics. Results: Ten patients entered the study. 9 were evaluable for safety. No dose limiting toxicity was observed. The combination regimen was well-tolerated. The most frequent adverse events were grade 1-2 fever/transient flu-like symptoms (n ¼ 10), grade 1-2 nausea (n ¼ 5), grade 1-2 anemia (n ¼ 4) and grade 1-2 fatigue (n ¼ 4). 2 patients (breast cancer, ovarian cancer) had stable disease as best overall response. Conclusions: IV JX-594 was well-tolerated in combination with low-dose CP. PK and PD (immunological profiling) will be presented at the meeting. Two phase 2 studies are ongoing in patients with advanced HER2 negative breast cancer and advanced softtissue sarcomas, respectively. Clinical trial identification: NCT02630368 Legal entity responsible for the study: Institut Bergonie´ Funding: INCA Disclosure: All authors have declared no conflicts of interest.

415P

Population pharmacokinetic analysis of OT-101 (trabedersen) in patients with advanced tumors

W. Wang1, K. Ng2, D. Nam2, V. Trieu1, L. Hwang1 Clinical, Oncotelic Inc, Agoura Hills, CA, USA, 2Clinical, Autotelic Inc., Costa Mesa, CA, USA

1

Background: OT-101 (Trabedersen) is a phosphorothioate antisense oligodeoxynucleotide specifically inhibiting the expression of transforming growth factor-beta 2 (TGF-b2), whose overexpression is a pivotal factor for malignant progression in solid tumors. In the clinical Phase I/II study, plasma pharmacokinetic (PK) profile of OT101 administered intravenously was evaluated in patients with advanced tumors. A population PK model was built to further understand the factors contributing to the variability in PK of OT-101. Methods: A total of 61 patients with pancreatic cancer (n ¼ 37), malignant melanoma (n ¼ 19), or colorectal carcinoma (n ¼ 5) were treated with OT-101 with escalating doses in 2 treatment schedules (1st schedule: 7-days-on/7-days-off; 2nd schedule: 4days-on/10-days-off; up to 10 cycles). The plasma concentration data of OT-101 were analyzed using nonlinear mixed-effect modeling (Phoenix NLME 7.0). The influence of age, gender, body mass index (BMI), body weight (BW), cancer type, treatment

138 | Developmental therapeutics

Annals of Oncology schedule, creatinine clearance (CrCl) and estimated glomerular filtration rate (eGFR) as covariates on PK was evaluated. Results: With exclusion of protocol deviations, the final analysis dataset contained 92 patient cycles and 1188 plasma samples. Twenty-six patient cycles were from 7-dayson/7-days-off schedule and 66 were from 4-days-on/10-days-off schedule. The concentration time course of OT-101 was best described by a two-compartment model with combination of additive and multiplicative error. The estimates of PK parameters were as follows: total body clearance, 41.79 mL/hr; distribution volume of the central compartment, 6.30 L; inter-compartmental clearance, 4.02 L/hr; distribution volume of the peripheral compartment, 5965.83 L. eGFR were identified as the covariates on OT-101 central compartmental clearance, with KeGFR as 11.45. Conclusions: The PK profile of OT-101 was best described by a two-compartment model. The model will be used with the sparse PK samples collected from the planned phase 3 clinical trial to calculate exposure measures for use in subsequent PK-PD analysis of efficacy. Legal entity responsible for the study: Oncotelic Funding: None Disclosure: W. Wang, V. Trieu, L. Hwang: Employee of Oncotelic Inc. K. Ng, D. Nam: Employee of Autotelic Inc.

417TiP

A phase 1 study of oral LOXO 292 in adult patients with advanced solid tumors, including RET-fusion non-small cell lung cancer, medullary thyroid cancer and other tumors with increased RET activity

A.E. Drilon1, T.M. Bauer2, V. Subbiah3, M. Cabanillas4, N. Lahkani5, L. Wirth6, G. Oxnard7, S. Smith8, T. Eary8, S. Cruickshank8, M. Nguyen8, S. Rothenberg8 1 Medical Oncology, Memorial Sloan-Kettering Cancer Center, New York, NY, USA, 2 Medical Oncology, Sarah Cannon Research Institute, Nashville, TN, USA, 3Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA, 4Endocrine Neoplasia & Hormonal Disorders, University of Texas M.D. Anderson Cancer Center, Houston, TX, USA, 5Medical Oncology, START Midwest, Grand Rapids, MI, USA, 6Hematology/ Oncology, Massachusetts General Hospital, Harvard University, Boston, MA, USA, 7 Thoracic Oncology, Dana Farber Cancer Institute, Boston, MA, USA, 8Research and Development, Loxo Oncology, Stamford, CT, USA Background: RET is a receptor tyrosine kinase with critical roles in normal physiology. Fusions of the RET kinase with a partner protein have been identified in 2% of nonsmall cell lung cancers (NSCLC) and a subset of papillary thyroid cancers and other tumors. RET mutations occur in the majority of medullary thyroid cancers (MTC). Although multikinase inhibitors with anti-RET activity are in the clinic, their activity is limited by incomplete RET inhibition in patients, toxicity from off-target effects (e.g.VEGFR2) and poor pharmacokinetics (PK). LOXO-292 is a potent and specific inhibitor of RET, including fusions, activating mutations and potential acquired resistance mutations, with minimal inhibition of off targets, including > 100-fold selectivity for VEGFR2. Trial design: This is an open label, multi-center, dose escalation and expansion Phase 1 study in adult patients with advanced solid tumors. Major eligibility criteria for dose escalation include prior cancer treatment (prior treatment with anti-RET TKIs allowed) and normal hematopoietic and major organ function. During dose escalation, when a dose level is achieved that is safe and consistent with RET target engagement, enrollment will be limited to patients with RET-fusion NSCLC, MTC and other tumors with RET alterations or increased RET activity, as identified in tumor or blood by prior molecular assays performed locally. Once the Maximum Tolerated Dose (MTD) or recommended dose for further study is identified, patients will be enrolled to one of five dose expansion cohorts, depending on tumor type (i.e. NSCLC, MTC, other cancer), prior TKI therapy and type of RET alteration. The starting dose of LOXO-292 is 20 mg orally once per day, and dose escalation is proceeding using a 3 þ 3 design. The primary endpoint is establishment of the MTD/recommended dose for further study. Key secondary endpoints include: safety and tolerability, PK parameters and preliminary assessment of anti-tumor activity. Patients undergo safety, clinical and PK assessments and radiographic evaluation for their disease at regular intervals. Clinical trial identification: Treatment of patients has begun. The study has been submitted to the NIH (https://clinicaltrials.gov) and the ClinicalTrials.gov Identifier is pending and will be provided as soon as available. Legal entity responsible for the study: Loxo Oncology Funding: Loxo Oncology Disclosure: S. Smith, T. Eary, S. Cruickshank, M. Nguyen, S. Rothenberg: Ownership interest in Loxo Oncology. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 418TiP

Phase 1/2 study of the selective TRK inhibitor larotrectinib, in pediatric patients with cancer

S. Gallego Me´lcon1, M. Casanova2, S. Bielack3, J. Chisholm4, C.S. van Tilburg5, N. Federman6, C.M. Albert7, L. Mascarenhas8, B. Turpin9, R. Nagasubramanian10, N. Shukla11, S. Spunt12, M.C. Cox13, D.S. Hawkins7, A.S. Pappo14, F. Doz15, G. Bisogno16, S.G. Dubois17, T.W. Laetsch18, B. Geoerger19 1 Pediatric Oncology, Hospital Universitario Vall d’Hebron, Barcelona, Spain, 2Pediatric Oncology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 3Oncology, Klinikum Stuttgart - Olgahospital, Stuttgart, Germany, 4Pediatric Oncology, The Royal Marsden Hospital, Sutton, UK, 5Pediatric Oncology, Heidelberg University Clinic and German Cancer Research Center, Heidelberg, Germany, 6Pediatric Oncology, University of California, Los Angeles, Los Angeles, CA, USA, 7Pediatric Oncology, Seattle Children’s Hospital, University of Washington, Fred Hutchinson Cancer Research Center, Seattle, WA, USA, 8Pediatric Oncology, Children’s Hospital Los Angeles, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA, 9Pediatric Oncology, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA, 10Pediatric Oncology, Nemours Children’s Hospital, Orlando, FL, USA, 11Pediatric Oncology, Memorial Sloan-Kettering Cancer Center, New York City, NY, USA, 12Pediatric Oncology, Stanford University, Palo Alto, CA, USA, 13Clinical Development, Loxo Oncology, South San Francisco, CA, USA, 14 Pediatric Oncology, St Jude Children’s Research Hospital, Memphis, TN, USA, 15 Pediatric Oncology, Institut Curie, Paris, France, 16Pediatric Oncology, Padova University, Padua, Italy, 17Pediatric Oncology, Dana-Farber/Boston Children’s Cancer and Blood Disorders Center, Boston, MA, USA, 18Pediatric Oncology, University of Texas Southwestern Medical Center at Dallas, Dallas, TX, USA, 19Pediatric Oncology, Institut Gustave Roussy, Villejuif, France Background: Neurotrophin ligands and their receptors TRKA, TRKB, and TRKC (encoded by NTRK1, NTRK2, and NTRK3) are important for growth regulation, differentiation and survival of neurons. Translocations involving the NTRK1/2/3 kinase domain have been described in a broad range of adult and pediatric tumors, including infantile fibrosarcoma (IFS), spindle-cell sarcoma, congenital mesoblastic nephroma, pediatric papillary thyroid cancer, high- and low-grade gliomas and Ph-like acute lymphoblastic leukemia. Larotrectinib is the first small-molecule selective inhibitor of TRKA, -B, and -C in clinical development and has demonstrated tumor growth inhibition in preclinical models and clinically meaningful and durable responses in patients with NTRK-translocated cancers in an adult phase 1 trial. Trial design: We have initiated an open-label, multi-center, international Phase 1/2 study with larotrectinib in pediatric patients with solid tumors and primary CNS tumors. A pediatric recommended phase 2 dose of 100mg/m2 (caped at 100mg BID) has been established. Enrollment to phase 2 began in April 2017 and is ongoing. For the phase 2 component, patients from 1-month of age with IFS or an NTRK-fusion positive tumor, including those who have not undergone definitive surgery are eligible. Patients who have not undergone definitive surgery are eligible as well. Larotrectinib is administered as an oral liquid formulation or capsules twice daily on a continuous 28-day schedule. Dosing is based on body surface area. The phase 2 portion enrolls patients with NTRK-translocated tumors and measurable disease into three cohorts: 1) infantile fibrosarcoma; 2) other extracranial solid tumors; and 3) primary CNS tumors. The primary endpoint is objective response rate, with duration of response and progression free survival as secondary efficacy endpoints. Quality of life measures and ctDNA are exploratory endpoints. Each phase 2 cohort will enroll in a single stage of up to 10 patients. Molecular abnormalities will be characterized through the analysis of archival tissue. Clinical trial identification: NIH: NCT02637687; EudraCT #: 2016-003498-16 Legal entity responsible for the study: Loxo Oncology, Inc. Funding: Loxo Oncology, Inc Disclosure: M.C. Cox: Employee and stockholder of Loxo Oncology, Inc. All other authors have declared no conflicts of interest.

419TiP

Chemosensitization of carboplatin by NOX66: Pharmacokinetics and safety

formulated using 400mg of idronoxil per 2.2g suppository. Patients are allocated to receive either one or two suppositories per day. Study Part 1: NOX66 PK: Patients receive NOX66 for 14 consecutive days as monotherapy, with a follow up period of 7 days post-dosing. Blood samples will be collected throughout the monotherapy arm to measure levels of idronoxil. If no significant adverse events are noted in this 21 day period, a patient will continue in the study. Study Part 2: NOX66 plus Carboplatin: Patients receive NOX66 at the same dose as in Part 1, for 7 days. Carboplatin is administered on Day 2 of treatment. Up to 6 cycles of chemotherapy are administered, at intervals of 28 days. For Cycles 1-3, low dose (AUC4) carboplatin is administered. Subject to safety review, standard dose (AUC6) carboplatin is administered for cycles 4-6. Safety assessment is continued throughout the study, with measures to identify efficacy signals (CT scan, ECOG) performed at baseline and after Cycles 3 and 6. Clinical trial identification: Trial Protocol Number: NOX66-001A Clinicaltrials.gov NCT02941523 Legal entity responsible for the study: Noxopharm Limited Funding: Noxopharm Limited Disclosure: I. Minns: Employee of Noxopharm Limited. G. Kelly: Member of the board of Directors, employee and a shareholder of Noxopharm Limited.

420TiP

M. Johnson1, B. Benson2, R. Wei3, R. Brachmann2, M.D. Galsky4 Tennessee Oncology, Sarah Cannon Research Institute, Nashville, TN, USA, 2Clinical Development, BeiGene USA, Inc, Emeryville, CA, USA, 3BioStatistics, BeiGene USA, Inc, Emeryville, CA, USA, 4Medical Oncology, Icahn School of Medicine at Mount Sinai, New York, NY, USA

1

Background: Poly (ADP-ribose) polymerase (PARP) proteins are a family of DNA binding and repair proteins and are thought to play a key role in the base excision repair of DNA damage generated by temozolomide (TMZ), a DNA-alkylating agent. PARP inhibitors (PARPis) represent a class of antitumor agents that exert their cytotoxic effects by inhibiting PARP activity. Some PARPis are capable of trapping PARP proteins on DNA, further augmenting cell death. BGB-290 is a potent and selective inhibitor of PARP1/2 and has demonstrated PARP trapping capacity. Synergistic cytotoxicity has been observed in vitro and in vivo when BGB-290 is combined with low dose TMZ. Trial design: This open-label, Phase 1b/2 dose-escalation/dose-expansion study is designed to evaluate BGB-290 at the recommended Phase 2 dose (60 mg administered orally twice daily [PO BID]) in combination with TMZ in patients with locally advanced and metastatic solid tumors. The phase 1b dose-escalation component will follow a 3 þ 3 design to establish the maximum tolerated dose (MTD) of TMZ in combination with BGB-290 in 50 patients with solid tumors. Dose escalation will evaluate the safety, tolerability and pharmacokinetics of BGB-290 (60 mg BID) plus escalating doses

Table: 420TiP Treatment arm

Tumor type

Estimated sample size

Cohort 1

Platinum-sensitive high grade epithelial, non-mucinous, ovarian cancer, fallopian cancer or primary peritoneal cancer with either known deleterious or suspected deleterious germline or somatic BRCA1/2 mutation or with DNA HRD Triple negative breast cancer with either known deleterious or suspected deleterious germline or somatic BRCA1/2 mutation or with DNA HRD Metastatic castration-resistant prostate cancer with either known deleterious or suspected deleterious germline or somatic BRCA1/2 mutation or with documented HRD Extended stage small cell lung cancer who have been treated with 2 prior regimens Gastric or gastroesophageal junction cancer who have been treated with 2 prior regimens

20

Cohort 2

I. Minns, G. Kelly Clinical Development and Medical Affairs, Noxopharm Limited, Sydney, Australia Background: The experimental anti-cancer drug, idronoxil, is a selective inhibitor of PI3K/AKT in tumor cells, with studies showing it to be a potent chemosensitizing agent of carboplatin in vitro and in animal studies across a wide range of cancer cell types. These results, however, have not translated to clinical efficacy, with a Phase 3 study of combined oral idronoxil and intravenous carboplatin in patients with late-stage platinum-refractory ovarian cancer discontinued with no efficacy seen. This lack of efficacy is now thought to be due to complete conversion of idronoxil to bio-inactive Phase 2 metabolites. NOX66 is a suppository formulation of idronoxil designed to protect the drug from Phase 2 metabolism. This first-in-human study will look at the ability of NOX66 to deliver relatively high levels of idronoxil in a bio-active form, investigating (a) PK and (b) safety of NOX66 administration both as a monotherapy and in combination with carboplatin. Trial design: This is an open label, Phase 1 PK and safety study of NOX66 as a monotherapy and in combination with carboplatin. Patients included have end stage, refactory solid tumours, and no further therapy options available. A total of 16 patients will be recruited into the study in two cohorts of 8 patients. NOX66 suppositories are

Volume 28 | Supplement 5 | September 2017

Phase 1b/2 study to assess the safety, tolerability, and clinical activity of BGB-290 in combination with temozolomide in patients with locally advanced or metastatic solid tumors

Cohort 3

Cohort 4 Cohort 5

20

20

20 20

doi:10.1093/annonc/mdx367 | 139

abstracts of TMZ administered once daily (QD) either on Days 1-7 (Arm A) or continuously (Arm B) of each 28-day cycle. The phase 2 component will further evaluate the safety, tolerability and antitumor activity of the recommended combination dose and schedule in 20 patients with one of five different tumor types (Table). Enrollment into these expansion cohorts will occur simultaneously and independent of each other. Subjects will continue to receive treatment in 28-day cycles until confirmed disease progression, intolerable toxicity, or discontinuation/withdrawal. Legal entity responsible for the study: Beigene Ltd. Funding: Beigene Ltd Disclosure: M. Johnson: Research funding compensation for consulting to the institution from OncoMed, BerGenBio, Lilly, and various Pharm/Biotech companies. Spouse is a contract lobbyist for Astellas and Otsuka Pharmaceuticals. B. Benson, R. Wei: Employee and stock holder of BeiGene USA. R. Brachmann: Employee of BeiGene USA. M.D. Galsky: Consulting fees from Genentech, Merck, Novartis, Astellas, Astra Zeneca, and Bristol-Myers Squibb outside the submitted work; stock options in Dual Therapeutics outside the submitted work.

421TiP

Phase 1b/2 study to assess the clinical effects of BGB-290 in combination with radiation therapy (RT) and/or temozolomide (TMZ) in patients with first-line or recurrent/refractory glioblastoma

P.Y. Wen1, K. Shih2, D. Schiff3, R. Brachmann4, R. Weitzman4, T. Cloughesy5 Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA, 2NeuroOncology, Sarah Cannon Research Insititute, Nashville, TN, USA, 3Neurology, University of Virginia Health System, Charlottesville, VA, USA, 4Clinical Development, BeiGene USA, Inc, Emeryville, CA, USA, 5Neurology-Oncology, University of California Los Angeles, Los Angeles, CA, USA

1

Background: Poly (ADP-ribose) polymerase (PARP) proteins are a family of DNA binding and repair proteins and are thought to play a key role in the base excision repair of DNA damage generated by TMZ. In glioblastoma (GB) cells, pharmacological modulation of PARP activity increased growth inhibition induced by TMZ in both p53-wild type and-mutant GB cells lowering the TMZ IC50. RT used in the clinical treatment of GB generates mostly single-strand breaks (SSBs). In non-replicating cells PARP inhibition only delays the repair of SSBs induced by radiation with a minimal impact on cell survival. On the contrary, PARP inhibition markedly enhances radiosensitivity of proliferating cells generating double-strand breaks. Thus, PARP inhibitors have the potential to increase the therapeutic index of RT by increasing DNA damage mainly in highly replicating tumor cells, but sparing non-cycling normal tissues. BGB290, a potent and selective inhibitor of PARP1/2, has demonstrated potent PARP trapping, brain penetrance and antitumor activity in preclinical intracranial xenograft models. Trial design: This open-label, dose-escalation/dose-expansion Phase 1b/2 study was designed to determine the safety, tolerability, pharmacokinetics, pharmacodynamics, and antitumor effects of BGB-290 at the recommended Phase 2 dose (60 mg PO BID) in combination with RT and/or TMZ. The Phase 1b component will consist of 3 doseescalation arms. Arm A: BGB-290 will be combined with RT in patients with first-line GB with unmethylated MGMT promoter (‘unmethylated GB’); Arm B: BGB-290 will be combined with both TMZ and RT in patients with first-line unmethylated GB; Arm C: BGB-290 will be combined with increasing doses of TMZ in patients with recurrent/ refractory methylated or unmethylated GB. Once a recommended Phase 2 regimen has been established, up to 60 patients may be enrolled in the dose-expansion (Phase 2) cohort for that arm. In Arm C, 2 expansion cohorts with up to 60 patients each may be opened: 1 for unmethylated GB and 1 for methylated GB. Legal entity responsible for the study: Beigene Ltd. Funding: Beigene Ltd. Disclosure: P. Wen: Grants, personal fees, and/or non-financial support from Agios, Angiochem, AstraZeneca, Genentech/Roche, GlaxoSmithKline, Immunocellular Therapeutics, Karyopharm, Merck, Novartis, and other biotech/pharma companies outside submitted work. D. Schiff: Grants from Cavion & Celldex, personal fees from VBI, Orbus, Monteris, Genentech-Roche, Heron Pharmaceuticals, Midatech, and Oxigene, outside the submitted work. R. Brachmann: Employee of Beigene USA, Inc. R. Weitzman: Consultant to BeiGene. T. Cloughesy: Personal fees from Pfizer, Tocagen, Roche, Novocure, Nektar, VBL, ABBVIE, Upshire Smith, Notable Labs, Oxigene, NewGen, Agios, Cortice, MedQia, PRoNai,and other pharma/biotech companies, outside the submitted work. All other authors have declared no conflicts of interest.

140 | Developmental therapeutics

Annals of Oncology 422TiP

PROCLAIM-CX-2009: A first-in-human trial to evaluate CX-2009 in adults with metastatic or locally advanced unresectable solid tumors

J. Garcia-Corbacho1, A. Spira2, V. Boni3, J. Feliu4, M. Middleton5, H. Burris6, A. Yang Weaver7, M. Will7, J. Harding8, F. Meric-Bernstam9, V. Heinemann10 1 Oncology, Hospital Clinic Barcelona, Barcelona, Spain, 2Oncology, Virginia Cancer Specialists Research Institute, and Oncology Research, Fairfax, VA, USA, 3Oncology, START Madrid-CIOCC, Madrid, Spain, 4Oncology, Hospital Universitario La Paz, Madrid, Spain, 5Oncology, Oxford NIHR Biomedical Research Centre, University of Oxford, Oxford, UK, 6Oncology, Sarah Cannon Research Institute, Nashville, TN, USA, 7Oncology, CytomX Therapeutics, South San Francisco, CA, USA, 8Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, USA, 9Oncology, University of Texas MD Anderson Cancer Center, Houston, TX, USA, 10Oncology, Ludwig-Maximilians-University of Munich - Campus Grosshadern, Munich, Germany Background: CX-2009 is a novel recombinant ProbodyTM drug conjugate (PDC) derived from a humanized monoclonal antibody (mAb) against CD166 and conjugated to N-succinimidyl 4-(2-pyridyldithio) butanoate-N20 -deacetyl-N20 -(4-mercapto-4methyl-1-oxopentyl)-maytansine (SPDB-DM4, licensed from Immunogen), a potent microtubule inhibitor. PDCs are fully recombinant mAb prodrugs designed to remain inactive until they are cleaved into an active mAb by tumor-associated proteases. This tumor-specific activation allows PDCs to target highly and homogeneously expressed tumor antigens while avoiding binding to these same targets on healthy tissue. An example is CD166 (also referred to as activated leukocyte cell adhesion molecule [ALCAM]), which is highly expressed in multiple cancers but also in healthy tissue. In preclinical studies, CX-2009 exhibited antitumor activity and reduced peripheral binding compared to the corresponding anti-CD166 ADC. Trial design: PROCLAIM-CX-2009 (PRObody CLinical Assessment In Man) is an open-label, multicenter, dose-escalation study to determine the maximum tolerated dose (MTD) and the recommended phase 2 dose (RP2D) of CX-2009 in 7 selected tumor types with high CD166 expression (breast, lung, prostate, ovarian, endometrial, head and neck, and biliary carcinomas). Part A (n 50) will initiate with accelerated dose titration, followed by a standard 3 þ 3 design to determine the MTD and ending in a modified toxicity probability interval 2-design cohort treated at the MTD to determine the RP2D. Part B of the study will be a dose expansion phase testing CX-2009 administered at the RP2D in the same 7 tumor types (up to 14 patients each, n 98). Eligibility is based on confirmed refractory metastatic or locally advanced unresectable tumor. Outcome measures include assessment of safety, tolerability, pharmacokinetics, and efficacy based on RECIST 1.1. Exploratory biomarkers will characterize tumor CD166 expression and mitotic markers as well as CX-2009 activation in tumor versus peripheral blood. Clinical trial identification: NCT03149549 Legal entity responsible for the study: CytomX Therapeutics, South San Francisco, CA, USA Funding: CytomX Therapeutics, South San Francisco, CA, USA Disclosure: M. Middleton: Grants: Roche, AstraZeneca, GSK. Advisory board: Amgen, Novartis, Rigontec, CytomX. Personal fees: Amgen, Roche, GSK, Novartis, BristolMyers Squibb, Eisai, Merck, CytomX. A. Yang Weaver, M. Will: Employee of CytomX Therapeutics. J. Harding: Consultant to Bristol-Myers Squibb. All other authors have declared no conflicts of interest.

423TiP

The first-in-human, dose-finding PROCLAIM-CX-072 trial to assess the antitumor activity and tolerability of the probody therapeutic CX-072 as monotherapy and in combination with ipilimumab or vemurafenib in solid advanced tumors and lymphomas

V. Boni1, J. Garcia-Corbacho2, J. Feliu3, J. Wydmanski4, Z. Horvath5, I. Bondarenko6, B. Irving7, M. Will7, E.G.E. de Vries8, F. Thistlethwaite9 1 Oncology, START Madrid-CIOCC, Madrid, Spain, 2Oncology, Hospital Clinic Barcelona, Barcelona, Spain, 3Oncology, Hospital Universitario La Paz, Madrid, Spain, 4Oncology, NZOZ Vegamed, Katowice, Poland, 5Oncology, Debreceni Egyetem Klinikai Kozpont, Debrecen, Hungary, 6Oncology, Dnipropetrovsk State Medical Academy, Dnepropetrovsk, Ukraine, 7Oncology, CytomX Therapeutics, South San Francisco, CA, USA, 8Oncology, Universitair Medical Center Groningen, Groningen, Netherlands, 9 Oncology, The Christie NHS Foundation Trust and University of Manchester, Manchester, UK Background: CX-072 is a novel protease activatable prodrug (ProbodyTM therapeutic) derived from a human monoclonal antibody against programmed cell death ligand 1 (PD-L1). CX-072 was designed to restrict its activity to the tumor microenvironment and remain largely inactive in nonmalignant tissue. A surrogate molecule for CX-072 displayed potent antitumor activity with reduced systemic immune activation and immune-related toxicities in preclinical tumor models, potentially enabling new and more effective combination therapies. The phase 1/2 PROCLAIM-CX-072 (PRObody CLinical Assessment In Man) study will assess the safety, tolerability, and antitumor activity of CX-072, as monotherapy and in combination, in adults with advanced or recurrent solid tumors and lymphomas. Trial design: In an open-label, multicenter, dose-escalation, 3 þ 3 design, CX-072 will be administered as monotherapy (Part A) in 2 combination schedules with ipilimumab 3 mg/kg 3 weekly 4 (Parts B1 and B2) and in combination with vemurafenib 960 mg/ kg twice daily (Part C). The expansion cohort (Part D) will include CX-072

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology monotherapy in PD-L1–responsive tumor types. Patient recruitment was initiated on January 11, 2017. Key inclusion criteria are: Parts A and B1—advanced, refractory solid tumor or lymphoma in checkpoint inhibitor-naive patients for whom approved PD agents are not available; Part B2—advanced, refractory solid tumors or lymphomas with measurable disease that progressed on previous treatment with a PD-1/PD-L1 inhibitor but patients did not discontinue due to toxicity; Part C—checkpoint inhibitor, BRAF-inhibitor, and MEK-inhibitor-naive metastatic V600E BRAF-mutated melanoma. Efficacy will be determined according to irRECIST v1.1 criteria, and safety and tolerability will be assessed based on the incidence and nature of dose-limiting toxicities, adverse events (AEs), and serious AEs. Exploratory biomarkers will be used to characterize tumor protease activity, immune response pattern within the tumor, and CX-072 activation in tumor vs peripheral blood. Clinical trial identification: NCT03013491 Legal entity responsible for the study: CytomX Therapeutics, South San Francisco, CA, USA Funding: CytomX Therapeutics, South San Francisco, CA, USA Disclosure: J. Wydmanski: Consultant to Bristol-Myers Squibb. B. Irving, M. Will: Employee of CytomX Therapeutics. F. Thistlethwaite: Personal fees from Novartis, Bristol-Myers Squibb, Pfizer, and Ipsen. All other authors have declared no conflicts of interest.

424TiP

Phase 1b multi-indication study of the antibody drug conjugate anetumab ravtansine in patients with mesothelin-expressing advanced or recurrent malignancies

A. A. Adjei1, A. Walter2, L. Cupit3, J. Siegel4, A. Holynskyj3, B.H. Childs3, C. Elbi3 Medical Oncology, Mayo Clinic, Rochester, MN, USA, 2Biomarker Strategy, Bayer AG, Berlin, Germany, 3Clinical Development, Bayer HealthCare Pharmaceuticals Inc., Whippany, NJ, USA, 4Clinical Statistics, Bayer HealthCare Pharmaceuticals Inc., Whippany, NJ, USA

1

Background: Mesothelin is expressed in a wide variety of tumours, including mesothelioma, ovarian, pancreatic, gastric/GEJ, NSCLC, triple-negative breast cancer, cholangiocarcinoma, and thymic carcinomas. Anetumab ravtansine (BAY 94-9343), is a novel fully human anti-mesothelin IgG1 antibody conjugated to the maytansinoid tubulin inhibitor DM4 and has shown encouraging anti-tumor activity in mesothelioma and ovarian cancer patients in a phase I study. We will therefore conduct a signal generating study with anetumab ravtansine in six additional high unmet medical need malignancies with mesothelin expression (NCT03102320). Trial design: Eligibility criteria include: 18 years, unresectable locally advanced or metastatic recurrent or relapsing disease, one or more prior lines of therapy, and availability of tumour tissue for mesothelin expression testing. Mesothelin-positive patients with selected adenocarcinomas (NSCLC, triple negative breast, gastric including gastroesophageal junction) and thymic carcinoma will receive anetumab ravtansine as monotherapy at 6.5 mg/kg IV on a 21-day cycle. Patients with cholangiocarcinoma will receive anetumab ravtansine in combination with cisplatin (25 mg/m2 IV day 1 and 8 on a 21-day cycle for up to 6 cycles) and patients with pancreatic adenocarcinoma will receive anetumab ravtansine in combination with gemcitabine (1000 mg/m2 IV day 1 and 8 on a 21-day cycle). A safety run-in phase (18-24 patients each) will be conducted for the combination regimens prior to enrolling patients in the main study phase. The primary objective of the main phase of the study is objective response rate (ORR) of anetumab ravtansine as monotherapy or combination therapy in patients with either of two mesothelin expression levels: high (30% positive tumor cells with moderate and stronger membrane staining intensity) and low-mid (5% all intensities and 3 previous lines of therapy (23.8% pazopanib; 80.9% sunitinib; 47.6% everolimus) beside somatostatin analogs. 20 pts were evaluable for ORR with a median follow up of 10m (4.23-12.43). No responses (0%) were observed; 11 (55%) pts had stable disease and 6 of them lasted more than 6m; 7 (35%) pts had progression as best response. 1 pt had tumor shrinkage of 8%. Median PFS was 1.9m (IC95% 0 - 13). Median OS was 16.6m (IC95% 9.3 – 23.9). Most frequent toxicities of any grade were: asthenia (16.2%), diarrhea (6.4%), abdominal pain (4.7%), nausea (4.3%) and neutropenia (11.5%). 5 pts developed G3-4 neutropenia (1 case of febrile neutropenia) and 2 pts G3-4 thrombocitopenia. Conclusions: Lack of activity was observed with palbociclib in 21 molecularly unselected and heavily pretreated patients with advanced G1/2 pNETs. Translational studies correlating activity with molecular tumor markers and Ki67 proliferation index are ongoing. Clinical trial identification: EudraCT: 2014-003924-34 Legal entity responsible for the study: GETNE (Spanish Taskforce Group for Neuroendorine Tumors) Funding: Pfizer Disclosure: E. Grande Pulido: Advisor for Ipsen, Pfizer, Lexicon. Scientific lectures for Pfizer and Ipsen. All other authors have declared no conflicts of interest.

430O

Different RNA expression profile defines prognosis in grade 1/2 neuroendocrine neoplasms of small intestine origin: The GETNENETSEQ study

J. Capdevila1, F. Mancuso2, S. Landolfi3, F. Salva4, C. Miguel5, P. Jime´nez-Fonseca6, opez L opez8, A. Casteras9, T. Sauri Nadal1, P. Nuciforo10, R. Garcia-Carbonero7, C. L A. Vivancos2, I. Matos1 1 Medical Oncology, Vall d’Hebron University Hospital, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 2Cancer Genomics Group, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 3Anatomic Pathology Department, Vall d’Hebron Hospital, o Hospital, Barcelona, Spain, 5Pathology, Barcelona, Spain, 4Medical Oncology, Matar H. U. Central de Asturias, Oviedo, Spain, 6Medical Oncology, H. U. Central de Asturias, Oviedo, Spain, 7Medical Oncology, University Hospital 12 de Octubre, Madrid, Spain, 8 Oncology, Hospital Universitario Marque´s de Valdecilla, Santander, Spain, 9 Endocrinology, Vall d’Hebron University Hospital, Barcelona, Spain, 10Molecular Oncology Group, Vall d’Hebron Institute of Oncology, Barcelona, Spain Background: Neuroendocrine neoplasms (NENs) are a heterogeneous group of tumors with different prognosis. Clinical and pathological factors are used to predict outcome but there are some patients that have a poor outcome even with good classical prognostic factors. We aimed to define a different RNA expression profile (EP) that could detect patients with worse prognosis and identify possible targetable new pathways. Methods: We identified 48 paraffin-embedded archival tumor material of patients with metastatic grade 1/2 NENs of small intestine origin for RNAseq. We generated on average 66 million paired-end reads for each sample on HiSeq2500 (Illumina). RNAseq reads were mapped against the human reference genome (hg19) with Tophat (v2.0.14) and quantified using Cufflinks tools suite (v.2.2.1). 41 samples had sufficient quality to be included in the analysis. We used multivariate Cox proportional models to study the association between EP, clinical variables (gender, age, location of metastases, hormone production and Ki67 index) and overall survival (OS). We defined as poor outcome those patients that died within the first 3 years of the diagnosis of advanced disease. Results: 9348 transcripts were quantified. A gene signature of 329 transcripts was defined by a two-way statistical analysis between poor and long term survivors. A pathway enrichment analysis of these genes showed a deregulation in the poor prognosis group on the PI3KCA-Akt-mTOR and the Toll-like receptor signaling pathways. The hazard ratio (HR) for mOS defined by EP comparing poor and long term survivor groups was 0.33, 95% CI 0.1-1.1, p ¼ 0.081 in the univariate analysis and HR of 0.05, 95% CI 0.005-0.51, p ¼ 0.011 in the multivariate analysis. Conclusions: We identified statistically different RNA-clusters and different deregulated pathways for those patients with advanced NENs of small intestine origin with poor prognosis. To our knowledge, this is the first time that Toll-like pathway is involved in the pathogenesis of NENs. These results are relevant as they may help improve the prognosis stratification of patients and involve novel targetable pathways of great clinical potential. Legal entity responsible for the study: Vall d’Hebron Institute of Oncology Funding: Spanish Task Force for Neuroendocrine and Endocrine Tumors (GETNE) Disclosure: All authors have declared no conflicts of interest.

431O

Genomic subtypes of pulmonary large cell neuroendocrine carcinoma (LCNEC) may predict chemotherapy outcome

J. Derks1, N. Leblay2, R.J. van Suylen3, E. Thunnissen4, M.A. den Bakker5, H.J.M. Groen6, E. Smit7, R.A. Damhuis8, E. van den Broek9, A. Charbrier2, M. Foll2, J. McKay2, L. Fernandez-Cuesta2, E-J. Speel10, A-M. Dingemans1 1 GROW School for Oncology and Developmental Biology, Department of Pulmonology, Maastricht University Medical Center (MUMC), Maastricht, Netherlands, 2Genetic Cancer Susceptibility Group, International Agency for Research on Cancer (IARC-WHO), Lyon, France, 3Pathology-DNA, Jeroen Bosch Hospital, s’Hertogenbosch, Netherlands, 4 Department of Pathology, Vrije University Medical Centre (VUMC), Amsterdam, Netherlands, 5Department of Pathology, Maasstad Hospital & Erasmus University Medical Center, Rotterdam, Netherlands, 6Department of Pulmonary Diseases, University of Groningen and University Medical Center Groningen, Groningen, Netherlands, 7Department of Pulmonary Diseases & Thoracic Oncology, VU Medical Center & Netherlands Cancer Institute, Amsterdam, Netherlands, 8Department Research, Comprehensive Cancer Association, Utrecht, Netherlands, 9PALGA Foundation, Utrecht, Netherlands, 10GROW School for Oncology and Developmental Biology, Department of Pathology, Maastricht University Medical Center (MUMC), Maastricht, Netherlands Background: Whether to treat LCNEC with non-small cell lung carcinoma type chemotherapy (NSCLC, i.e. platinum-gemcitabine/taxanes or pemetrexed) or small cell lung carcinoma type (SCLC, i.e. platinum-etoposide) is subject of debate. Recent molecular studies have identified two mutually exclusive genomic LCNEC subtypes, the co-mutated TP53 and RB1 (i.e. SCLC like) and the STK11/KEAP1 (predominantly RB1 wild-type, i.e. NSCLC like) subtype. We determined if genomic LCNEC subtypes are clinically relevant for chemotherapy (CT) outcome. Methods: Clinical data and tumour specimens were retrospectively obtained from the Netherlands Cancer Registry and Pathology Registry (PALGA, 2003-2012). All firstline CT treated patients with panel-consensus diagnosed LCNEC were included for next-generation sequencing (NGS) analysis for TP53, RB1, STK11, and KEAP1 genes. Furthermore, immunohistochemistry for RB1 (pRB1, 13A10) was analysed (H-score, 50 considered as positive). NGS and pRB1 results were correlated with overall survival (OS) and progression free survival (PFS) by Kaplan Meier plots and Log-rank test. Results: LCNEC was panel-consensus diagnosed in 148/232 patients; 79 passed quality control for NGS and 109 for pRB1. RB1 mutations were found in 47% (n ¼ 37) and loss of pRB1 expression in 72% (n ¼ 78) of the cases. Mutations in RB1 were mutually exclusive with mutations in STK11 (n ¼ 8; P ¼ 0.006). Due to reported resistance in neuroendocrine carcinomas, we analysed NSCLC-CT without pemetrexed-CT; OS was significantly longer for NSCLC-CT (n ¼ 15, 9.6 [7.7-11.6] months) compared to SCLCCT (n ¼ 13, 5.8 [5.5-6.1] months, P ¼ 0.026). LCNEC tumours expressing pRB1 also had longer OS when treated with NSCLC-CT (n ¼ 14, 9.6 [7.4-11.8] vs. n ¼ 9, 1.9 [1.7-2.1] months, P ¼ 0.001). PFS of RB1 wild-type NSCLC-CT treated patients was significantly longer than SCLC-CT (P ¼ 0.018) also for pRB1 (P ¼ 0.023). In patients with a RB1 mutation OS and PFS were not significantly different for NSCLC-CT vs. SCLC-CT. Conclusions: In LCNEC with RB1 wild-type, NSCLC-CT correlates with a more favourable outcome compared to SCLC-CT. However, RB1 mutated LCNEC treated with NSCLC-CT have similar clinical outcomes as compared to SCLC-CT. Prospective studies should be initiated. Legal entity responsible for the study: Maastricht University Medical Center Funding: This study was supported by grants from the Dutch Cancer Society (UM2014-7110) granted to Prof. Dr. A-M. C. Dingemans. Mr. J.L. Derks is the recipient of an ERS/EMBO Joint Research Fellowship – Nr. STRTF 2016 – 7178. The research leading to these results has received funding from the European Respiratory Society (ERS) and European Molecular Biology Organization (EMBO). Disclosure: H.J.M. Groen: Other from Lilly, GSK, Merck, Pfizer, Roche, BMS, outside the submitted work. E-J. Speel: Other from Pfizer, Roche, Amgen, outside the submitted work. A-M. Dingemans: Personal fees from Roche, BMS, Boehringer Ingelheim, AstraZeneca, Eli Lilly, MSD, Pfizer and Amgen, outside the submitted work. All other authors have declared no conflicts of interest.

432PD

Prognostic impact of RNA expression profile (EP) in the phase III DECISION trial for patients with advanced radioactive-iodine refractory differentiated thyroid cancer (DTC)

I. Matos1, F. Mancuso2, C. Iglesias3, P. Nuciforo4, C. Zaf on5, H.G. Palmer6, Z. Ogbah2, na8, M.S. Brose9, M. Schlumberger10, L. Mu~ nos2, G. Villacampa Javierre7, C. Pe~ A. Vivancos2, J. Capdevila1 1 Oncology, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 2Cancer Genomics Group, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 3Pathology, Vall d’Hebron University Hospital, Barcelona, Spain, 4Molecular Oncology Group, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 5Endocrinology, Vall d’Hebron University Hospital, Barcelona, Spain, 6Stem Cells and Cancer Group, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 7Oncology Data Science (ODysSey) Group, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 8Biomarker Strategy, Bayer HealthCare Pharmaceuticals, Whippany, NJ, USA, 9Medical Oncology, University of Pennsylvania- CRB, Philadelphia, PA, USA, 10Oncology, Institut de Cance´rologie Gustave Roussy, Villejuif, France Background: Sorafenib has demonstrated a significant impact in progression free survival (PFS) in patients with advanced DTC in the DECISION trial. BRAF and RAS

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx368 | 143

abstracts

Annals of Oncology

Table: 432PD Clinical variables in each expression profiles

Treatment Sex ECOG Tumor histology

BRAF status RAS status

BRAF like (n 5 56)

RAS like (n 5 28)

non-BRAF/RAS-like (n 5 41)

37(66.1%) 19 (33.9%) 26 (46.4%) 30 (53.6%) 36 (64%) 20 (36%) 53 (94.6%) 2 (3.6%) 1 (1.8%) 31 (55.4%) 25 (44.6%) 5 (8.9%) 51 (91.1%)

9 (32%) 19 (68%) 19 (68%) 9 (32%) 15 (53.6%) 13 (46.4%) 14 (50%) 7 (25%) 7 (25%) 2 (7%) 26 (93%) 14 (50%) 14 (50%)

22 (53.7%) 19 (46.3%) 15 (36.6%) 26 (63.4%) 29 (70.7%) 12 (29.3%) 6 (14.6%) 24(58.5% 10 (24.4%) 0 (0%) 41 (100%) 1 (2.4%) 40 (97.6%)

Sorafenib Placebo Female Male 0 1-2 Papillary Follicular Poorly differentiated Mutated Wild type Mutated Wild type

mutation status (MS) have not shown prognostic significance on PFS and overall survival (OS) in multivariate model. The aim of this study was to correlate different RNA expression profiles with PFS and OS in a multivariate model. Methods: We previously identified 3 expression profiles, BRAF, RAS and non-BRAF/ RAS-like based on RNA-seq analysis (77 million paired-end reads for each sample on HiSeq2000) of 125 tumour samples of patients included in the DECISION trial. RNAseq reads were mapped against the human reference genome (GRCh38) with STAR (v2.5.1b) using ENCODE parameter. We used multivariate Cox proportional models to study the association between clinical variables (sex, age, thyroid cancer histology, Eastern Cooperative Oncology Group performance status), arm of treatment and biomarkers (EP and MS) with PFS and OS. Results: The clinical variables in each expression profiles are shown in Table. Multivariable analysis indicated that only sorafenib treatment (HR: 0.39, 95% CI 0.230.66, p < 0.001), age (HR: 0.97, 95% IC 0.94-0.99, p ¼ 0,002) and BRAF-like EP (HR ¼ 0.41, 95% IC 0.17-0.99, p ¼ 0.046) were independent prognostic factors for PFS. No significant prognostic factors we identified for OS. However, in papillary histology (PTC), only the BRAF-like EP was associated with outcome (HR ¼ 0.32, 95% IC 0.1180.876, p ¼ 0.026). Conclusions: RNA-seq analysis identifies 3 different expression profiles in DTC: BRAF-like, RAS-like and non-BRAF/RAS-like. BRAF-like EP includes almost all BRAF mutant tumors but also a 45% of tumors with no mutation in BRAF gene. In the multivariate analysis, BRAF-like EP has shown a better prognostic factor for PFS in DTC and for OS in PTC. Legal entity responsible for the study: Vall d’Hebron Institute of Oncology Funding: Bayer HealthCare Pharmaceuticals, Inc. Disclosure: C. Pe~ na: Bayer employee. All other authors have declared no conflicts of interest.

433PD

Comprehensive genomic profiling of metastatic and relapsed thyroid gland carcinomas is associated with tumor type and reveals new routes to targeted therapies

J.S. Ross1, L.M. Gay2, P. Vanden Borre2, N. Almog2, A.B. Schrock3, J-A. Vergilio4, J. Suh4, S. Ramkissoon4, E. Severson4, S. Daniel4, S.M. Ali4, V.A. Miller4, P.J. Stephens4, J.A. Elvin4, D. Bowles5 1 Pathology, Albany Medical Center, Albany, NY, USA, 2Pathology, Foundation Medicine, Cambridge, MA, USA, 3Clinical Development, Foundation Medicine, Cambridge, MA, USA, 4R&D, Foundation Medicine, Cambridge, MA, USA, 5Medical Oncology, University of Colorado Denver, South Aurora, CO, USA Background: We queried whether CGP could differentiate the mTC subtypes of papillary (PTC), follicular (FTC), medullary (MTC) and anaplastic (ATC) carcinomas and their potential responses to targeted and immunotherapies. Methods: CGP was performed on FFPE samples using hybridization-captured, adaptor ligation based libraries to a mean coverage depth of > 500X for up to 315 cancerrelated genes. Total mutational burden (TMB) was determined on 1.1 megabases of sequenced DNA as previously described (PMID: 28420421). Results: 778 clinically advanced mTC were studied. Median age across the subtypes was similar with male patients slightly more frequent than females except for PTC. ATC had the highest median GA/sample at 4.34. BRAF was a target option for 73% of PTC and 39% of ATC, and RET for MTC in 81% of cases. Oncogenic rearrangements of RET, BRAF, ALK, and NTRK1 were detected in 8%, 3%, 1%, and 1% of PTC cases, respectively, and in 1%, 1%, 0%, and 1% of ATC cases, respectively, but not detected in any cases of FTC or MTC. At 66%, TP53 GA were most frequent in ATC. TERT GA were 58-67% in PTC, FTC and ATC and 0% in MTC. TMB was extremely low for all

Table: 433PD PTC

FTC

MTC

ATC

No. Patients Median age (years) Gender (F/M) GA/tumor Signiﬁcant genes altered

408 59 215/193 2.75 BRAF TERT TP53 RET

TP53 GA Frequency RET GA Frequency hTERT Frequency BRAF GA Frequency BRAF, RET, ALK, or NTRK rearrangemens Total Mutational Burden 10 mut/Mb Opportunity for Targeted Therapies

11% 9% 58% 73% 13% 2% High (BRAF, oncogenic fusions)

77 60 37/40 3.01 TERT NRAS TP53 PTEN HRAS 16% 0% 67% 7% 0% 1% Low

113 54 46/67 1.96 RET VHL MEN1 HRAS CCND1 2% 81% 1% 0% 0% 2% High (RET)

180 64 88/92 4.34 TP53 TERT BRAF NRAS PIK3CA NF1 NF2 PTEN 66% 2% 61% 39% 3% 3% Moderate (BRAF, oncogenic fusions)

144 | Endocrine and neuroendocrine tumours

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 4 mTC tumor types with only 3 mTC (50% of parenchyma) of liver metastases [OR 13.86 (2.57 – 74.68), p ¼ 0.002]. Time from symptoms to diagnosis of NET was borderline significant (p ¼ 0.08). When CHD was defined as at least mild valve regurgitation, the frequency of CHD was 45% (19 patients) and we observed a significant association between the presence of cardiovascular comorbidities and CHD [OR 6.58 (1.09; 39.78), p ¼ 0.040]. Conclusions: CHD is highly frequent among pts with CS. We found that high liver tumor burden and possibly, longer time of symptoms until diagnosis of NET were associated with CHD. Such findings probably imply that a delayed diagnosis negatively affects CS patients, increasing the risk of CHD. Interestingly, we found that concurrent cardiovascular disease was associated with CHD, as a potential predisposing factor. Legal entity responsible for the study: Instituto do C^ancer do Estado de S~ao Paulo Funding: None Disclosure: All authors have declared no conflicts of interest.

448P 446P

Benefit of oral monotherapy with pazopanib in metastatic gastroenteropancreatic neuroendocrine tumours

D. Katalinic, I. Aleric, A. Vcev Department of Internal Medicine, Faculty of Medicine Osijek, Osijek, Croatia Background: Although standard therapy for gastroenteropancreatic neuroendocrine tumors (GEP-NETs) can provide symptom relief and delay tumour progression, new strategies are needed for patients with metastatic disease. The aim of this study was to investigate the antitumour activity and safety profile of pazopanib - a selective multitargeted receptor tyrosine kinase inhibitor of vascular endothelial growth factor receptor and platelet-derived growth factor receptor. Methods: We enrolled 124 patients with metastatic GEP-NETs. Pazopanib was administered orally at a dose of 800 mg daily with a 28-day cycle. The primary endpoint was overall response rate (ORR) according to Response Evaluation Criteria in Solid Tumors. The secondary endpoints were overall survival (OS), progression-free survival (PFS) at 6 months and safety profile of pazopanib (general tolerability and toxicity). The third endpoint was to compare the clinicopathological features of tumours and biomarker analysis with survival of the patients. Results: The mean follow-up time was 196687 days with a range of 67-268 days; 26 patients died within the observation time. 69% of the patients had confirmed pancreatic GEP-NET and 51% had colorectal, gastric and duodenal GEP-NET. 59 (47.6%) patients had G1, 34 (27.4%) G2 and 31 (25%) had G3 GEP-NET. ORR was 24% (19 of 124 patients), stable disease was achieved in 49 patients (39.5%) and PFS at 6 months was 36%. Median OS was 10.2 months (95% CI, 5.4-13.2 months). The most common grade 3-4 adverse events attributed to therapy were neutropenia (11%), proteinuria (14%), diarrhea (7%), and fatigue (12%). Patients with high CgA levels had the highest mortality risk (hazard ratio 3.478, 94% confidence interval 1.313-4.727, p < 0.001) and worse outcomes. Furthermore, extremely high CgA levels (>3000 ng/mL, range 8300800 ng/mL) were associated with low survival independently from the Ki-67 score in a multivariate Cox regression model. Conclusions: Pazopanib demonstrated a comparable therapeutic efficiency as well as a satisfying safety profile compared to other targeted agents in the treatment of patients with metastatic GEP-NETs. Legal entity responsible for the study: Faculty of Medicine Osijek Funding: Faculty of Medicine Osijek Disclosure: All authors have declared no conflicts of interest.

447P

High hepatic tumor burden and cardiovascular comorbidities linked to carcinoid heart disease

M.C. Mesquita1, C.A.C. Silva1, C.M.P.D.C. Silva2, M.C. Feres Almeida Soeiro2, L.A. Hajjar2, P.M. Gehm Hoff1, R.S.P. Riechelmann1 1 Oncology, Instituto do C^ ancer do Estado de S~ ao Paulo, Sao Paulo, Brazil, 2Cardiology, Instituto do C^ ancer do Estado de S~ ao Paulo, Sao Paulo, Brazil Background: One of the most common functioning syndromes associated with neuroendocrine tumors (NET) is the carcinoid syndrome (CS). By releasing vasoactive substances, these tumors can cause fibrotic complications, including right-sided valve heart fibrosis, named carcinoid heart disease (CHD). Factors associated with the onset and progression of CHD are poorly understood. We aimed to investigate prognostic factors associated with CHD. Methods: Retrospective study of consecutive patients (pts) with advanced NET and CS and/or elevated 24h-urinary 5HIAA who performed an echocardiogram to screen for CHD. CHD was defined as echocardiographic evidence of moderate to severe tricuspid or pulmonar regurgitation. Results: From 2009 to 2017 42 pts were included: Median age was 54.4 (19 – 85) years, 24 were female, 69% had midgut NET. The frequency of CHD was 38% (16 pts) CHD was not associated with age (p ¼ 0.79), sex (p ¼ 0.38), bone metastasis (p ¼ 0.66),

Volume 28 | Supplement 5 | September 2017

Bone metastases in patients with neuroendocrine neoplasms: A survey of natural history and clinical management

N. Fazio1, P. Maisonneuve2, A.M. Frezza3, T. Ibrahim4, A. La Salvia5, S. Tafuto6, S. Pusceddu7, R. Marconcini8, F. Silvestris9, D. Campana10, D. Santini11, A. Faggiano12, S. Massironi13, L. De Marinis14, G. Rubini15, E. Merola16, L. Antonuzzo17, V. Amoroso18, I. Puliafito19, F. Spada1 1 Gastrointestinal Medical Oncology and Neuroendocrine Tumors, European Institute of Oncology, Milan, Italy, 2Epidemiology, European Institute of Oncology, Milan, Italy, 3 Adult Mesenchymal Tumour and Rare Cancer Medical Oncology, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy, 4Osteoncology and Rare Tumors Center, IRCCS Istituto Scientiﬁco Romagnolo per lo Studio e la Cura dei Tumori (IRST), Meldola, Italy, 5Medical Oncology, Ospedale S. Luigi, Orbassano, Italy, 6Medical Oncology, Istituto Nazionale Tumori – I.R.C.C.S - Fondazione Pascale, Naples, Italy, 7Medical Oncology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 8Oncology, Azienda Ospedaliera Universitaria S.Chiara, Pisa, Italy, 9Biomedical Sciences and Human Oncology, Universit a di Bari, Bari, Italy, 10Medical Oncology, Policlinico S. Orsola, Bologna, Italy, 11Oncology, Campus Bio-Medico di Roma, Rome, Italy, 12 Endocrinology, Policlinico di Napoli, Naples, Italy, 13Gastroenterology, Policlinico di Milano, Milan, Italy, 14Endocrinology, Policlinico Gemelli, Rome, Italy, 15Nuclear Medicine, Universit a di Bari, Bari, Italy, 16Gastroenterology, Ospedale S. Andrea, Rome, Italy, 17Medical Oncology, Azienda Ospedaliera Careggi, Florence, Italy, 18Medical Oncology, Spedali Civili, Brescia, Italy, 19Medical Oncology, Istituto Oncologico del Mediterraneo, Viagrande, Italy Background: Bone metastases (BM) in neuroendocrine neoplasms (NEN) represent a poorly defined issue. Methods: This is a nationwide survey among Italian institutions dealing with NEN patients. Characteristics of BM, clinical management, skeletal related events (SREs) and disease outcome were recorded. Results: We analysed 321 patients with histological diagnosis of NEN and BM collected from 18 Italian Centers. Mean age was 59 y.o. (range 13-86). Primary sites were 47% gastroenteropancreatic (GEP), 36% lung, 4% Paraganglioma/Pheocromocytoma (Par/ Pheo), 7% unknown, 5% others. The vast majority (72%) of NEN were already metastatic at diagnosis and the liver represented the second most frequent site of metastasis (in 77% of patients) during follow-up, in addition to BM. Bone was the first metastatic site in 41% of cases. Neoplasms were low/intermediate grade in 80% and high grade in 20%. SREs occurred in 32% of cases, mainly in lung and others. Median time to SRE was 4 months. It strictly correlated with the high grade, irrespective of the primary site. Bisphosphonates were administered in 32% of patients. Median survival from BM diagnosis was 65 months (range 45-78) in the whole population, with Par/Pheo at the best and high grade GEP at the worst limit. SRE, high grade (or in alternative high Ki67) and prior lung metastases resulted significantly associated with worse overall survival at the multivariable analysis. After adjustment for tumor grade, survival of patients with GEP and lung NENs were similar. Conclusions: This is one of the largest series of NEN patients with BM reported so far. This survey mirrors the Italian real clinical practice in this setting, as it included most Centers involved in NET patients’ management. It showed that overall, BM from NEN are associated with a relatively long survival. Bisphosphonates were used in a low percentage of cases, probably related to SRE. Tumor grade confirmed its value in separating two survival categories, irrespective of primary site. The results of this analysis generated hypotheses for prospective trials in homogeneous clinical settings. Legal entity responsible for the study: Nicola Fazio Funding: None Disclosure: All authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx368 | 149

abstracts 449P

Financial toxicity in patients with neuroendocrine tumors: Impact of a chronic disease on patients’ economic situation

L. Apostolidis1, K. Mehlis1, M. Kudlich1, J. Witte2, J. Walther1, W. Greiner2, E.C. Winkler1 Department of Medical Oncology, National Center for Tumor Diseases, University Hospital Heidelberg, Heidelberg, Germany, 2Department of Health Economics and Health Care Management, School of Public Health, Bielefeld University, Bielefeld, Germany

1

Background: The diagnosis of cancer imposes physical, emotional and financial burdens on patients. So far, the socio-economic impact of cancer for patients in Germany is poorly understood. The aim of the project is to provide an overview on patients’ financial losses due to a neuroendocrine tumor (NET) diagnosis as well as possible psychosocial effects. Methods: This prospective quantitative study recruited n ¼ 123 patients with NET from November 2016 to March 2017 at the National Center for Tumor Diseases, University Hospital Heidelberg. They completed a survey on patients’ income, cancerrelated out-of-pocket costs, disease burden (Distress Thermometer), quality of life (EORTC-LQ 29/30), health status (EQ-5D) and demographic data. Results: 78.0% (n ¼ 96) of the patients stated to have higher out-of-pocket costs because of their disease, mostly in terms of travel expenses and co-payments for medication. With regard to loss of income, 29.3% (n ¼ 36) of the participants reported a minus, which is beyond 800eper month in almost half (44.4%) of these cases. 61.5% of the persons affected by income losses cannot compensate these by savings or credits. 33.3% (n ¼ 41) of the responding patients indicated that they have to cut back on their expenses of daily living as a result of their disease. Higher cancer-related out-of-pocket costs per month were associated with lower estimation of patient’s quality of life (p ¼ 0.003), lower self-reported health status (p ¼ 0.013) and a more severe perception of disease burden (p ¼ 0.036) whereas a higher monthly income was strongly correlated with better quality of life (p¼.008)/health status (p ¼ 0.008) and lower disease burden (p ¼ 0.006). Conclusions: Given the fact that the majority of surveyed patients has to face financial losses due to their cancer diagnosis which is accompanied by the experience of distress as well as worsened quality of life and health status, there is a need for targeted measures that could prevent financial problems and reduce emotional burdens. Further research is required to address this aim. Clinical trial identification: The trial was approved by the Instituional Research Ethics Committee (approval S-458/2016). Legal entity responsible for the study: National Center for Tumor Diseases, University Hospital Heidelberg Funding: Ipsen Pharma GmbH Disclosure: All authors have declared no conflicts of interest.

450P

Pancreatic exocrine insufficiency (PEI) in patients (pts) with welldifferentiated neuroendocrine tumours (wd-NETs) treated with somatostatin analogues (SSAs): Incidence and impact on quality of life

A. Lamarca1, L. McCallum1, C. Nuttall1, J. Barriuso2, M. Frizziero1, R. Leon1, W. Mansoor1, M.G. McNamara2, R.A. Hubner1, J.W. Valle2 1 Medical Oncology, The Christie NHS Foundation Trust, Manchester, UK, 2Medical Oncology, The University of Manchester/The Christie, Manchester, UK Background: Advanced wd-NET patients (pts) are commonly treated with SSAs. PEI may be under-estimated in trials due to difficulties in distinguishing carcinoid syndrome-related diarrhoea and PEI. Methods: In this single-institution, prospective, observational study, sequential pts with advanced wd-NET were commenced on SSAs and followed for a minimum of 12 months (or until disease progression). Toxicity was prospectively assessed monthly. Faecal elastase testing (FE) (for diagnosis of PEI) and quality of life (QoL) questionnaires (QLQ-C30 and QLQ-GI.NET21) were performed 3-monthly. Results: Of 52 pts recruited (Jan 15-Apr 16), 50 were eligible: median age 65.8 yrs; 58% male; ECOG performance status 0 (42%), 1 (46%) or 2 (12%); primary: small bowel (60%), pancreas (22%), lung (12%) and other (6%). Baseline median Ki-67 was 3.1% (range 0.7-25), serum 5HIAA: 195 nmol/L (95%CI 145-318) and chromogranin A (CgA): 327 ng/mL (95%CI 140-582). Most pts were metastatic (92%), non-functional (66%) and started SSA first-line (88%); depot SSA was octreotide in 60%, lanreotide in 40%. Forty-one pts (82%) started full-dose SSA (4-weekly octreotide 30mg or lanreotide 120mg); 96% achieved full dose; 3 pts required dose reduction due to toxicities. Grade (G) 1-2 toxicities were flatulence (50%), abdominal pain (32%), diarrhoea (30%), fatigue (20%), PEI (22%), nausea (16%), hyperglycaemia (6%), anorexia (4%) and constipation (2%). G 3-4 toxicities were few (G3 hyperglycaemia (n ¼ 1) and G3 PEI (n ¼ 1); no G4). Twelve pts (24%) developed SSA-related PEI (4 clinical diagnosis, 8 FE-confirmed) at a median of 2.9 mo (95%CI 1.7-8.6) after starting SSA; 11/12 (92%) pts received enzyme replacement. Questionnaires identified fatigue, insomnia and diarrhoea as the most important baseline symptoms; SSA therapy did not negatively-affect QoL. Estimated median progression-free survival (PFS) was 29.9 mo (95%CI 21.4-not reached). High baseline CgA was an independent factor for shorter PFS (HR 1.01 (95%CI 1.001-1.1); p-value 0.001) after adjustment for other factors (baseline 5HIAA, Ki-67).

150 | Endocrine and neuroendocrine tumours

Annals of Oncology Conclusions: SSA-induced PEI occurs in 1:4 pts; clinicians should actively identify and treat. Legal entity responsible for the study: N/A Funding: Dr. Lamarca was part-funded by the European Society for Medical Oncology Translational Fellowship Programme and by the Spanish Society of Medical Oncology (SEOM) Fellowship Programme. Dr Barriuso was funded by the Spanish Society of Medical Oncology (SEOM) Fellowship Programme. Disclosure: All authors have declared no conflicts of interest.

451P

Final analysis of time to subsequent disease progression/death in patients with metastatic enteropancreatic neuroendocrine tumours progressing under placebo and switched to lanreotide autogel/depot 120mg in the CLARINET open-label extension

J.B. Cwikla1, E.M. Wolin2, M. Pavel3, A.T. Phan4, M. Raderer5, E. Sedlackova6, G. Cadiot7, J. Capdevila8, G. Rindi9, C. Lombard-Bohas10, N. Liyanage11, X-M. Truong Thanh12, P. Ruszniewski13, M. Caplin14 1 University of Warmia and Mazury, Olsztyn, Poland, 2Monteﬁore Einstein Center for atsklinikum Erlangen, Erlangen, Germany, Cancer Care, New York, NY, USA, 3Universit€ 4 Medical Oncology, University of New Mexico Cancer Center, Albuquerque, NM, USA, 5 Medicine I, Clinical Division of Oncology, University Hospital, Vienna, Austria, 6 Oncology of the First Faculty, Medicine and General Teaching Hospital, Prague, Czech Republic, 7Gastroenterology and Digestive Oncology, Robert-Debre´ Hospital, Reims, France, 8Gastrointestinal and Endocrine Tumor Unit, Vall d’Hebron University Hospital, a Cattolica del Barcelona, Spain, 9Diagnostica e Medicina di Laboratorio, Universit Sacro Cuore, Rome, Italy, 10Edouard-Herriot Hospital, Hospices Civils de Lyon, Lyon, 11 12 France, R&D Statistics, Ipsen Innovation, Les Ulis, France, Ipsen, BoulogneBillancourt, France, 13Beaujon Hospital, Clichy, France, 14Neuroendocrine Tumour Unit, Royal Free Hospital, London, UK Background: The CLARINET core study established the antitumour activity of lanreotide Autogel 120mg/28 days (LAN) in metastatic enteropancreatic neuroendocrine tumours (NETs). The vast majority of the core study population (96%) had stable disease (SD) at baseline, but the LAN open-label extension (OLE) also included patients with progressive disease (PD; while receiving placebo [PBO] in the core study). Here, we report the final analysis of time to subsequent death/PD for patients with PD switched to LAN. Methods: In the core study, patients with metastatic well-/moderately differentiated non-functioning enteropancreatic NETs received LAN/PBO for 96 weeks or until death/PD (RECIST 1.0). Eligible patients for the OLE (NCT00842348) had SD at corestudy end or PD (with PBO only) during the core study. Adverse events (AEs) were recorded at 4-weekly visits. CT/MRI scans from OLE baseline (week 1) and every 24 weeks subsequently were assessed locally for PD (RECIST 1.0). Primary objective: longterm safety. Secondary objective: long-term efficacy, with assessments including PFS and time to subsequent death/PD (from Kaplan–Meier analyses; months approximated as 4 weeks). Results: 89 patients were treated in both core and OLE studies (42 LAN–LAN [SD, n ¼ 41]; 47 PBO–LAN [SD, n ¼ 15]); 40% of the LAN–LAN vs. 47% of the PBO–LAN group had treatment-related AEs. Overall median LAN PFS, based on the intent-totreat population (n ¼ 101), was 38.5 months. Seven PD events (no deaths) occurred during the OLE in 15 patients entering with SD from the PBO arm of core study. In total, 32 patients with PD whilst receiving PBO in the core study entered OLE (of 59 potentially eligible); NETs were in pancreas in 17 patients, midgut in 10, hindgut in one, and of other/unknown origin in four. Of these patients, 20 had subsequent PD during the OLE and three died; median time to subsequent death/PD was 19.0 months [95% CI: 10.1; 26.7]. Conclusions: The final analysis of the CLARINET OLE study suggests benefit with LAN in patients who had experienced PD when receiving no NET-specific treatment (PBO), with median time to subsequent death/PD of 19 months. Clinical trial identification: NCT00842348 Legal entity responsible for the study: Ipsen Pharma Funding: Ipsen Pharma Disclosure: J.B. Cwikla: Travel expenses paid by Ipsen. E.M. Wolin: Advisory boards for Ipsen and Advanced Accelerator Applications. M. Pavel: Advisory boards and honoraria from Ipsen, Novartis, Pfizer and Lexicon; funding from Ipsen and Novartis for scientific research. A.T. Phan: Consultant for Ipsen and Novartis; Speaker’s bureau for Lilly, Genentech, Celgene, Lexicon, Novartis and Ipsen. M. Raderer: Honoraria from Ipsen, Novartis, Celgene, Roche and EASAI. E. Sedlackova: Consultant role for Ipsen; Speaker’s bureau for Ipsen, Novartis and Merck; funding from Ipsen and Novartis for scientific research; travel expenses paid by Ipsen, Novartis, Roche and Merck. G. Cadiot: Consultant role for Ipsen, Novartis, Advanced Accelerator Applications, and Keocyt; funding from Ipsen and Novartis for scientific research. J. Capdevila: Research, advisory role and speaker for Ipsen, Pfizer and Novartis. G. Rindi: Speaker’s bureau for Ipsen and Novartis. C. Lombard-Bohas: Consultant role for Ipsen, Pfizer and Novartis; funding from Ipsen and Novartis for scientific research. N. Liyanage, X-M. Truong Thanh: Employee of Ipsen P. Ruszniewski: Honoraria from Ipsen and Novartis; consultant role for Ipsen; Speaker’s bureau for Ipsen and Novartis; funding from Ipsen and Novartis for scientific research. M. Caplin: Honoraria from, consulted for, and participated in Speaker’s bureau for Ipsen and Novartis.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 452P

Temozolomide-capecitabine (TemCap) chemotherapy for neuroendocrine neoplasms (NENs): Time to maximum response and optimal treatment duration

A. Lamarca1, J. Barriuso2, L. McCallum1, G. Papaxoinis1, M. Nasralla1, C. Nuttall1, M. Frizziero1, Z. Kordatou1, M.G. McNamara2, R.A. Hubner1, P. Manoharan3, W. Mansoor1, J.W. Valle2 1 Medical Oncology, The Christie NHS Foundation Trust, Manchester, UK, 2Medical Oncology, The University of Manchester/The Christie, Manchester, UK, 3Radiology/ Nuclear Medicine, The Christie NHS Foundation Trust, Manchester, UK Background: TemCap is an option for treatment of NENs; benefit of treatment until progression rather than a fixed 6-month (mo) course remains unclear. Methods: Patients (pts) diagnosed with advanced NEN (pathology-confirmed), treated with TemCap with follow-up and available radiological response data were eligible for this retrospective study. Efficacy was assessed by RECIST v1.1. Results: Of 62 pts identified (Jan’12-Jan’17), 60 were eligible. Median (med) age at starting TemCap was 63.6 yrs; 50% were male; Performance Status (PS) 0-1 (83.3%), 2 (16.7%); with NEN of lung (33.3%), pancreas (21.7%), small bowel (16.7%), colorectal (3.3%) and other (25.0%) origin. The med Ki-67 was 12% (range 1-29); most (83.3%) were well-differentiated [grade (G)1/typical (18.3%); G2/atypical (65%); G3 (16.7%)], non-functional (75.0%) and metastatic (90.0%). Pts received TemCap as first- (33.3%) or second- (35.0%) line, for a med of 5.58 mo (95%CI 5.33-5.78). After 6 cycles, 38 pts (63.3%) were progression-free (i.e. eligible for maintenance TemCap [mTemCap]); 11 received mTemCap, 27 did not. Rationale for mTemCap was good response (n ¼ 7), good tolerance (n ¼ 3) or pts’ wishes (n ¼ 1). Overall, 29 pts (48.3%) had stable disease and 14 pts (23.3%) achieved a partial response (PR); med reduction in responding pts was -56.7% (95%CI -76.4 to -33.3); 4 additional pts (6.67%) achieved a reduction >20% but 55

G1 (Ki67 < ¼2) G2(220%: 14.5%). When considering Ki67 as continuous variable, and by including also morphology and beta-catenin in the multivariable OS model, patient-specific estimates were obtained, thereby improving the prognostic classification, particularly for G3 patients, which could be split in further sub-groups (Table). Harrell’s C index was 0.864. Similar results were obtained for DFS. Conclusions: WHO 2010 classification stratifies the risk of OS and DFS for G1 and G2 diseases. On the other hand, the risk of death for G3 disease varies according to Ki67 values, morphology and beta-catenin. Morphology has the strongest predictive power, segregating two macro groups in which beta-catenin has a lower differential effect while a prognostic gradient by Ki67 (up to Ki67 55) is evident. Legal entity responsible for the study: FONDAZIONE IRCCS Istituto Nazionale Tumori, Milano Ethical Committee Approved 48/16 Funding: None Disclosure: All authors have declared no conflicts of interest.

458P

The preoperative blood lymphocyte-to-monocyte ratio acts as a superior prognostic factor and predicts tumor metastasis in gastric neuroendocrine neoplasms after surgery

computed tomography (aCT), magnetic resonance imaging (aMRI) and ultrasound (aUS)) were the main outcomes. Results: Nine hundred and thirty-six patients were identified with median follow-up 47 months. The mean age was 59, 51% were female, and distribution of primary cancers was: small intestine 47%, pancreas 20%, large intestine 21%, rectum 6.4%, stomach 6.0%. The median survival time to a composite outcome of recurrence or death was 7.2 years, and 9.5 years if censoring on death. The cumulative incidence of recurrence was 8.4% (95% CI 6.8% to 10.3%) within one year, 33.7% (95% CI 30.4% to 36.9%) within five years, and 48.5% (95% CI 44.4% to 52.4%) within 10 years. The rate of recurrence significantly increased with age (HR ¼ 1.529 for age 50-70 compared to < 50, p ¼ 0.0003) and pancreatic primary (HR ¼ 1.463, p ¼ 0.0006), but not income quintile (p ¼ 0.1071), rurality (p ¼ 0.1931) or gender (p ¼ 0.3787). The rate of use of aCTs, aMRIs and aUS decreased over time, from 1.04 per 100 patient-days in months 1-3 to 0.22 at months 49-60. On average, 1.59 abdominal CTs per patient were performed in the first year, 0.83 in the second year and 0.52 in years 3-5. Conclusions: Unlike colon cancer, significant numbers of NETs recur between 5-10 years after curative surgical resection. These data support the lengthening of follow-up for resected NETs to a minimum of 10 years. Future research should focus on the impact of imaging on early detection of recurrence and survival outcomes. Legal entity responsible for the study: Sunnybrook Research Institute Funding: AGITG Disclosure: D. Chan: Travel support from Novartis; honoraria from Ipsen. E. Segelov: Honoraria from Roche, Bayer, Ipsen and Pfizer; research funding from Merck Serono and Ipsen; travel support from Roche and Ipsen. S. Singh: Honoraria and travel funding from Ipsen, Pfizer and Novartis. All other authors have declared no conflicts of interest.

460P L. Cao, H. Chang-Ming, J. Lin, J. Lu, C. Zheng Department of Gastric Surgery, Fujian Medical University Union Hospital, Fuzhou, China Background: The aim of this study is to investigate the prognostic significance of the preoperative blood lymphocyte-to-monocyte ratio (LMR) in gastric neuroendocrine neoplasms (g-NENs). Methods: We enrolled 177 patients who had been diagnosed with g-NENs and undergone radical surgery. Receiver operating characteristic curve analysis was used to identify the optimal value for the LMR. Univariate and multivariate survival analyses were used to identify prognostic factors. A nomogram was adopted to predict recurrence free survival (RFS) and overall survival (OS) after surgery. Results: The LMR was significantly lower in patients with g-NENs than in matched normal volunteers (NVs) (P < 0.05) and was associated with age, tumor site, tumor size, depth of invasion, the lymph node ratio (LNR) and lymphovascular invasion (all P < 0.05). Multivariate analysis demonstrated that the LMR was an independent prognostic factor for RFS and OS. The concordance index (C-index) of the nomograms for RFS (OS), which included the lymph node ratio, histological type and the LMR, was 0.776 (0.760), which was higher than the C-index of the traditional TNM staging system [0.678 (0.667)]. The recurrence rate was 38.9% (69/177), and the median time to recurrence was 10 months. We noted a significant correlation between the LMR and tumor recurrence, especially liver, peritoneal and lymph node metastases (all P < 0.05). Conclusions: As an independent prognostic factor for survivals in patients with gNENs, the LMR combined with the lymph node ratio and histological type had a more superior ability to predict clinical outcomes in post-surgery patients than the traditional TNM staging system. Patients with low LMRs require close surveillance to identify tumor recurrence early. Legal entity responsible for the study: Changming Huang Funding: None Disclosure: All authors have declared no conflicts of interest.

459P

Follow-up and recurrence in resected gastroenteropancreatic neuroendocrine tumours: A population-based study

D. Chan1, E. Segelov2, L. Moody3, N. Liu4, H. Fischer4, P. Austin4, S. Singh1 Medical Oncology, Sunnybrook Odette Cancer Center, Sunnybrook HSC, Toronto, ON, Canada, 2Medical Oncology, Monash University and Monash Health, Melbourne, Australia, 3IHPME, University of Toronto, Toronto, ON, Canada, 4Biostatistics, Institute of Clinical Evaluative Sciences, Toronto, ON, Canada

1

Background: Neuroendocrine tumours (NETs) are uncommon. Little data exist to guide follow-up in resected disease, with no consensus regarding the optimal follow-up frequency or modality. Follow-up imaging regimens are extrapolated from other gastrointestinal tumours. As NETs are heterogeneous, this may result in both over-use and underuse of investigations in patients. Methods: A population-based retrospective cohort study using linked data from the Institute for Clinical Evaluative Sciences and the Ontario Cancer Registry (capturing more than 99% of incident cases in Ontario) was conducted to evaluate patients diagnosed with gastroenteropancreatic NETs in Ontario, Canada from 1994 to 2012. Recurrence-free survival and the frequency of cross sectional imaging (abdominal

Volume 28 | Supplement 5 | September 2017

Elevated levels of 5-HIAA and CgA in patients with PanNETs from the CLARINET Study

A.T. Phan1, E.M. Wolin2, N. Liyanage3, B. Mirakhur4, S. Pitman Lowenthal4, A. Vinik5, G.A. FisherJr6, M. Pavel7 1 Medical Oncology, University of New Mexico Comprehensive Cancer Center, Albuquerque, NM, USA, 2Medical Oncology, Monteﬁore Einstein Center for Cancer Care, Bronx, NY, USA, 3R&D Statistics, Ipsen Innovation, Les Ulis, France, 4Medical Affairs, Ipsen Biopharmaceuticals, Basking Ridge, NJ, USA, 5Diabetes Research, Eastern Virginia Medical School, Norfolk, VA, USA, 6Medical Oncology, Stanford University, Stanford, CA, USA, 7Gastroenterology and Hepatology, Endocrinology & Metabolic Diseases, Charite´ University Medicine Berlin, Berlin, Germany Background: While carcinoids frequently synthesize, and secrete serotonin into the circulation, and 5-HIAA is a common biomarker in the carcinoids, measurement of 5HIAA in non-carcinoid PanNET patients (i.e. no hormone-related symptoms or nonfunctional) is not routinely recommended by international guidelines. The incidence of serotonin-producing PanNETs may be underestimated, with potential impact on clinical outcome when serotonin levels remain elevated. We sought to characterize 5HIAA and CgA levels in PanNET patients who participated in the large placebocontrolled phase III CLARINET Study. Methods: Evaluable data available for urinary 5-HIAA and serum CgA for patients with PanNET in CLARINET study were analyzed. Urinary 5-HIAA and Serum CgA were assessed at baseline and every 12 weeks thereafter through Week 96. Changes in urinary 5-HIAA and serum CgA levels were calculated using a non-parametric Wilcoxon 2-sample test. Biochemical response for urinary 5-HIAA or serum CgA was defined as baseline >upper limit of normal (ULN, 41.6 mmol/d 5-HIAA; 98.1 mg/L CgA) and 50% decrease from baseline or a decrease to a value ULN on study. Results: 91/204 patients in CLARINET had PanNETs. Evaluable data for urinary 5HIAA and serum CgA concentrations were available in 79 and 88 patients, respectively. A substantial number of patients with PanNET had elevated (>ULN) urinary 5-HIAA levels (21/79; 27%) and/or serum CgA (63/88; 72%). Among the 21 PNET patients with baseline 5-HIAA >ULN, biochemical response was achieved in 85% (11/13) lanreotide-treated patients compared with 63% (5/8) in patients on placebo at the last available value (p ¼ 0.33). Among patients with baseline CgA >ULN, biochemical response was achieved in 66% (19/29) of lanreotide vs. 18% (6/34) of placebo-treated patients (p ¼ 0.0002). Limited sample sizes precluded robust analysis for statistically significant differences in the lanreotide vs. the placebo group among patients with elevated biomarkers at baseline and biochemical response. Conclusions: The percentage of patients with elevated urinary 5-HIAA was unexpected. The concept of PanNET and secretion of serotonin may need to be redefined. The potential of 5-HIAA and CgA as biomarkers of response and follow-up in nonfunctioning PanNET is alluring, but requires further study. Data from additional prospective studies are needed to impact clinical practice guidelines. Clinical trial identification: NCT00353496 Legal entity responsible for the study: Ipsen Biopharmaceuticals Funding: Ipsen Biopharmaceuticals r A.T. Phan: Consultant/Advisor: Ipsen, Novartis; speakers’ bureau: Celgene, Genentech, Eli Lilly, Ipsen, Novartis; research support: Incyte, Ipsen, Lexicon, Novartis, Sanofi. E.M. Wolin: Consultant/Advisor: Celgene, Ipsen, Novartis, Pfizer; research support: Ipsen (Inst), Novartis (Inst), Pfizer (Inst). N. Liyanage: Employee of Ipsen. B. Mirakhur: Employee of Ipsen Biopharmaceuticals. S. Pitman Lowenthal: Employee of Ipsen Biopharmaceuticals. A. Vinik: Const/Advisor: Isis, Merck, Neurometrix, Pamlab, Pfizer; speakers bureau: Merck, Pamlab; research support: BMS, Celgene, Eli Lilly, Medivation, Newlink Genetics, Novartis,

doi:10.1093/annonc/mdx368 | 153

abstracts Pharmaceutical Research Associates, Polaris, PRA International, Tercica, XBiotech. G.A. Fisher, Jr: Consultant/Advisor: Genentech, Ipsen; research support: BMS, Celgene, Lilly, Ipsen, Medivation, Newlink Genetics, Novartis, Pharmaceutical Research Associates, Polaris, PRA International, Tercica, XBiotech; stock ownership: Seattle Genetics. M. Pavel: Consultant/Advisor: Ipsen, Lexicon, Novartis, Pfizer; research support: Novartis.

461P

The prognostic value of cytokeratin 7, 19, thyroid transcription factor1 and CD117 expression in lung neuroendocrine tumors of various grades

L. Gurevich1, N. Korsakova1, I. Voronkova1, I. Kazantseva1, V. Ashevskaya1, A. Titov2, L. Kogoniya2, M. Byakhova1 1 Pathology, Moscow Regional Research and Clinical Institute (“MONIKI”), Moscow, Russian Federation, 2Surgery, Moscow Regional Research and Clinical Institute (“MONIKI”), Moscow, Russian Federation Background: Neuroendocrine tumors of the lung (NETL) are a wide range of tumors with various malignancy grades and prognosis. Methods: We performed immunohistochemical assessment of the diagnostic biopsies and surgical specimens from 205 patients with NETL aged 55 6 14 years and identified 61 (29,8%) typical carcinoids (TC), 44 (21,5%) atypical carcinoids (ATC), 84 (41%) small cell neuroendocrine carcinomas (SCNEC) and 16 (7,8%) large cell neuroendocrine carcinomas (LCNEC). Markers of neuroendocrine differentiation (synaptophysine, chromogranine A and CD56) and cytokeratins (CK) 7 and 19, thyroid transcription factor-1 (TTF-1), CD117 were used. Results: Most often, the expression of CK7 and CK19 was found in LCNEC (71.4%, 10/ 14 and 91.7%, 11/12 respectively), less frequently, in ATC and SCNEC (52,8%, 19/36 and 52.4%, 22/42; 43,9%, 29/66 and 68,2%, 45/66 of cases, respectively), whereas in TC it was rare (13,3%, 6/45 and 19,3%, 11/57 respectively). The rates of CK7 and 19 expression were significantly lower in the TC, compared to the SCNEC and LCNEC (h < 0.01, v.) The expression of TTF-1 was very rare in the TC (11,6%, 5/43 of cases) and significantly more often in ATC (60.5%, 23/38) and in SCNEC and LCNEC (79,2%, 57/72 and 75%, 9/12 of cases, respectively). TTF-1 expression was significantly less frequent in typical than in ATC, SCNEC and LCNEC (h < 0.01, v.). The expression CD117 was absent in the TC (0%, 0/27), very rare in the ATC (17,4%, 4/23) and significantly more often in SCNEC and LCNEC (95,7%, 43/47 and 42,8%, 3/7 of cases, respectively). Conclusions: Expression of TTF-1, CK7, 19 and CD117 in the NETL is characteristic for a less differentiated cell immunophenotype and allows for identification of the risk group with unfavorable clinical outcome among low-grade TC and ATC. Legal entity responsible for the study: L. Gurevich Funding: None Disclosure: All authors have declared no conflicts of interest.

461P

An open-label, multicenter Phase 1b study of radium-223 1 paclitaxel in cancer patients with bone metastases

R. Geva1, J. Lopez2, S. Danson3, H. Joensuu4, A. Peer5, S.J. Harris6, F. Souza7, B.A. Ploeger8, K.M.C. Pereira9, R. Perets5 1 Research Unit, Division of Oncology, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel, 2 Drug Development Unit, The Royal Marsden Hospital and The Institute of Cancer Research, Sutton, UK, 3Department of Oncology and Metabolism, Shefﬁeld Experimental Cancer Medicine Centre, Weston Park Hospital, Shefﬁeld, UK, 4Department of Oncology, Helsinki University Hospital, Helsinki, Finland, 5Oncology Division, Rambam Health Care Campus, Haifa, Israel, 6Department of Medical Oncology, The Royal Marsden NHS Foundation Trust and Institute of Cancer Research, Sutton, UK, 7 Clinical Pharmacodynamics Expert Oncology, Bayer HealthCare Pharmaceuticals, Whippany, NJ, USA, 8Drug Discovery, Pharmaceuticals, Clinical Pharmacometrics, Bayer ao Pharma AG, Berlin, Germany, 9Pharmaceuticals Development, Bayer Pharma AG, S~ Paulo, Brazil

Annals of Oncology neuroendocrine (1 pt). 7 pts had received 3 prior chemotherapy regimens. In the 13 pts who completed cycle 3, grade 3 neutropenia rates in cycles 2 and 3 were 31% and 8%, respectively, vs 23% in cycle 1; there were no cases of grade 4 neutropenia or grade 3/4 thrombocytopenia. No pts discontinued treatment due to toxicity from the treatment combination. Safety data for the breast cancer pt subset will be presented. The myelosuppression model showed an additive effect of Ra-223 to PTX-induced neutropenia, with an additional 10% average decrease in absolute neutrophil count vs PTX alone. Conclusions: In pts with solid tumors and bone mets, Ra-223 was well tolerated when combined with PTX, with an additional 10% average decrease in neutrophil levels compared with PTX monotherapy. The combination should be explored further in pts with bone mets. Clinical trial identification: NCT02442063 Legal entity responsible for the study: Bayer HealthCare Pharmaceuticals Funding: Bayer HealthCare Pharmaceuticals Disclosure: R. Geva: Honoraria from BMS, Lilly, Medison, Roche, Novartis; consultant or advisory role for MSD and Novartis. S. Danson: Consultant/advisor: Incanthera; speakers bureau: BMS, Boehringer Ingelheim; research funding: Lilly, GSK, BMS, Astellas, Merke, Incyte, Novartis, Virtua, Boehringer Ingelheim; travel, accommodations, expenses: BMS, MSD, Boehringer Ingelheim. H. Joensuu: Stock/ownership: Orion Pharma, Sartar Therapeutics; honoraria: Orion Pharma; consultant/advisor: Orion Pharma, Sartar Therapeutics. A. Peer: Consultant/advisor: AstraZeneca, Novartis, Sanofi, Merck. S.J. Harris: Travel, accommodations, expenses: BMS (travel by car to and from conference). F. Souza: Employment, research funding, and travel, accommodations, and expenses: Bayer B.A. Ploeger: Employment and stock/ownership with Bayer K.M.C. Pereira: Employment: Bayer. R. Perets: Consultant or advisor: Musli Thyropeutics. All other authors have declared no conflicts of interest.

462P

L-L. Cao, H. Chang-Ming, J. Lu, J. Xie, P. Li Department of Gastric Surgery, Fujian Medical University Union Hospital, Fuzhou, China Background: This study investigated the predictive value of the tumor-associated neutrophil-to-lymphocyte ratio (TA-NLR) on clinical outcomes for patients with gastric neuroendocrine neoplasms (g-NENs) after radical surgery. Methods: Data from 142 patients who were diagnosed with g-NENs and underwent radical gastrectomy at our department from March 2006 to March 2015 were prospectively collected and retrospectively analyzed. Receiver operating characteristic curve analysis was used to identify the optimal value for TA-NLR. Univariate and multivariate survival analysis were used to identify prognostic factors for g-NENs. A nomogram was adopted to predict RFS and OS after surgery. Results: TA-NLR was not significantly associated with clinical characteristics (all P > 0.05). TA-NLR significantly correlated with tumor recurrence, especially with liver and lymph node metastasis (both P < 0.05). A multivariate Cox regression analysis identified the TA-NLR as an independent prognostic factor for recurrence-free survival (RFS) and overall survival (OS) (both P < 0.05). The concordance index (C-index) of the nomograms, including the TA-NLR, Ki-67 index and lymph node ratio, for RFS (OS) was 0.788(0.759) and was higher than the C-index of the traditional TNM staging system [0.672(0.663)]. Conclusions: TA-NLR was an independent prognostic factor for patients with g-NENs regarding RFS and OS. Nomograms with the TA-NLR, Ki-67 index and lymph node ratio had a superior ability to predict clinical outcomes for postoperative g-NENs patients, as well as the traditional TNM staging system. Legal entity responsible for the study: Changming Huang Funding: None Disclosure: All authors have declared no conflicts of interest.

463P Background: Concomitant radium-223 (Ra-223) and chemotherapy is a possible option for cancer patients (pts) with bone metastases (mets). Both treatments impact hematologic parameters, and myelosuppression risk during coadministration is unknown. This phase 1b study (NCT02442063) in cancer pts with bone mets evaluated the safety of Ra-223 þ paclitaxel (PTX) and the mode of interaction between treatments regarding myelosuppression. Methods: Eligible pts had a confirmed malignant solid tumor with 2 bone mets and were candidates for PTX treatment. Treatment included 7 PTX cycles (90 mg/m2 IV per wk as per local standard of care; 3 wk on/1wk off) combined with 6 Ra-223 cycles (55 kBq/kg IV; 1 injection every 4 wk, starting at PTX cycle 2). The primary end point was percentage of pts with neutropenia and thrombocytopenia during treatment with Ra223 þ PTX (cycles 2 and 3) vs PTX alone (cycle 1). A previously developed doseexposure-response model describing the time course of PTX and Ra-223–induced suppression of absolute neutrophil counts was used to evaluate the mode of interaction (additive or synergistic) between Ra-223 and PTX. Results: Of 22 enrolled pts, 15 were treated; 13 completed cycles 1-3 and were included in the pharmacodynamics analysis. Tumors in treated pts were breast (7 pts), prostate (4 pts), bladder (1 pt), non-small cell lung (1 pt), myxofibrosarcoma (1 pt), and

154 | Endocrine and neuroendocrine tumours

A nomogram based on tumor-associated neutrophil-to-lymphocyte ratio to predict survival prognosis for patients with gastric neuroendocrine neoplasms

Plasma protein fingerprinting and machine learning for the diagnosis of small intestinal neuroendocrine tumors: The nordic NET biomarker group EXPLAIN study

M.M. Kjellman1, S. Welin2, U. Knigge3, H. Gronbaek4, H. Sorbye3, E. Thiis-Evensen5, M. Thyregod Jørgensen6, V. Johanson7, C. Schalin-J€antti8, P. Myrenfors9, T. Stro¨m9, K. Becker9, R. Belusa9 1 Department of Endocrine Surgery, Karolinska Institutet, Stockholm, Sweden, 2 Department of Endocrine Oncology, Uppsala Akademiska, Uppsala, Sweden, 3 Department of Surgery, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark, 4Department of Clinical Medicine, Aarhus University Hospital, Aarhus, Sweden, 5Department of Gastroenterology, Oslo University Hospital, Oslo, Sweden, 6 Department of Medical Gastroenterology, Odense University Hospital, Odense, Denmark, 7Department of Surgery, Sahlgrenska Hospital, Gothenburg, Sweden, 8 Department of Endocrinology, Helsinki University Hospital, Helsinki, Finland, 9Medical Department, Ipsen, Stockholm, Sweden Background: Small intestinal neuroendocrine tumors (siNETs) are notoriously difficult to diagnose, especially in an early stage. The EXPLAIN study aimed to investigate

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 92 plasma proteins (PP), previously shown to be cancer related, in an attempt to improve the accuracy in diagnosis of siNETs. Methods: This non-interventional exploratory study in the nordic countries analysed 136 patients with siNET from 17 hospitals and 144 age and sex matched controls (all with written consent). Exclusion criteria: NET not confirmed, previously treated for NET, other malignant diseases, chronic inflammatory disease, kidney or liver failure. Blood samples (4 ml) were obtained at first visit. Samples analysis used the Proseek Multiplex Oncology II assay (OLink) to measure relative levels of the 92-cancer related PP. In addition, chromogranin A (CgA) was analyzed centrally (Akademiska Lab. Uppsala). Data was subjected to statistical supervised learning techniques (SSLT): random forest and support vector machine. Results: This is the first interim analysis. Patient characteristics: age 65610 (mean6SD), 58% male, 48% G1 and 52% G2, 88% N1 and 65% M1, 23% >3 bowel mov/d and 11.5% >3 flushes/d. CgA (mean (SD), nmol/L) in 115 patients free from proton pump inhibitor treatment (PPI): 42.37 (86.62), in 21 NET patients with PPI: 68.41 (74.21), in 132 controls free from PPI: 3.67 (3.57) and in 12 controls treated with PPI: 11.83 (8.97). Several PP (>20) showed significant p5M in the 6M arm) populations. Subgroups and long lasting neuropathy results are also reported here. Results: From May 2009 to May 2014, 2022 pts were randomized from 129 centers and 2010 (99%) and 1757 (87%) were included in the mITT and mPP populations, respectively. With a median follow-up of 4.3 years, the 3-year DFS rate was 72% and 76% (HR ¼ 1.24; 95% CI 1.05–1.46, p ¼ 0.01) for the 3M and 6M mITT populations, respectively and 72% and 78% (HR ¼ 1.36; 95% CI 1.14–1.63, p ¼ 0.0008) for the 3M and 6 M mPP populations. In the mITT FOLFOX treated population (90% of pts), 3year DFS was 81% (3M) and 83% (6M) for T1-3/N1 pts (N ¼ 1106, HR ¼ 1.15 95%CI 0.89-1.49) and 58% (3M) and 66% (6M) for T4/N2 pts (N ¼ 702, HR ¼ 1.44 95%CI 1.14-1.82). Grade >1 neuropathy was observed in 36% and 67% of pts (p < 0.0001) in the 3M and 6M arms, respectively. With a median follow-up of 3.6 years, final residual grade >1 neuropathy was 2.8% and 7.4% (p < 0.0001), in the 3M and 6M arms, respectively. Conclusions: The IDEA France study, with 90% of pts treated with mFOLFOX6, shows that 6M adjuvant treatment is superior to 3M. However, this difference was not significant in the mITT and mPP T1-3N1 populations suggesting that 3M of the mFOLFOX6 regimen could be an option for these pts. Clinically relevant (grade>1) neuropathy was significantly higher in the 6M arm, with long-lasting neuropathy in 7.4% of pts. Clinical trial identification: Registration Number (European Union Drug Regulating Authorities Clinical Trials): 2009-010384-16 Legal entity responsible for the study: GERCOR - Groupe Coope´rateur Multidisciplinaire en Oncologie Funding: French Ministry of Health and French National Cancer Institute (INCa) Disclosure: J. Taieb: Advisory board or Board of directors: Sanofi, Baxalta, Roche, Merck, Amgen, Lilly, Celgene. F. Bonnetain: Advisory board or Board of directors: Roche, Ipsen, Amgen, Nestle, Novartis; corporate-sponsored research: Novartis, Roche. L. Mineur: Advisory board or Board of directors: Amgen, Sanofi, Bayer, Roche; corporate-sponsored research: Sanofi, Merck, Chugai. J. Bennouna: Advisory board or Board of directors and honorarium: BMS, Roche, Boehringer Ingelheim, AstraZeneca. D. Vernerey: Honorarium: Janssen, Celgene. Advisory board: HallioDx. C. Lepere: Advisory board or Board of directors: Ipsen. O. Bouche: Advisory board or Board of directors: Merck Serono, Roche, Amgen. M. Ychou: Advisory board or Board of directors: Roche, Bayer, Amgen, Lilly. T. Andre´: Advisory board or Board of directors: BMS, Amgen, Roche; corporate-sponsored research: BMS, Roche; honoraria: Baxter, Bayer, Lilly, MSD, Sanofi, Mundipharma, Novartis. All other authors have declared no conflicts of interest.

C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

474O

Annals of Oncology 28 (Supplement 5): v158–v208, 2017 doi:10.1093/annonc/mdx393

Treatment outcome according to tumor RAS mutation status in TRICOLORE trial: A randomized phase 3 trial of S-1 and irinotecan plus bevacizumab versus mFOLFOX6 or CapeOX plus bevacizumab as firstline treatment for metastatic colorectal cancer

Y. Komatsu1, A. Takashima2, T. Denda3, M. Gamoh4, I. Iwanaga5, H. Shimodaira6, M. Nakamura7, T. Yamaguchi8, H. Takahashi9, K. Kobayashi10, M. Tsuda11, Y. Kobayashi12, H. Baba13, M. Kotake14, C. Ishioka6, A. Sato15, S. Yuki16, S. Morita17, S. Takahashi6, K. Shimada18 1 Cancer Center, Hokkaido University Hospital, Sapporo, Japan, 2Gastrointestinal Medical Oncology, National Cancer Center Hospital, Tokyo, Japan, 3Gastroenterology, Chiba Cancer Center, Chiba, Japan, 4Medical Oncology, Osaki Citizen Hospital, Osaki, Japan, 5Medical Oncology, Japanese Red Cross Kitami Hospital, Kitami, Japan, 6Medical Oncology, Tohoku University Hospital, Sendai, Japan, 7Aizawa Comprehensive Cancer Center, Aizawa Hospital, Matsumoto, Japan, 8Surgery, Tokyo Metropolitan Cancer and Infectious Diseases Center Komagome Hospital, Tokyo, Japan, 9Clinical Oncology, Japanese Red Cross Ishinomaki Hospital, Ishinomaki, Japan, 10Medical Oncology, Showa University Northern Yokohama Hospital, Yokohama, Japan, 11Medical Oncology, Hyogo Cancer Center, Akashi, Japan, 12Medical Oncology, Kushiro Rosai Hospital, Kushiro, Japan, 13Surgery, Kumamoto University, Kumamoto, Japan, 14 Surgery, Kouseiren Takaoka Hospital, Takaoka, Japan, 15Medical Oncology, Hirosaki University Graduate School of Medicine, Hirosaki, Japan, 16Gastroenterology and Hepatology, Hokkaido University Hospital, Sapporo, Japan, 17Biomedical Statistics and Bioinformatics, Graduate School of Medicine, Kyoto University, Kyoto, Japan, 18Medical Oncology, Showa University Kouto-Toyosu Hospital, Tokyo, Japan Background: Combination therapy with oral fluoropyrimidine and irinotecan (CPT11) has not yet been established as first-line treatment for metastatic colorectal cancer (mCRC). However, several studies of S-1 and CPT-11 plus bevacizumab (Bmab) combination therapy have shown promising efficacy in mCRC, suggesting the potential to replace mFOLFOX6 or CapeOX plus Bmab. We performed a randomized phase 3 trial to determine whether S-1 and CPT-11 plus Bmab is non-inferior or superior to mFOLFOX6 or CapeOX plus Bmab in terms of progression-free survival (PFS). Methods: The TRICOLORE trial was a randomized, open-label, phase 3 trial. Chemotherapy-naı¨ve patients with mCRC were randomized to receive either mFOLFOX6 or CapeOX plus Bmab (group A) or S-1 and CPT-11 plus Bmab (group B; 3-week regimen: 7.5 mg/kg Bmab, 150 mg/m2 CPT-11 on day 1, and 40 60 mg S-1 twice daily for 2 weeks, followed by a 1-week rest; or 4-week regimen: 5 mg/kg Bmab, 100 mg/m2 CPT-11 on days 1 and 15, and 40 60 mg S-1 twice daily for 2 weeks, followed by a 2-week rest). The primary endpoint was PFS. The non-inferiority margin was a hazard ratio (HR) of 1.25 based on the assumption of a median PFS of 11/12 months in group A/group B (power 0.85, 1-sided alpha 0.025). The primary tumor RAS status of patients consented to submit tissue sample were centrally analyzed. Results: A total of 487 patients were enrolled from June 2012 to September 2014. Data were analyzed after confirming >374 events as planned. All demographic factors were well balanced. Median PFS was 10.8 months in group A and 14.0 months in group B (HR 0.85, 95% CI: 0.70–1.03, p < 0.001 for non-inferiority, p ¼ 0.087 for superiority). The RAS mutation status was evaluable in 67.6%. In the RAS wild-type subgroup, median PFS was 11.6 months in group A and 15.9 months in group B. In the RAS mutanttype subgroup, median PFS was 9.3 months in group A and 11.3 months in group B. Conclusions: S-1 and CPT-11 plus Bmab was non-inferior to mFOLFOX6 or CapeOX plus Bmab with respect to PFS and has now become a recommended 1st-line treatment for mCRC irrespective of RAS status. Clinical trial identification: UMIN000007834 2012/05/11 Legal entity responsible for the study: The Tokyo Cooprative Oncology Group Funding: The Tokyo Cooprative Oncology Group (with funding from Taiho) Disclosure: Y. Komatsu: Other substantive relationships: Taiho Pharmaceutical, Lilly, MSD, Ono Pharmaceutical, Novartis, Chugai Pharma, Yakult, Merck Serono, Pfizer, Bayer. A. Takashima: Corporate-sponsored research: Taiho Pharmaceutical, Gilead Sciences, Merck Serono. M. Gamoh: Corporate-sponsored research: Taiho Pharmaceutical, Gilead Sciences, Merck Serono. H. Shimodaira: Corporate-sponsored research: Taiho, Eisai, Bayer. H. Baba: Advisory board or Board of directors: Taiho Pharmaceutical Co., Ltd.; corporate-sponsored research: Taiho Pharmaceutical Co., Ltd. C. Ishioka: Mochida, Kyowa-K, Eisai, Chugai, Tsumura, Novartis, Merck Serono, Daiichi Sankyo, Takeda, Nihon-Kayaku, Yakult, Taiho, Ono, Astellas, Asahi Kasei, Kissei, Bristol-Myers Squibb, Mochida, Chugai, Novartis, Lilly, Bayer. A. Sato: Advisory board or Board of directors: Taiho Pharmaceutical Co., Ltd. Chugai Pharma Co., Ltd., Sumitomo Dainippon Pharma Co., Ltd.; corporate-sponsored research: Taiho Pharmaceutical Co., Ltd.Chugai Pharma Co., Ltd. S. Yuki: Speakers’ bureau: Chugai Pharmaceutical, Eli Lilly Japan, Bayer Yakuhin, Takeda Pharmaceutical, Taiho Pharmaceutical, Merck Serono. S. Morita: Corporatesponsored research: Taiho; honorarium: Taiho. All other authors have declared no conflicts of interest.

abstracts

Annals of Oncology

475O

mFOLFOXIRI 1 panitumumab versus FOLFOXIRI as first-line treatment in patients with RAS wild-type metastatic colorectal cancer m(CRC): A randomized phase II VOLFI trial of the AIO (AIO-KRK0109)

M. Geissler1, U.M. Martens2, R. Knorrenschield3, J. Greeve4, A. Florschuetz5, A. Tannapfel6, S. Wessendorf1, T. Seufferlein7, S. Kanzler8, V. Heinemann9, S. Held10, A. Reinacher-Schick11 1 Oncology, Hematology, Gastroenterology, Infectious Diseases, Klinikum Esslingen, Esslingen, Germany, 2Medical Clinic III, SLK-Kliniken Heilbronn GmbH, Heilbronn, atsklinkum Marburg, Marburg, Germany, Germany, 3Hematology/Oncology, Universit€ 4 Medical Clinic I, St. Vincenz Krankenhaus, Paderborn, Germany, 5Internal Medicine, 6 at, Bochum, Germany, Klinikum Dessau, Dessau, Germany, Pathology, Ruhr Universit€ 7 Internal Medicine I, Ulm University, Ulm, Germany, 8Medical Clinic II, Leopoldina 9 Krankenhaus Medizinische Klinik II, Schweinfurt, Germany, Medicine III, Ludwig Maximilians University - Grosshadern, Munich, Germany, 10ClinAssess GmbH, at, Bochum, Germany Leverkusen, Germany, 11Hematology/Oncology, Ruhr Universit€ Background: Triple chemotherapy with an anti-EGFR reported promising activity with some safety concerns in single arm phase II trials. The randomized VOLFI trial evaluated activity and safety of mFOLFOXIRI þ panitumumab versus FOLFOXIRI in ECOG 0-1, primarily non-resectable mCRC patients. Methods: Prospective 2:1 randomized, multi-center, phase II trial comparing mFOLFOXIRI (Oxaliplatin 85 mg/m2, Irinotecan 150 mg/m2, 5-FU 3000mg/m2 cont. 48h, LV 200 mg/m2) þ Panitumumab 6 mg/KG (arm A) with FOLFOXIRI (Ox 85 mg/ m2, Iri 165 mg/m2, 5-FU 3200mg/m2 cont. 48h, LV 200 mg/m2; arm B), both arms q2w. Cohort 1: irresectable mCRC; cohort 2: chance of secondary resection of metastatic lesions. Primary endpoint was ORR, secondary endpoints were secondary resection rate (cohort 2), DCR, PFS, OS, toxicity, quality of life. Financially supported by an unrestricted grant from Amgen. Results: A total of 96 patients were randomized (63 arm A, 33 arm B). In arm A and B 20 (31.7%) and 11 (33.3%) patients belonged to cohort 2, respectively. ORR was 85.7% in arm A and 54.5% in arm B (p ¼ 0.0013, OR 5.000; 95%-CI 1.870-13.370). DCR was 96.8% in arm A and 78.8% in arm B (p ¼ 0.0071, OR 8.212). In arm A and B 53 (84,1%) and 25 (75.8%) tumors were left sided, 10 (15.9%) and 6 (18.2%) were located in the right colon, respectively. ORR in Arm A was 90.6% versus 60.0% (p ¼ 0.0288, OR 6.400) and in Arm B 60.0% versus 50% (p¼n.s.) for left and right located CRC, respectively. ORR between arms A and B comparing left and right sided CRC was 90.6% versus 60.0% (p ¼ 0.0039, OR 6.400; 95%-CI 1.889-21.679) and 60.0% versus 50.0% (p¼ n.s.), respectively. Secondary resections in cohort 2 were 60% (n ¼ 12) and 36.4% (n ¼ 4) in arms A and B, respectively. Serious adverse advents grade 3-5 occured in 45.3% and 24.2% in arms A and B, respectively (p ¼ 0.0496). Conclusions: mFOLFOXIRI plus panitumumab results in significantly higher response rates compared to FOLFOXIRI in RAS wild-type mCRC. Response rates, however, are differential according to tumor sidedness. High secondary resection rates were observed. Toxicity is manageable in younger fit patients with ECOG 0-1. PFS, OS, QL and TR data are still immature and will be presented at the meeting. Clinical trial identification: NCT01328171 Legal entity responsible for the study: AIO Funding: Amgen Disclosure: M. Geissler: Honoraria and advisory board from Amgen All other authors have declared no conflicts of interest.

476O

Neoadjuvant FOLFOX 4 versus FOLFOX 4 plus cetuximab versus immediate surgery for high-risk stage II and III colon cancers: A phase II multicentre randomised controlled trial (PRODIGE 22)

M. Karoui1, A. Rullier2, C. Mariette3, E. Maillard4, A. Bardier5, F. Poizat6, A. Luciani7, A. Sarran8, J-L. Legoux9, C. De Chaisemartin10, C. Lecaille11, O. Bouche12, F. Mauvais13, F. Brunetti14, M. Prudhomme15, J-F. Seitz16, C. Lepage17, J. Taieb18 1 Department of Digestive Surgery, CHU Pitie´-Salpe´trie`re, Paris, France, 2Department of Anatomopathology, Pellegrin Hospital, Bordeaux, France, 3Department of Digestive and Oncologic Surgery, Lille University Hospital, Lille, France, 4Biostatistics, FFCD, Burgundy University, INSERM U866, Dijon, France, 5Department of Pathologic Anatomy, CHU Pitie´-Salpe´trie`re, Paris, France, 6Department of Anatomopathology, Institut Paoli Calmettes, Marseille, France, 7Department of Medical Imaging, CHU Henri Mondor, Cre´teil, France, 8Department of Radiology, Institut Paoli Calmettes, Marseille, France, 9 Department of Gastroenterology, CHR La Source, Orle´ans, France, 10Department of Digestive Surgery, Institut Paoli Calmettes, Marseille, France, 11Department of Oncology, Polyclinique Bordeaux Nord, Bordeaux, France, 12Department of Medicine-Oncology, CHU Robert Debre´, Reims, France, 13Department of Digestive Surgery, Centre Hospitalier Beauvais, Beauvais, France, 14Department of Digestive Surgery, CHU Henri Mondor, Cre´teil, France, 15Department of Digestive Surgery, CHU Care´meau, Nıˆmes, France, 16 Department of Gastroenterology and Oncology, CHU La Timone, Marseille, France, 17 Department of Gastroenterology, CHU Le Bocage, FFCD, Burgundy University, INSERM U866, Dijon, France, 18Department of Gastroenterology, European Georges Pompidou Hospital, Paris, France Background: Neoadjuvant chemotherapy has proven valuable in several tumors, but not in colon cancer (CC). The present randomized phase II trial addressed this issue in patients (pts) with locally advanced CC.

Volume 28 | Supplement 5 | September 2017

Methods: Pts with resectable CC deemed as high risk T3 (extramural tumor invasion > 5 mm), T4 and/or N2 (3 or more visible lymph nodes or 1 node >10mm diameter) on initial abdominopelvic CT-scan were randomized to either receive 6 months of adjuvant FOLFOX after colectomy (arm A control), or neoadjuvant FOLFOX for 4 cycles before surgery and 8 cycles after (arm B). In RAS wild-type pts a third arm testing perioperative FOLFOX þ cetuximab has been added prior to colectomy (arm C). The primary endpoint of the study was the rate of major pathological Tumor Regression Grade (TRG) as defined by Ryan centrally assessed by 2 pathologists blinded to the pts treatment. The secondary endpoints included toxicity, perioperative morbidity, carcinologic quality and completeness of the surgery. Analysis was by intention to treat. Results: 120 pts from 37 French centres were enrolled, 94% completed preoperative chemotherapy. All but 5 pts (disease progression n ¼ 2, metastatic disease at inclusion n ¼ 1, non resectable tumor n ¼ 1, death n ¼ 1) in the preoperative arms were resected. 95% and 98% of patients underwent R0 resection in the preoperative arms and control arm, respectively. No significant differences in severe postoperative morbidity rates (Dindo Grade >3) were seen between arm A (13.7%), B (8.2%) and C (14.3%) (p ¼ 0.64). Major pathological responses (TRG 1-2) were observed in 7.7%, 44.2%, and 6.3% in arm A, B and C respectively (p < 0.001). Conclusions: Preoperative FOLFOX for locally advanced resectable CC is feasible with acceptable toxicity/morbidity and high TRG. A phase III trial to establish whether these encouraging results translate into improved long-term oncological outcome is now warranted. Clinical trial identification: NCT01675999 Legal entity responsible for the study: AP-HP Funding: Merck, PHRC 2010 Disclosure: O. Bouche: Roche, Merck-Serono, Amgen, Lilly, Boehringer Ingelheim, Bayer J-F. Seitz: Merck, Sanofi. J. Taieb: Abbvie, Amgen, Baxalta, Celgene, Lilly, Merck, Roche. All other authors have declared no conflicts of interest.

477O

Bevacizumab (Bev) or cetuximab (Cet) plus chemotherapy after progression with bevacizumab plus chemotherapy in patients with wild-type (WT) KRAS metastatic colorectal cancer (mCRC): Final analysis of a French randomized, multicenter, phase II study (PRODIGE 18)

J. Bennouna1, S. Hiret2, C. Borg3, A. Bertaut4, O. Bouche5, G. Deplanque6, E. Francois7, T. Conroy8, F. Ghiringhelli9, G. des Guetz10, J-F. Seitz11, P. Artru12, T. Stanbury13, S. Charpentier14, M. Denis14, A. Adenis15 1 IMAD - Digestive Oncology, CHU Nantes, Nantes, France, 2Medical Oncology, Institut de Cance´rologie de l’Ouest, Saint-Herblain, France, 3Medical Oncology, CHU de Besanc¸on, Besanc¸on, France, 4Biostatistics, Georges-Francois Leclerc Cancer Center, Dijon, France, 5Gastroenterology, CHU Reims, Reims, France, 6Medical Oncology, Hoˆpital Riviera Chablais, Vevey, Switzerland, 7Medical Oncology, Centre Antoine Lacassagne, Nice, France, 8Medical Oncology, Institut de Cance´rologie de Lorraine, Vandoeuvre-Le`s-Nancy, France, 9Medical Oncology, Georges-Francois Leclerc Cancer Center, Dijon, France, 10Medical Oncology, CHU Saint-Denis, Limoges, France, 11 Gastroenterology, APHM, Marseille, France, 12Gastroenterology, Hoˆpital Prive´ Jean Mermoz, Lyon, France, 13Investigation Center, UNICANCER, Paris, France, 14Molecular Biology, CHU Nantes, Nantes, France, 15Medical Oncology, Centre Oscar Lambret, Lille, France Background: Second-line treatment with chemotherapy plus Bev or Cet is now established as a valid option in mCRC. The main objective of this French multicenter, randomized open phase II trial, was to evaluate the Progression Free Survival (PFS) rate at 4 months with chemotherapy plus Bev or Cet in patients with disease progression after Bev plus chemotherapy. Methods: The main eligibility criterion was disease progression after bevacizumab þ 5FU with irinotecan or oxaliplatin in patients with WT KRAS exon 2 mCRC. Patients were randomized in Arm A (FOLFIRI or mFOLFOX6 plus Bev) or in Arm B (FOLFIRI or mFOLFOX6 plus Cet); the chemotherapy doublet was chosen according to the first line (cross over). Analyses were performed in ITT population. They were repeated on the KRAS þ NRAS WT population and in the triple negative population (KRAS, NRAS, and BRAF negative). Results: From October 2010 to May 2015, 133 patients were included in 25 sites (1 patient ineligible): 85 males (64%), PS 0 (74, 56%), 1 (54, 41%), unknown (4, 3%). The 4month PFS rate was 80.3% [95%CI (68.0% - 88.3%)] in Arm A and 66.7% [95%CI (53.6% - 76.8%)] in Arm B. Median PFS was 7.1 months in Arm A vs 5.6 months in Arm B (p ¼ 0.060). Median OS reached 15.8 months in Arm A vs 10.4 months in Arm B (p ¼ 0.073). Tumors samples were collected by a central laboratory and 95 were analysed using the KRAS/BRAF mutation analysis panel kit (KRAS exon 2,3,4 and BRAF V600E) and NRAS mutation detection kit (exons 2,3,4; Entrogen). On the whole, 81 patients were KRAS and NRAS WT (41 in Arm A and 40 in Arm B). Median PFS was respectively 7.8 months and 5.6 months in Arm A and Arm B (p ¼ 0.076); median OS was 21.0 months in Arm A vs 10.7 months in arm B (p ¼ 0.324). 73 were negative for the 3 genes (n ¼ 36 and 37). Their median PFS were 8.2 months in Arm A) vs 5.7 months in arm B (p ¼ 0.100). Median OS was 21.1 months vs 12.6 months (p ¼ 0.365). Conclusions: PRODIGE18 study is in favour of bevacizumab continuation beyond progression with chemotherapy cross over in WT RAS mCRC initially treated with first-line Bev plus chemotherapy. Legal entity responsible for the study: UNICANCER

doi:10.1093/annonc/mdx393 | 159

abstracts

Annals of Oncology

Funding: Roche Disclosure: J. Bennouna: Advisory Board for Roche, Boehringer Ingelheim, AstraZeneca, Servier, BMS. S. Hiret: Roche, Boehringer Ingelheim, AstraZeneca. C. Borg: Roche, Sanofi, Servier. O. Bouche: Roche, Merck, Amgen, Lilly, Pierre Fabre, Boehringer Ingelheim, Novartis. E. Francois: Advisory Board: Roche, Merck. F. Ghiringhelli: Roche, Sanofi, Amgen, BMS. J-F. Seitz: Roche, Merck, Sanofi. P. Artru: Roche, Merck, Amgen. A. Adenis: Roche. All other authors have declared no conflicts of interest.

478O

Efficacy and safety of Sym004 in refractory metastatic colorectal cancer with acquired resistance to anti-EGFR therapy: Results of a randomized phase II study (RP2S)

J. Tabernero1, F. Ciardiello2, C. Montagut3, C. Ding4, S. Kopetz5, T. Tuxen Poulsen6, A. Bardelli7, L. Wyrwicz8, A. Cubillo9, C. Santos10, G. Fumi11, V. Zagonel12, J. Bennouna13, S. Siena14, A. Falcone15, M. Benavent16, G. Argiles17, M. Kragh18, I.D. Horak6, M. Dvorkin19 1 Medical Oncology, Vall d’Hebron University Hospital and Institute of Oncology (VHIO), a di Napoli, Naples, Italy, 3University Hospital del Barcelona, Spain, 2Seconda Universit Mar, Barcelona, Spain, 4Symphogen, Somerville, NJ, USA, 5GI Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA, 6Symphogen, Ballerup, Denmark, 7Oncology, IRCCS University of Turin School of Medicine, Candiolo, Italy, 8Medical Oncology - GI Cancer Department, M. Sklodowska-Curie Memorial Cancer Center, Warsaw, Poland, 9Medical Oncology, Hospital Madrid Norte San Chinarro Centro Integral Oncologico Clara Campal, Madrid, Spain, 10Medical Oncology, Institut Catal a d’Oncologia Hospital Duran i Reynals, Barcelona, Spain, 11 Azienda Ospedaliera S. Maria, Terni, Italy, 12Istituto Oncologico Veneto IRCCS, Padua, 13 a degli Italy, CHU de Nantes, Nantes, France, 14Niguarda Cancer Center and Universit Studi di Milano, Milan, Italy, 15Oncology, University of Pisa, Pisa, Italy, 16Medical 17 Oncology, H. U. Virgen del Rocıo, Seville, Spain, Medical Oncology, Vall d’Hebron University Hospital, Barcelona, Spain, 18Symphogen, Denmark, 19Oncology, BHI of Omsk region Clinical Oncology Dispensary, Omsk, Russian Federation Background: Sym004, a mixture of 2 anti-EGFR monoclonal antibodies (mAbs), was shown to be active in a prior P1/2 trial in refractory mCRC. Due to its unique mode of action, Sym004 was developed for overcoming of acquired resistance to anti-EGFR antibodies. Methods: 254 patients (pts) were entered to an open label, multinational, 3-arm (1:1:1) RP2S comparing 2 regimens (12 mg/kg [A] or 9 mg/kg loading dose followed by 6 mg/ kg [9/6; B]) of weekly Sym004 vs investigator choice (IC) of 5-FU, capecitabine, or best supportive care [C]. Standard eligibility criteria were used; pts were to be refractory to chemotherapy and have responded to, and progressed on anti-EGFR mAb-based therapy; RAS exon 2 wild type in tumour archival sample. The study was designed to detect a 3-month (M) improvement in overall survival (OS) (6 vs 9 M) between either Arm A or B and Arm C. Results: Demographic and baseline parameters were well balanced. The Sym004 adverse event (AE) profile was typical although frequency/severity of dermatologic AEs and hypomagnesemia was higher and GI AEs appeared lower than with approved antiEGFR mAbs. Arm B was better tolerated than Arm A. OS in the ITT population and exploratory subgroups are presented. The primary outcome of the study was negative due to unexpected outcomes of Arm C. Arm B (9/6 mg dose) was not only better tolerated over 12/6 mg (Arm A), but also was associated with improved survival. Biomarker-specific analyses evaluating pts with double-negative (DN) (no RAS mutant allele frequency >20% in circulating tumor [ct]DNA; no BRAF V600E) or triplenegative (TN) (DN þ no EGFR extracellular domain mutation in ctDNA) mCRC demonstrated markedly prolonged survival and established the 9/6 regimen as welltolerated and active in DNmCRC (OS increased 3.5 M) and TNmCRC (OS increased 5.5 M).

Conclusions: Although the study was negative in ITT population, treatment with Sym004 was associated with remarkable response when compared with any 4th-line treatment of mCRC. The promising results in the molecularly selected population provide guidance to design a pivotal ctDNA-guided pivotal trial in EGFR inhibitor refractory mCRC. Clinical trial identification: NCT02083653 or EMR200637-002 Legal entity responsible for the study: Symphogen Funding: Symphogen Disclosure: J. Tabernero: Advisory boards: Amgen, Bayer, Boehringer Ingelheim, Celgene, Chugai, Genentech, Lilly, MSD, Merck Serono, Novartis, Pfizer, Roche, Sanofi, Symphogen, Taiho, Takeda. F. Ciardiello: Advisory boards: Roche, Merck, Lilly, BMS, Pfizer, Amgen, Bayer. C. Montagut: Advisory boards: Amgen, Bayer, Merck Serono, Sanofi, Symphogen. C. Ding, T. Tuxen Poulsen, M. Kragh, I.D. Horak: Employee of Symphogen. S. Kopetz: Advisory boards: Amgen, Merrimack, Bayer, Sanofi, Array BioPharma, Genentech, Molecular Match, Symphogen, Guardant Health, EMD Serono, Merck. V. Zagonel: Advisory boards: Celgene, Bayer, Roche, Amgen, Novartis, Pfizer. J. Bennouna: Honoraria: Roche, Boehringer Ingelheim, AstrZeneca, Shire, MSD, BMS; consulting or advisory role: Roche, Boehringer Ingelheim, AstraZeneca, Shire, MSD, BMS. S. Siena: Advisory boards: Amgen, Roche, Bayer, Merck-Serono, Sanofi, Merrimack. A. Falcone: Advisory boards and research grants to Institution: Amgen, Merck, Roche, Bayer, Servier, Lilly, Sanofi. All other authors have declared no conflicts of interest.

479O

Consensus molecular subtypes (cms) as predictors of benefit from bevacizumab in first line treatment of metastatic colorectal cancer: Retrospective analysis of the MAX clinical trial

J. Mooi1, P. Wirapati2, R. Asher3, C. Lee3, P. Savas4, T.J. Price5, S. Tejpar6, J. Mariadason1, N. Tebbutt7 1 Oncogenic Transcription Laboratory, Olivia Newton-John Cancer Research Institute, Heidelberg, Australia, 2Bioinformatics Core Facility, Swiss Institute of Bioinformatics, Lausanne, Switzerland, 3NHMRC Clinical Trials Centre, University of Sydney, Sydney, Australia, 4Medical Oncology, Peter MacCallum Cancer Center, Melbourne, Australia, 5 Medical Oncology, Queen Elizabeth Hospital and University of Adelaide, Woodville, Australia, 6Oncology, University Hospital Leuven, Leuven, Belgium, 7Medical Oncology, Olivia Newton-John Cancer Centre and Austin Health, Heidelberg, Australia Background: CMS is a transcriptome-based classification of colorectal cancer (CRC) with prognostic implications, but its association with treatment outcomes, especially in the metastatic setting, remains unknown. We investigated whether CMS classification was predictive of bevacizumab treatment benefit using data from the phase 3 MAX trial. MAX previously reported progression-free survival (PFS) benefit for the addition of bevacizumab (B) to chemotherapy (capecitabine (C) þ/- mitomycin (M)) in first line treatment of metastatic CRC. Methods: Archival tumours from 256 patients (54% of trial population) were available for gene expression profiling using Almac Xcel microarray. Tumours were classified into CMS groups 1 to 4 using previously published methods. We correlated CMS groups with PFS in the MAX trial. The predictive value of CMS was demonstrated as the interaction between CMS and bevacizumab treatment, assessed by Cox proportional hazards model. Results: After data quality control, primary tumours from 239 patients (51% of trial population) were suitable for survival analysis. Distribution of CMS groups were CMS1 18%, CMS2 48%, CMS3 12%, CMS4 23%. Hazard ratios (HR)(95% CI) of PFS in C vs CBþCBM arms for CMS 1,2,3 and 4 were 0.83 (0.43-1.62), 0.50 (0.33-0.76), 0.31 (0.13-0.75) and 1.24 (0.68-2.25) respectively (test for interaction between CMS and treatment, p ¼ 0.03). CMS remained a significant independent predictor of PFS after adjustment for prognostic factors in a multivariate analysis (p ¼ 0.04).

Table: 478O Population

Arm A

Arm B

Arm C

ITT N ¼ 254 US&EUc N ¼ 224 US&EU with biomarker data N ¼ 193 US&EU with DNmCRC N ¼ 170 (88%)d US&EU with TNmCRC N ¼ 131 (68%)d

7.9a (6.5, 9.9)b N ¼ 83 7.7 (6.1, 11.3) N ¼ 75 7.7 (5.5, 11.3) N ¼ 70 8.9 (6.2, 12.4) N ¼ 62 10.6 (6.8, 13.1) N ¼ 47

10.3 (9.0, 12.9) N ¼ 86 9.9 (8.0, 12.8) N ¼ 74 9.9 (7.1, 12.9) N ¼ 67 11.9 (9.7, 13.8) N ¼ 57 12.8 (9.7, 14.7) N ¼ 46

9.6 (8.3, 12.2) N ¼ 85 8.5 (6.8, 10.2) N ¼ 75 8.5 (6.4, 9.9) N ¼ 56) 8.4 (6.4, 10.0) N ¼ 51 7.3 (6.3, 8.8) N ¼ 38

a

median survival in M 95% conﬁdence intervals c the subgroup analyses excluded pts due to different medical practice d with biomarker data b

160 | Gastrointestinal tumours, colorectal

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Conclusions: In metastatic CRC, CMS 2 and 3 subtypes preferentially benefit from the addition of bevacizumab to chemotherapy, compared to CMS 1 and 4. Validation of these findings in independent cohorts is required. Once validated, CMS classification could be used to guide patient selection for bevacizumab therapy. Legal entity responsible for the study: Olivia Newton-John Cancer Research Institute, Australia Funding: Olivia Newton-John Cancer Research Institute, Australia and Australian Gastrointestinal Trials Group (AGITG) Disclosure: P. Wirapati: Funding: Roche, Bayer and Sanofi for research in predictive biomarker for colorectal cancer. T.J. Price, N. Tebbutt: Advisory boards: Roche All other authors have declared no conflicts of interest.

480O

Prognostic value of methylator phenotype in stage III colon cancer treated with oxaliplatin-based adjuvant chemotherapy

C. Gallois1, J. Taieb1, D. Le Corre2, K. Le Malicot3, J. Tabernero4, C. Mulot2, J-F. Seitz5, T. Aparicio6, G. Folprecht7, C. Lepage8, E. Mini9, J-L. Van Laethem10, J.F. Emile11, P. Laurent-Puig12 1 Department of Gastroenterology and Digestive Oncology, Hoˆpital Europe´en Georges Pompidou, Paris, France, 2Department of Biology, Paris Descartes University, Sorbonne Paris Cite´, France; Assistance Publique Hoˆpitaux de Paris, INSERM-UMR-S1147, Paris, France, 3Biostatistics, FFCD, Dijon, France, 4Medical Oncology, Vall d’Hebron University Hospital, Barcelona, Spain, 5Department of Gastroenterology and Oncology, CHU La Timone, Marseille, France, 6Gastroenterology, Saint Louis Hospital, Paris, France, 7 University Cancer Center, Medical Department I, University Hospital Carl Gustav Carus, Dresden, Germany, 8Gastroenterology, CHU Le Bocage, FFCD, Burgundy University, INSERM U866, Dijon, France, 9Department of Experimental and Clinical Medicine, University of Florence, Florence, Italy, 10Department of Gastroenterology, Hoˆpital Universitaire Erasme, Brussels, Belgium, 11Anatomie Cytologie Pathologiques, Hopital Ambroise Pare, Boulogne-Billancourt, France, 12Department of Biology, Paris Descartes University, Paris, France Background: There are conflicting results concerning the prognostic value of the methylator phenotype (CIMPþ for “CpG island methylator phenotype”) in non-metastatic colon cancer (CC) patients (pts). We studied this phenotype in stage III CC pts having undergone R0 resection, characterized for MSI, RAS and BRAF mutation status and treated with adjuvant FOLFOX-based treatment. Methods: Tumor samples of 1910 pts enrolled in the PETACC-8 adjuvant phase 3 trial were analysed. The method used was methylation-specific PCR where CIMPþ status was defined by methylation of at least three of the five following genes: IGF2, CACNA1G, NEUROG1, SOCS1 and RUNX3. Association between CIMP status and overall survival (OS), disease-free survival (DFS), and survival after relapse (SAR) was assessed by Cox model and adjusted for prognostic factors (including MSI, BRAF and RAS mutation status) and treatment arm (FOLFOX or FOLFOX plus cetuximab). CIMP status was analyzed according to treatment efficacy. Results: Determination of CIMP status was successful in 1870 pts (98%): 275 (14.7%) tumors were classified CIMPþ. Compared to CIMP- pts, CIMPþ pts were significantly older (p ¼ 0.002), with more frequently women (p ¼ 0.04). CIMPþ tumors were more frequently right-sided (p < 0.0001)), with histopathology grade 3-4 (p < 0.0001), pN2 (p ¼ 0.001), MSI (p < 10e-4), BRAF mutated (p < 0.0001) and RAS wild-type (p < 0.0001). In multivariate analysis, CIMPþ status was associated with shorter OS (HR: 1.4; 95% CI 1.02 – 1.9; p ¼ 0.04) and SAR (HR: 1.8; 95% CI 1.2 – 2.6; p < 0.0004); but DFS was not significantly different between CIMPþ and CIMP– pts (HR: 1.1; 95% CI 0.8 1.5; p ¼ 0.34). Theses results were independent of the treatment received. No benefit or detrimental effect of cetuximab was observed in CIMPþ patients for OS and DFS. Conclusions: In a large clinically and molecularly well defined stage III CC population treated with standard adjuvant therapy, methylator phenotype is a prognostic biomarker for OS and SAR. However, no impact of CIMP status on DFS was observed. Finally, we did not find any predictive value of the CIMP status for the efficacy of FOLFOX versus FOLFOX plus cetuximab. Clinical trial identification: PETACC8 Trial (EuDRACT number: 2005-003463-23) Legal entity responsible for the study: FFCD Funding: Merck, Sanofi Disclosure: J. Taieb: Consulting or/and advisory boards: Merck KGaA, Sanofi, Roche Genentech, Pfizer, Amgen. J. Tabernero: Consulting or/and advisory boards: Amgen, ImClone Systems, Lilly, Millennium, Novartis, Roche/Genentech, Sanofi, Celgene, Chugai Pharma, Taiho Pharmaceutical, Boehringer Ingelheim, Merck KGaA. J-F. Seitz: Grants for consultancy: Celgene, Lilly, Merck, Novartis Oncology, Pfizer, Sanofi, Roche; grants: Roche; payments for development of educational presentations: Amgen, Lilly; travel grants: Ipsen Pharma, Merck. T. Aparicio: Personal grants consultancy: Pierre Fabre; grants: Roche, Amgen; payments for development of educational presentations: Novartis Oncology, Pfizer, Sanofi, Roche; travel grants: Ipsen Pharma, Novartis Oncology, Sanofi, Roche. G. Folprecht: Consulting or/and advisory boards: Merck KGaA, Roche/Genentech, Sanofi-Aventis, Bayer, Lilly, Servier, BMS. C. Lepage: Personal grants for board membership: AAA; grants: Novartis; travel grants: Ipsen Pharma, Amgen, Bayer. J.F. Emile: Honoraria: Amgen, Merck KGaA. P. Laurent-Puig: Consulting or/and advisory boards: Sanofi, Merck KGaA, Amgen, Roche, Genomic Health, Myriad Genetics, Pfizer. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

481PD

Sidedness influences prognosis in stage III but not in stage II colon cancer patients receiving an adjuvant therapy: A GISCAD analysis from three randomized trials including 5234 patients

S. Cascinu1, D. Poli2, A. Zaniboni3, R. Labianca4, A. Sobrero5, V. Torri6 Department of Oncology/Hematology, Azienda Ospedaliero - Universitaria Policlinico di Modena, Modena, Italy, 2Laboratory of Methodology for Clinical Research, IRCCSIstituto Ricerche Farmacologiche Mario Negri, Milan, Italy, 3Oncology Unit, Fondazione Poliambulanza, Brescia, Italy, 4Department of Oncology, Ospedale Papa Giovanni XXIII, Bergamo, Italy, 5Medical Oncology Unit, IRCCS San Martino, Genoa, Italy, 6Department of Oncology, IRRCS Mario Negri Institute for Pharmacological Research, Milan, Italy

1

Background: While in the advanced setting right colon cancer is associated with a worse outcome, this negative prognostic effect has been not definitively demonstrated in the adjuvant setting. We have analyzed the outcome data from 3 large randomized trials (SITAC-1; SMAC and TOSCA) assessing adjuvant therapy in colon cancer patients with stage II and III. Furthermore, since previous trials were not powered to assess the prognostic role of transversum colon cancer, we analyzed this site independently. Methods: In order to define the prognostic effect of sidedness we assessed three randomized trials of adjuvant therapy (SITAC, 5FU/FA vs control, 821 patients; SMAC, intraportal 5FU vs 5FU/FA, 990 patients; TOSCA, FOLFOX or XELOX three vs six months 3513 patients) carried out in Italy from 1987 to 2013 and including 5324 patients. Survival and disease-free survival, overall and in each trial, were analyzed according to right, transversum and left colon location. Right-sided was considered caecum to hepatic flexure, left-sided splenic flexure to rectum and trasversum hepatic to splenic flexure. Statistical analysis considered all randomized patients according to allocation arm, with available data on putative prognostic factors. Analysis was planned in order to provide overall and by stage results. Results: 5324 patients were included in this analysis; 2490 patients were males and 2834 females. Median age was 64 years. 2240 patients had a stage II colon cancer and 3084 a stage III. Right tumors were 1573 (30%), trasversum 822 (15%) and left 2929 (55%). Patients characteristics were well balanced among the three trials. In all the 5234 patients DFS was not affected by tumor location (right colon versus left, HR ¼ 1.01; 95% CI ¼ 0.89-1.15) while right tumor was associated to a worse OS compared to left tumor (HR ¼ 1.21; 95% CI ¼ 1.05-1.40) In stage II patients there was no difference in terms of DFS and OS among the three different tumor location while in stage III patients, right colon cancer had a worse outcome both in DFS and OS than left tumor (HR: 1.37 95% CI ¼ 1.16-1.64, p < 0.001). Conclusions: This is the largest analysis demonstrating the prognostic effect of tumor location in colon cancer patients receiving adjuvant chemotherapy. The effect however is present only in stage III but not in stage II colon cancer. Legal entity responsible for the study: GISCAD Fundation Funding: None Disclosure: All authors have declared no conflicts of interest.

482PD

Robot-assisted vs laparoscopic vs open abdominoperineal resections for low rectal cancer: Short-term outcomes of a single-center prospective randomized controlled trial

J. Xu, Y. Wei, L. Ren, Q. Feng, J. Chen, D. Zhu, W. Chang, T. Yi, L. Yang, X. Qin General Surgery Department, Zhongshan Hospital, Fudan University, Shanghai, China Background: Currently, robotic surgery for rectal cancer using da Vinci System is common. However, there is almost no clinical trial reported. This randomized controlled trial aims to compare the safety and efficacy of robot-assisted, laparoscopic and open abdominoperineal resection (APR) for low rectal cancer. Methods: From 2013-09 to 2017-03, patients aged from 18 to 75 years, with low rectal cancer within 5 cm from anal verge, clinical T1 to T3, no distant metastases, were randomly assigned to receive either robot-assisted procedures (RAP), laparoscopic procedures (LAP) or open surgery (OS) for APR in 1:1:1 ratio. The primary endpoint was postoperative complication rate. Results: Totally 506 patients were enrolled in this study, randomly assigned to RAP (n ¼ 169), LAP (n ¼ 169), and OS (n ¼ 168). Actually, 3 patients refused surgery, 173 finished RAP, 176 finished LAP, and 154 finished OS (including 4 convert from LAP to OS). The open conversion rate was 0 in RAP and 2.4% in LAP, with no significant difference (P ¼ 0.123). In per-protocol analysis, no significant difference was observed in tumor location, size, differentiation and pathological TNM stage, among the three groups. RAP had significantly lower postoperative complication rate (10.4%) than both LAP (18.8%, P ¼ 0.027) and OS (26.0%, P < 0.001). Also, RAP reduced intraoperative hemorrhage (median, 100 ml) than LAP (130 ml, P < 0.001) and OS (200 ml, P < 0.001). And RAP promoted postoperative recovery, with shorter days to first flatus (1.0 day) than LAP (2.0 day, P < 0.001) and OS (3.0 day, P < 0.001), shorter days to first automatic urination (2.0 day) than LAP (3.0 day, P < 0.001) and OS (3.0 day, P < 0.001), and shorter days to discharge (5.0 days) than LAP (6.0 days, P < 0.001) and OS (6.0 day, P < 0.001). There was no significant difference in resection margin involvement and number of lymph node harvested. More details are shown in the table. Conclusions: Robot-assisted APR was safe, and reproduce equivalent surgical quality of conventional laparoscopic and open surgery. Also, it provided less injury and faster functional recovery.

doi:10.1093/annonc/mdx393 | 161

abstracts

Annals of Oncology

Table: 482PD Study results in per-protocol analysis

Operating time, min (median, IQR) Intraoperative hemorrhage, ml (median, IQR) Patients with perioperative transfusion, n (%) Lymph node harvested, n (median, IQR) Circumferential resection margin positive, n (%) Days to ﬁrst ﬂatus (median, IQR)# Days to ﬁrst automatic urination (median, IQR)# Days to discharge (median, IQR)# Postoperative mortality, n (%) Postoperative morbidity, n (%) Morbidity of Clavien-Dindo Grade III-IV, n (%)

RAP (n ¼ 173)

LAP (n ¼ 176)

OS (n ¼ 154)

P value RAP vs. LAP

P value RAP vs. OS

205 (200-220) 100 (90-110) 0 (0) 16 (13-20) 1 (0.6) 1.0 (1.0-2.0) 2.0 (2.0-3.0) 5.0 (5.0-5.0) 0 (0) 18 (10.4) 2 (1.2)

195 (160-240) 130 (100-150) 2 (1.1) 16 (12-19) 3 (1.7) 2.0 (2.0-3.0) 3.0 (2.0-4.0) 6.0 (6.0-7.0) 0 (0) 33 (18.8) 6 (3.4)

160 (140-180) 200 (120-220) 3 (1.9) 15.5 (13-19) 3 (1.9) 3.0 (2.0-4.0) 3.0 (2.0-4.3) 6.0 (5.0-7.0) 0 (0) 40 (26.0) 5 (3.2)

0.002 65 years.

Table: 539P Parameters

ECOG Primary location Metastatic sites Liver Lung Bone Peritoneum Lymph nodes N organ involved Stage at diagnosis CEA levels

CTCs> 3 p value

RASmut p value

BRAFmut p value

0.0069 –

0.002 –

– < 0.0001

< 0.0001 0.02 0.0002 – – 0.001 0.002 < 0.0001

0.01 0.005 0.002 – – 0.007 0.02 0.02

0.02* 0.002 0.004 0.005 – – –

*BRAF mutant less frequently associated to liver involvement

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Conclusions: CTCs and RAS mutation are significantly associated with other clinical poor prognostic factors. The poor prognosis of BRAF-mutated tumors reported in the literature cannot be explained by its correlation with poor prognostic clinical characteristics. Clinical trial identification: VISNÚ 1: NCT01640405 VISNÚ 2: NCT01640444 Legal entity responsible for the study: Spanish Cooperative Group for the Treatment of Digestive Tumors (TTD) Funding: Roche Farma SA Disclosure: J.M. Vie´itez: Consultant or advisory relationship and research funding: Roche. E. Aranda Aguilar: Honoraria for advisory role from Amgen, Bayer, Celgene, Mecrk, Roche, Sanofi. All other authors have declared no conflicts of interest. 540P

Analysis of liquid biopsies from metastatic colorectal carcinoma (mCRC) patients (pts) enrolled in the CAPRI GOIM clinical trial

N. Normanno1, R. Esposito Abate1, M. Lambiase1, L. Forgione1, C. Cardone2, A. Iannaccone1, A. Sacco1, A.M. Rachiglio1, E. Martinelli2, D. Rizzi3, S. Pisconti4, G. Cartenı5, R. Bordonaro6, T. Troiani2, F. Giuliani7, S. Leo8, S. Romito9, A. Rinaldi10, E. Maiello11, F. Ciardiello12 1 Cell Biology Unit, Istituto Nazionale Tumori – I.R.C.C.S - Fondazione Pascale, Naples, Italy, 2Oncologia Medica, Dipartimento di Internistica Clinica e Sperimentale “F. Magrassi”, Universit a degli Studi della Campania Luigi Vanvitelli, Naples, Italy, 3Trial ofﬁce, GOIM, Bari, Italy, 4Oncologia Medica, Ospedale S. Giuseppe Moscati, Statte, Italy, 5 Medical Oncology, Azienda Ospedaliera di Rilievo Nazionale “Antonio Cardarelli”AORN A. Cardarelli, Naples, Italy, 6Medical Oncology, Azienda Ospedaliera ARNAS Garibaldi, Catania, Italy, 7Medical Oncology, Istituto Tumori Giovanni Paolo II, Bari, Italy, 8Geriatric Oncology, Ospedale Vito Fazzi, Lecce, Italy, 9Oncology, Ospedale Riuniti di Foggia, Foggia, Italy, 10Oncologia, ASL Taranto, Castellaneta, Italy, 11Medical Oncology, IRCCS Casa Sollievo della Sofferenza, San Giovanni Rotondo, Italy, 12 Dipartimento Medico-Chirurgico di Internistica Clinica e Sperimentale “F. Magrassi E. A. Lanzara”, Universit a degli Studi della Campania “Luigi Vanvitelli”, Naples, Italy Background: Liquid biopsy can represent an alternative to tissue biopsy for biomarker testing in cancer pts. In addition, liquid biopsy can be used to monitor the response to treatment and the molecular evolution of the disease. Methods: In the CAPRI GOIM trial, KRAS exon 2 wild type (wt) mCRC pts received first line cetuximab plus FOLFIRI. Tumor samples were assessed by Next Generation Sequencing (NGS) with the Ion AmpliSeqTM Lung and Colon Cancer Panel (Thermofisher). Plasma samples at baseline (n ¼ 96), at 3 weeks of treatment (n ¼ 54), at 6 weeks (n ¼ 14) and at progression of disease (n ¼ 24) were collected from 96 patients and analyzed for exon 2, 3 and 4 KRAS and NRAS mutations using BEAMing Digital PCR (Sysmex Inostics). Results: Analysis of basal plasma samples from the 96 pts included in this study showed a concordance of 79,2% with the tissue RAS status as defined by NGS. The 11 cases that were RAS mutant (mut) in tissue and wt in plasma had suboptimal plasma volume available for analysis (1 mut genes 9 (14.3%), double mut 5 and triple mut 4, specifically double KRAS 1, KRAS/NRAS 2, KRAS/BRAF 1, NRAS/BRAF 1, double KRAS/NRAS 1, KRAS/ NRAS/BRAF 3. BRAF15 mut were all atypical and concomitant with KRAS and/or NRAS mutations. Prevalence of KRAS2-4, NRAS2-4, BRAF15 >1 mut samples were 19%, 53.8%, and 100% of each mut gene. At median follow-up 21 months (m), PFS and OS overall, and of KRAS2 genotype were consistent with previously reported; in c.35 G > A KRAS2 mut trendly worse PFS 8 m and OS 14m. Differential clinical outcome of MCRC patients wt and mut were not significantly different: KRAS2-4, PFS 13 and 12m, OS 27m equivalently; NRAS2-4, PFS 16 and 12m, OS 28 and 22m; BRAF15 PFS 14 and 8 m, OS 28 and 11 m; KRAS2-4/NRAS2-4/BRAF15 PFS 18 and 12m, OS 28 and 22m. Conclusions: Clinical outcome of MCRC patients treated with FIr-B/FOx is not significantly affected by KRAS2-4/NRAS2-4/BRAF15 genotype; efficacy may be increased in triple wt patients; the prevalent c.35 G > A KRAS2 and BRAF15 mut may show worse prognosis. Legal entity responsible for the study: Enrico Ricevuto Funding: None Disclosure: All authors have declared no conflicts of interest.

Prevalence of KRAS/NRAS/BRAF mutations detected by massive parallel sequencing and differential outcomes in MCRC patients (pts) treated with first line FIr-B/FOx adding bevacizumab (BEV) to triplet chemotherapy

G. Bruera1, F. Pepe2, U. Malapelle2, P. Pisapia2, A. Dal Mas3, G. Calvisi3, G. Troncone2, E. Ricevuto1 1 Oncology Territorial Care, Department of Biotechnological and Applied Clinical Sciences, S. Salvatore Hospital, ASL1 Abruzzo, University of L’Aquila, L’Aquila, Italy, 2 Department of Public Health, University Federico II, Naples, Italy, 3Pathology, S. Salvatore Hospital, ASL1 Abruzzo, L’Aquila, Italy Background: KRAS/NRAS/BRAF genotypes guide tailoring of first and subsequent lines of MCRC treatment strategy. First line triplet chemotherapy/BEV regimens significantly improved progression-free survival (PFS) and overall survival (OS) in MCRC patients. OS may be significantly worse in KRAS c.35 G > A and BRAF mutant (mut) MCRC. Prevalence and differential clinical outcome according to KRAS/NRAS/ BRAF genotype was evaluated in MCRC patients treated with FIr-B/FOx intensive regimen. Methods: Tumoral samples of 67 MCRC pts treated with FIr-B/FOx (77% overall) were analyzed through a 50 genes panel (PGM/Colon Lung Cancer) by ION Torrent. KRAS exons 2-4 (KRAS2-4), NRAS exons 2-4 (NRAS2-4), and BRAF exon 15 (BRAF15) were evaluated. Molecular diagnostic criteria for mutation detection: >500sequence coverage; >1% mutant allelic fraction. Clinical outcomes (PFS and OS) were evaluated and compared by log-rank. Results: KRAS2-4 mut were 42 (66.7%), 4 not evaluable; NRAS2-4mut 13 (19.4%); BRAF15 mut 5 (7.5%). KRAS2-4/NRAS2-4/BRAF15 mut MCRC patients were 49

Analysis of angiogenesis biomarkers for ramucirumab (RAM) efficacy in patients with metastatic colorectal cancer (mCRC) from RAISE, a global, randomized, double-blind, Phase 3 study

J. Tabernero1, R.R. Hozak2, T. Yoshino3, A.L. Cohn4, R. Obermannova5, G. Bodoky6, R. Garcia-Carbonero7, T-E. Ciuleanu8, D.C. Portnoy9, K. Muro10, H. Ouyang11, S. Melemed12, D. Ferry13, F. Nasroulah14, E. Van Cutsem15 1 Medical Oncology, Vall d’Hebron University Hospital and Institute of Oncology (VHIO), CIBERONC, Universitat Aut onoma de Barcelona, Barcelona, Spain, 2Late-Phase Oncology Tailoring Biomarker Statistics, Eli Lilly and Company, Indianapolis, IN, USA, 3 Gastroenterology and Gastrointestinal Oncology, National Cancer Center Hospital East, Kashiwa, Japan, 4Medical Oncology/Hematology, Rocky Mountain Cancer Center, Denver, CO, USA, 5Comprehensive Cancer Care, Masaryk Memorial Cancer Institute, an and Szent L aszl o Brno, Czech Republic, 6Clinical Oncology, Combined Szent Istv Hospitals, Budapest, Hungary, 7Medical Oncology, University Hospital 12 de Octubre, Madrid, Spain, 8Medical Oncology, Ion Chiricuta Oncology Institute-IOCN, Cluj-Napoca, Romania, 9Medical Oncology, West Clinic, Memphis, TN, USA, 10Clinical Oncology, Aichi Cancer Center Hospital, Nagoya, Japan, 11Medical Oncology, Eli Lilly and Company, Indianapolis, IN, USA, 12Oncology, Eli Lilly and Company, Indianapolis, IN, USA, 13 Oncology, Eli Lilly and Company, Bridgewater, NJ, USA, 14Medical Oncology, Eli Lilly and Company, Buenos Aires, Argentina, 15Digestive Oncology, University Hospital Gasthuisberg-University of Leuven, Leuven, Belgium Background: The RAISE trial (NCT01183780) demonstrated that RAM plus FOLFIRI (leucovorin, fluorouracil, and irinotecan) significantly improved overall survival (OS) and progression-free survival (PFS) compared with placebo plus FOLFIRI (PBO) as second-line mCRC treatment. Despite multiple approved anticancer treatments targeting angiogenesis, there are currently no predictive markers to guide patient selection. The extensive RAISE biomarker program assessed the association of multiple candidate biomarkers with RAM efficacy outcomes. Methods: Plasma and tumor tissue collection was mandatory in the RAISE trial. Analyses were performed using exploratory assays to assess the correlations of the baseline marker levels (vascular endothelial growth factor [VEGF]-C and D; soluble vascular endothelial growth factor receptor [sVEGFR]1, 2, and 3; and VEGFR2 immunohistochemistry in tumor tissue) with clinical outcomes. Cox regression analyses adjusted for stratification factors were performed for each marker. Results: Biomarker results were available from >80% of patients. Among the candidate biomarkers analyzed, only VEGF-D levels had a consistent and statistically significant association with OS and PFS, suggesting a predictive relationship. Higher levels were associated with improved RAM efficacy (Table). This relationship was consistent across the full range of VEGF-D levels. Conclusions: These analyses from RAISE identified VEGF-D as a potential predictive marker for RAM efficacy in mCRC. Further investigation of this relationship is being pursued. Clinical trial identification: NCT01183780

Table: 555P Correlation of VEGF-D with Efficacy Outcomes: based on cut point from exploratory subset. Results below from combined exploratory þ confirmatory groups OS Prespecified cut point Patients Median (months) (95% CI) HR (95% CI) p-value

115 pg/mL RAM N ¼ 270 PBO N ¼ 266 13.9 (12.5, 15.6) 11.5 (10.1, 12.4) 0.73 (0.60, 0.89) 0.0022

188 | Gastrointestinal tumours, colorectal

PFS 0.2ng/mL (after RP) or > 1ng/mL above

Table 783O Time to Progression and Relative Risk for HRQoL End Points Measure

ADTþAAþP (months)

ADTþPBOs (months)

Median time to worst pain intensity progression (BPI-SF) Median time to pain interference progression (BPI-SF) Median time to fatigue intensity progression (BFI) Median time to fatigue interference progression (BFI) Median time to FACT-P degradation Total Score Pain-related Subscale Prostate Cancer Subscale

NR NR NR NR 12.9 10.2 8.3

HR, hazard ratio; CI, conﬁdence interval.

C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

HR (95% CI)

p Value

16.6 18.4 NR NR

0.695 (0.583-0.829) 0.671 (0.561-0.803) 0.652 (0.527-0.805) 0.594 (0.470-0.750)

50%, including 11/30 (37%) with decline >80%. In 17 pt with soft tissue disease, objective response (RECIST PRþCR) occurred in 12 pt (71%). Most common adverse events were grade 1 xerostomia (19 pt, 63%) and nausea (15 pt, 50%). Grade 3 or higher hematoxicity occurred in 5 pt (17%); all had baseline thrombocytopenia and were reversible. Following the first cycle of LuPSMA, global health score improved significantly (10 points) in 11/30 pt (37%), while in those with bone pain, mean severity score improved significantly ( 10 points) in 9/21 pt (43%). Conclusions: The LuPSMA Phase II trial provides evidence of high response rates and low toxicity with improved QoL and pain reduction in men with mCRPC who have failed conventional therapies. Clinical trial identification: Australian New Zealand Clinical Trials Registry: ACTRN12615000912583. Universal Trial Number (UTN): U1111-1172-4095. Legal entity responsible for the study: Peter MacCallum Cancer Centre, Melbourne, Australia

270 | Genitourinary tumours, prostate

HR (95%) 0.85 (0.62-1.16) 1.01 (0.72-1.40) 1.16 (0.76-1.77)

p-value 0.31 0.95 0.49

Funding: Investigator-initiated trial. Lutetium-177 (no carrier added) supplied by Australian National Nuclear Research and Development Organisation (ANSTO). PSMA617 supplied by Advanced Biochemical Compounds (ABX). Disclosure: All authors have declared no conflicts of interest.

786PD

DNA repair gene panel mutations in young onset and aggressive vs non aggressive prostate cancer cases in the UK

R. Eeles1, E. Saunders1, S. Wakerell1, I. Whitmore1, C. Cieza-Borrella1, K. Govindasami1, T. Dadaev1, Z. Kote-Jarai1, D. Leongamornlert2 1 Genetics & Epidemiology, The Institute of Cancer Research, Sutton, UK, 2Cancer, Ageing & Somatic Mutations Programme, Wellcome Trust Sanger Institute, Cambridge, UK Background: Prostate cancer (PrCa) is the most common solid tumour in men in the Western world. There is evidence that PrCa predisposition is due to germline common and rare variation. Methods: We sequenced 175 genes in the DNA damage response and repair pathways using an Agilent custom capture kit and Illumina technology in PrCa cases diagnosed at < 65 years compared with controls in the UK (mean coverage 76X). Data were analysed from 1346 PrCa cases and 1186 controls using a GATK 2.8 analysis pipeline. Results: We identified 5,118 single nucleotide variants (SNVs) and 172 indels; 216 unique protein truncating variants (PTVs) were in 96 genes of the 175 gene panel. The total number of PTVs in cases was significantly higher (181) than in controls (122); in particular, in the BROCA gene set of 22 tumour suppressor genes (P ¼ 0.002). Mutations in BRCA1, BRCA2, ATM, MSH5 and CHEK2 were 3 times more common in cases compared with controls (P ¼ 0.0018). To investigate if aggressive cases had a different mutation burden we compared 204 aggressive (Gleason score>8) versus 1049 non-aggressive (Gleason score 7) cases. In the single variant analysis, one variant in BRCA2, rs28897754 (K2950N) showed association with a more aggressive phenotype (P ¼ 0.0016). Gene burden testing showed BRCA2, MSH2, PALB2 and CHEK2 had an OR > 3 in aggressive v non aggressive cases (14% v 4% respectively). Men who died of PrCa had a 17% incidence of mutation in a subset of the 175 gene panel. Conclusions: We have shown that there is a higher percentage of DNA damage response and repair gene germline mutations in PrCa cases occurring at < 65 years, in those with aggressive and lethal disease and this result will enable us to develop a testing panel for use in clinical care in the near future. Clinical trial identification: UKGPCS - CCR0848 & 06/MRE02/4 Legal entity responsible for the study: The Institute of Cancer Research Funding: None Disclosure: All authors have declared no conflicts of interest.

787PD

Prognostic associations of prostate-specific antigen (PSA) decline with survival, radiographic response and progression in chemotherapy-naı¨ve men with metastatic castration-resistant prostate cancer (mCRPC) treated with enzalutamide

A.J. Armstrong1, P. Lin2, C.S. Higano3, P. Iversen4, C.N. Sternberg5, B. Tombal6, D. Phung7, T. Parli8, A. Krivoshik9, T.M. Beer10 1 Divisions of Medical Oncology and Urology, Duke Cancer Institute, Durham, NC, USA, 2 Biometrics, Medivation, Inc., San Francisco, CA, USA, 3Medicine and Urology, University of Washington Seattle Cancer Care Alliance, Seattle, WA, USA, 4Urology, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark, 5Department of Oncology, San Camillo and Forlanini Hospitals, Rome, Italy, 6Urology, Cliniques Universitaires St. Luc, Brussels, Belgium, 7Biostatistics, Astellas Pharma Europe B.V., Leiden, Netherlands, 8 Clinical Development, Medivation, Inc., San Francisco, CA, USA, 9Medical Sciences Oncology, Astella Pharma, Inc., Northbrook, IL, USA, 10Oregon Health & Science University, OHSU Knight Cancer Institute, Portland, OR, USA Background: In the PREVAIL clinical trial, enzalutamide provided significant improvements vs placebo in radiographic progression-free survival (rPFS) and overall survival (OS) in chemotherapy-naı¨ve men with mCRPC. This post hoc analysis aimed to evaluate the prognostic association between the magnitude of PSA decline from baseline and clinical outcomes in PREVAIL. Methods: Men from the enzalutamide and placebo arms of PREVAIL were grouped into categories of confirmed maximal PSA decline from baseline at month 3 of

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology

Table: 787PD Outcome

Best objective soft-tissue response (CR or PR), % (95% CI) Median (95% CI) time to PSA progression, mo Median (95% CI) rPFS, mo HR (95% CI) for rPFS Median (95% CI) OS, mo HR (95% CI) for OS

Maximal PSA Decline From Baseline at Month 3 in the Enzalutamide Arm (N 5 872) No Decline/ Decline < 30% (n 5 94/872)

30% Decline (n 5 701/872)

50% Decline (n 5 639/872)

90% Decline (n 5 307/872)

12.0 (4.5-24.3) 3.7 (3.7-4.6) 7.9 (3.7-NYR) 1.0 (ref) 23.1 (17.8-28.0) 1.0 (ref)

70.6 (65.1-75.6) 13.8 (11.3-14.0) NYR (13.8-NYR) 0.20 (0.13-0.31) 32.4 (31.5-NYR) 0.31 (0.22-0.42)

74.8 (69.2-79.9) 13.9 (13.8-16.6) NYR (13.8-NYR) 0.17 (0.11-0.27) NYR (31.5-NYR) 0.28 (0.20-0.39)

89.7 (82.8-95.0) 22.5 (16.8-NYR) NYR (13.8-NYR) 0.10 (0.05-0.19) NYR (NYR-NYR) 0.19 (0.12-0.28)

Abbreviations: CI, conﬁdence interval; CR, complete response; HR, hazard ratio; mo, months; NYR, not yet reached; OS, overall survival; PR, partial response; PSA, prostate-speciﬁc antigen; ref, reference, rPFS, radiographic progression-free survival.

treatment: no decline/decline < 30% and 30%, 50% or 90% decline. Confirmation required PSA decline on 1 consecutive visit after month 3. Best overall soft-tissue response (per RECIST v1.1) was determined for patients with measurable disease at baseline (data cutoff: 16 Sep 2013). Time to PSA progression (data cutoff: 16 Sep 2013), rPFS (per PCWG2; data cutoff: 6 May 2012) and OS (data cutoff: 16 Sep 2013) were estimated using the Kaplan-Meier method. Results: In PREVAIL, men were randomized to enzalutamide (n ¼ 872) or placebo (n ¼ 845). Most men in the placebo arm (66%, 558/845) had no PSA decline/decline < 30%, in contrast to 11% (94/872) in the enzalutamide arm. In the enzalutamide arm, 81% (701/872) of men had a PSA decline of 30% from baseline at week 13, 73% (639/ 872) had a PSA decline of 50% and 35% (307/872) had a PSA decline of 90%. Key outcomes for the enzalutamide arm are provided by PSA decline category in the Table. PSA flare (rise followed by a fall) after 3 months was rare with enzalutamide (< 1%). Conclusions: PSA declines after 3 months of enzalutamide therapy are strongly associated with soft-tissue response and improvements in rPFS and OS. Providing updated prognostic information to chemotherapy-naı¨ve men with mCRPC can be of clinical value given the heterogeneity of long-term outcomes. Clinical trial identification: NCT01212991 Legal entity responsible for the study: This study was sponsored by Medivation, Inc. (which was was acquired by Pfizer, Inc. in September 2016) and Astellas Pharma, Inc., the co-developers of enzalutamide. Funding: This study was sponsored by Medivation, Inc., (which was acquired by Pfizer, Inc. in September 2016) and Astellas Pharma, Inc., the co-developers of enzalutamide. Disclosure: A.J. Armstrong: Consultant: Bayer, Sanofi, Novartis, Dendreon, Medivation, Janssen Biotech, Eisai Bureau: Dendreon, Sanofi, Medivation, Janssen Biotech Grant/Patent (inst) Dendron, Sanofi, Bayer, Pfizer, Novartis, BMS, Janssen Oncology, Medivation, Astellas, Gilead. P. Lin, T. Parli: Employment: Pfizer, Inc. C.S. Higano: Consulting/Travel: Dendreon, Bayer, Medivation, Ferring, J&J, AbbVie, Genentech, Pfizer, BHR, Orion, Sanofi, Amgen, Ockham, Teva, Astellas. P. Iversen: Consultant/Advisor, Meeting Participant/Lecturer, Scientific Study/Trial and Clinical Research Collaboration: Astellas Pharma, Medivation. C.N. Sternberg: Honoraria: Pfizer, Bristol-Myers Squibb, Novartis, Janssen, Bayer, Astellas Pharma, Sanofi, Eisai, Ipsen, GlaxoSmithKline, MSD. B. Tombal: Consulting: Astella, Bayer, Ferring, Janssen, Takeda, Steba Biotech, Sanofi Speakers Bureau: Amgen, Janssen Travel/Honoraria: Amgen, Astellas, Bayer, Ferring, Janssen, Sanofi. D. Phung: Employment: Astellas Pharma. A. Krivoshik: Employment and Travel/Expenses: Astellas Pharma Stock/ Ownership Interests: Abbott Laboratories, AbbVie. T.M. Beer: Consulting: Astellas, Bayer, Dendreon, Janssen Japan, Novartis, AstraZeneca, Churchill, Proacta Stock/ Ownership: Salarius.

788PD

Randomized controlled trial comparing radiotherapy 1/- endocrine therapy versus endocrine therapy alone for PSA failure after radical prostatectomy: Japan Clinical Oncology Group Study JCOG0401

A. Yokomizo1, T. Satoh2, K. Hashine3, T. Inoue4, K. Fujimoto5, S. Egawa6, T. Habuchi7, K. Kawashima8, O. Ishizuka9, N. Shinohara10, M. Sugimoto11, Y. Yoshino12, M. Wakabayashi13, K. Nihei14, H. Fukuda13, K-I. Tobisu15, Y. Kakehi11, S. Naito1 1 Urology, Harasanshin Hospital, Fukuoka, Japan, 2Urology, Kitasato University School of Medicine, Sagamihara, Japan, 3Urology, National Hospital Organization Shikoku Cancer Center, Matsuyama, Japan, 4Urology, Kyoto University Hospital, Kyoto, Japan,

Volume 28 | Supplement 5 | September 2017

5 Urology, Nara Medical University, Nara, Japan, 6Urology, Jikei University Hospital, Tokyo, Japan, 7Urology, Akita University Graduate School of Medicine, Akita, Japan, 8 Urology, Tochigi Cancer Center, Utsunomiya, Japan, 9Urology, Shinshu University School of Medicine, Matsumoto, Japan, 10Renal and Genitourinary Surgery, Hokkaido University Hospital, Sapporo, Japan, 11Urology, Faculty of Medicine, Kagawa University, Kagawa, Japan, 12Urology, Nagoya University Graduate School of Medicine, Nagoya, Japan, 13JCOG Data Center/Operations Ofﬁce, National Cancer Center, Tokyo, Japan, 14 Division of Radiation Oncology, Tokyo Metropolitan Cancer and Infectious diseases Center Komagome Hospital, Tokyo, Japan, 15Urology, Tokyo Metropolitan Cancer and Infectious disease Center Komagome Hospital, Tokyo, Japan

Background: A standard therapy has not been established on PSA failure after radical prostatectomy for localized prostate cancer. Therefore, the randomized controlled trial was designed to confirm the superiority of radiotherapy 6 endocrine therapy over endocrine therapy alone for PSA failure after radical prostatectomy. Methods: Patients were randomly assigned to arm A [endocrine therapy only: bicalutamide (BCL) monotherapy followed by LH-RH agonist in case of BCL failure], or arm B [64.8 Gy of salvage radiotherapy (SRT) followed by same regimen of arm A in case of treatment failure of SRT]. The primary endpoint is time to treatment failure (TTF) of BCL, and secondary endpoints are TTF of protocol treatment, clinical relapse-free survival (RFS), overall survival (OS), adverse events. The planned sample size was 210 to detect improvement of median TTF of BCL from 5 years to 8.3 years with one-sided alpha of 5% and power of 80%. This trial is registered with UMINCTR (C000000026). Results: A total of 210 patients (105 patients in each arm) were registered from May 2004 to May 2011. The TTF of BCL was significantly better in arm B as shown in Table 1 (Hazard ratio 0.56 90% CI (0.40–0.77); one-sided p ¼ 0.001). The 33 patients (32%) of 102 patients with SRT of arm B had no treatment failure of SRT, resulting in being free from hormonal therapy. In addition, TTF of protocol treatment was also significantly better in arm B. However, clinical RFS and OS were similar between the arms. Grade 4 adverse event was reported in one patient in arm B. Conclusions: The first SRT had advantage in both TTF of BCL and protocol treatment. Although the clinical outcomes of both arms of salvage therapy were similar with each other in terms of clinical PFS and OS, the SRT was effective in 32% of the patients, which contributed to avoiding the salvage endocrine therapy. Clinical trial identification: UMIN-CTR (C000000026) Legal entity responsible for the study: Japan Clinical Oncology Group, JCOG Funding: Ministry of Health, Labor and Welfare of Japan Disclosure: A. Yokomizo: Lecture fee from Astellas Pharma Inc., AstraZeneca K.K.and Takeda Pharmaceutical Company Limited. T. Satoh: lecture fee from Bayer AG, Astellas Pharma Inc., and AstraZeneca K.K. K. Hashine: lecture fee from Astellas Pharma Inc., Sanofi, Takeda Phrmaceutical Co.Ltd., Brystol Myers, AstraZeneca and Janssen Pharma. T. Inoue: lecture fee from Astellas Pharma Inc., AstraZeneca K.K.and Janssen Pharma K.K. K. Fujimoto: lecture fee from AstraZeneca K.K.and Takeda Pharmaceutical Company Limited S. Egawa: AstraZeneca, Astellas, Takeda Pharma Co. O. Ishizuka: Research grant from Astellas Pharma Inc., Takeda Pharmaceutical Company Limited, Pfizar Japan Inc. Lecture fee from Astellas Pharma Inc. N. Shinohara: Lecture fee from Bayer Co., Astellas Pharma Inc., Pfizer, and Novartis. .Advisary fee from Ono Pharma Inc. and Takeda Pharma. Co. S. Naito: Personal financial interest from Astellas Pharma Inc., Takeda Pharmaceutical Co., Ltd. and Green

doi:10.1093/annonc/mdx370 | 271

abstracts

Annals of Oncology

Table: 788PD

5-year TTF of BCL 5-year TTF of protocol treatment 5-year clinical RFS 5-year OS

Arm A (endocrine therapy only) (95% CI)

Arm B (radiation þ/- endocrine therapy) (95% CI)

Hazard Ratio (95% CI)

57.0% (46.7%–66.0%) 67.0% (56.9%–75.3%) 93.8% (86.8%–97.2%) 99.0% (93.4%–99.9%)

69.7% (59.6%–77.7%) 76.8% (67.1%–83.9%) 88.9% (80.9%–93.7%) 91.4% (84.2%–95.4%)

0.56 (0.38-0.82) 0.66 (0.44-1.00) 0.90 (0.45-1.81) 1.03 (0.46-2.29)

Peptide Pharmaceutical Co., Ltd. All other authors have declared no conflicts of interest.

789PD

Abiraterone acetate (AA) 1 prednisolone (P) for metastatic castration-resistant prostate cancer (mCRPC) with early progression or non-response to androgen deprivation therapy (ADT)

G. Arai1, M. Ogi2, K. Kobayashi3, N. Okuno4, T. Takahara2, K. Fukushima2, K. Yoshizawa2 1 Department of Urology, Dokkyo Medical University Koshigaya Hospital, Saitama, Japan, 2Medical Affairs Division, Janssen Pharmaceutical K.K., Tokyo, Japan, 3Urology department, Yokosuka Kyosai Hospital, Kanagawa, Japan, 4Department of Urology, Sagamihara Hospital, Kanagawa, Japan Background: mCRPC with early progression (1year) or non-response to initial ADT carries a poor prognosis, and there is no consensus regarding second-line therapy for these patients (pts) [ADT poor responders]. Although AAþP is effective for chemonaı¨ve mCRPC, limited data is available for ADT poor responders; thus, we conducted this study to evaluate the efficacy and safety of AAþP as secondary treatment for this population. Methods: This was a multicenter, open-label, single arm, 2-stage trial according to Simon’s minimax design [hypothesis: p0¼0.150, p1¼0.350, a ¼ 0.025, b ¼ 0.100], and 48 pts were required to efficacy analysis. Key eligibility: Chemo-naı¨ve mCRPC (testosterone level 150 ng/dL, no concomitant ADT and no evidence of metastases. Safety, immunogenicity and efficacy were evaluated in 4 treatment arms in 60 planned pts (A: 16, 2mg INO-5150; B: 15, 8.5 mg INO-5150; C: 15, 2mg INO-5150 þ 1mg INO-9012; D: 16, 8.5mg INO-5150 þ 1mg INO-9012) treated with 4 IM doses followed by electroporation on day 0, wks 3, 12 and 24 who were followed for a total of 72 Wks. Results: Median age, Gleason score and time since diagnosis were 69.5 yrs (range: 55.487.7), 7 (5-10) and 8.4 yrs (0.4-23.8) respectively. Of 61 evaluable pts, 38 (62%) had PSADT 12 mos and 23 (38%) had DT > 12 mos at Day 0. For pts with DT 12 mos, Day 0 and week 27 median DT were 6.0 (1.5, 11.6) mos and 8.1 (2.2, 100.0) mos respectively. Flow cytometry analysis revealed antigen specific upregulation of CD38 and Perforin on CD8þ T cells in 19/50 (38%) pts across the trial, with the greatest proportion in arm A, 8/14 (57%). Additional analysis for this cell subset showed a high PD-1 expression of 68.6% in this arm at week 27. Of note, in 8/15 (53%) arm A pts with DT 12 mos, their median DT at Day 0 was 6.2 (2.9, 10.2) mos and 19.2 (6.6, 100.0) mos at Wk 27. Safety: 7 Grade (Gr) 3 SAEs in 5 pts and 0 Gr 4-5 SAEs reported. Most AEs were Gr 1-3 in 51/62 (82%) pts and majority of those were associated with injection site reactions. Conclusions: INO-5150 þ/- INO-9012 was safe at dosages examined. Data demonstrated both PSA and PSMA are immunogenic and INO-5150 induced cellular immune responses. Higher proportion of arm A pts showed immunological responses as well as improvements in PSA DT, specifically pts with DT 12 mos suggesting correlation of immunological efficacy and clinical benefit. Continued analyses are planned as patient follow-up is ongoing. (NCT02514213) Clinical trial identification: NCT02514213; July 29, 2015 Legal entity responsible for the study: Inovio Pharmaceuticals Funding: Inovio Pharmaceuticals Disclosure: K. Bhatt: Inovio (study sponsor employee), own stocks M. Morrow, T. McMullan, K. Kraynyak, J. Lee, B. Sacchetta, L. Liu, S. Rosencranz: Inovio (study sponsor) employee, own stocks in company. Y. Whang: Research funding from Janssen, Astellas, Tokai, Inovio. I. Csiki, M. Bagarazzi: Employed by Inovio Pharmaceuticals. All other authors have declared no conflicts of interest.

791PD

Re-education of tumor-associated macrophages by CXCR2 blockade drives senescence enhancement and tumor inhibition in advanced prostate cancer

D. Di Mitri1, J. Vasilevska1, A. Calcinotto1, V. Gil2, G. Boysen2, A. Revandkar1, D. Waugh3, S. Barry4, J. de Bono5, A. Alimonti6 1 Molecular Oncology, IOR, Bellinzona, Switzerland, 2Cancer Biomarkers Team, Royal Marsden NHS Foundation Trust, Sutton, UK, 3Centre for Cancer Research and Cell Biology, Queen’s University Belfast - Centre for Cancer Research and Cell Biology, Belfast, UK, 4Li Ka Shing Centre, IMED Oncology AstraZeneca, Cambridge, UK, 5Division of Clinical Studies, Prostate Cancer Targeted Therapy Group, Institute of Cancer Research Royal Marsden Hospital, Sutton, UK, 6Molecular Oncology, IOR/IOSI, Bellinzona, Switzerland Background: Tumor-associated macrophages (TAMs) represent a major component of the tumor microenvironment that supports tumorigenesis. TAMs re-education instead that eradication has been recently proposed as a strategy to promote tumor inhibition.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Methods: We performed an immunophenotyping of Pten null prostate murine models by flow cytometry and gene expression analysis. Immunoistochemistry and immunofluorescence sainings were utilized to detect macrophages infiltration in the tumor and marker of proliferation and senescence in the tissue. Results: We have found that aggressive prostate tumors are strongly infiltrated by TAMs that express alternatively activated M2 markers. Unexpectedly chemokines binding to the C-X-C chemokine receptor type 2 (CXCR2) were among the most upregulated factors secreted by Pten null tumors and controlled the functional polarization of TAMs toward an “M2-like” functional status. Pharmacological blockade of the CXCR2 receptor in different tumor models in vivo promoted the re-education of TAMs toward a pro-inflammatory phenotype, which resulted in induction of senescence and tumor inhibition. Strikingly, infusions of CXCR2 knockout monocytes in Ptenpc-/-; Trp53pc-/mice demonstrated that inhibition of CXCR2 does not interfere with the tumor recruitment of monocytes but prevented the polarization of TAMs in M2-like resulting in an increased percentage of TNFa-releasing M1-like macrophages in the tumor microenvironment. Moreover, tumor cells harboring PTEN deletion were more sensitive to TNFa-induced senescence when compared to PTEN WT tumors due to increased levels of TNFR1. Conclusions: Taken together our results identify TAMs as a target for prostate cancer therapy and describe new therapeutic strategies to harness the anti-tumorigenic potential of macrophages in cancer. Legal entity responsible for the study: Institute of Oncology Research Funding: ERC/Steiner Disclosure: All authors have declared no conflicts of interest.

792PD

Phase I, open-label, dose-finding study of GSK2636771, a phosphoinositide 3-kinase (PI3K)b inhibitor, in combination with enzalutamide in male subjects with metastatic castration-resistant prostate cancer (mCRPC)

P. Rescigno1, J. de Bono2, A. Aparicio3, S. Chowdhury4, P. Twardowski5, N. Dawson6, U. Vaishampayan7, A.J. Pantuck8, Y. Zhou9, D. Fecteau10, G. Ganji11, J. Tolson12, D.A. Smith13, J. Medina14, L. Yan15 1 Division of Clinical Studies, Prostate Cancer Targeted Therapies, The Institute of Cancer Research and The Royal Marsden Hospital, Sutton, UK, 2Division of Clinical Studies, Prostate Cancer Targeted Therapy Group, Institute of Cancer Research Royal Marsden Hospital, Sutton, UK, 3Department of Genitourinary Medical Oncology, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX, USA, 4Medical Oncology, Guy’s and St. Thomas’ Hospital NHS Trust, London, UK, 5 Department of Medical Oncology and Experimental Therapeutics, City of Hope, Duarte, CA, USA, 6Prostate Cancer Research and Treatment Center, Georgetown Univerisity, Lombardi Comprehensive Cancer Center, Washington, DC, USA, 7Solid Tumor Oncology, Wayne State University, Karmanos Cancer Center, Detroit, MI, USA, 8 Department of Urology, Ronald Reagan UCLA Medical Center, Los Angeles, CA, USA, 9 Oncology - Biostatistics, GlaxoSmithKline USA, Collegeville, PA, USA, 10Oncology Clinical Development, GlaxoSmithKline USA, Collegeville, PA, USA, 11Biology, GlaxoSmithKline USA, King Of Prussia, PA, USA, 12Oncology - Clinical Development, GlaxoSmithKline, Rtp, NC, USA, 13Clinical Pharmacology, PAREXEL International, Durham, NC, USA, 14Oncology, GlaxoSmithKline, Collegeville, PA, USA, 15Oncology Clinical Development, GlaxoSmithKline USA, Collegeville, PA, USA Background: GSK2636771 is a potent, adenosine triphosphate (ATP) competitive, selective, oral inhibitor of the PI3Kb isoform that inhibits the growth of phosphatase and tensin homolog (PTEN) deficient tumor cells in preclinical models. Resistance to the R ) may be mediated through androgen receptor (AR) inhibitor enzalutamide (XtandiV upregulation of the PI3K pathway. Thus, GSK2636771 may enhance PTEN deleted prostate cancer cell kill. Methods: Within 30-days of documented progression on enzalutamide, subjects with PTEN deficient mCRPC are enrolled and treated with enzalutamide plus GSK2636771 in either dose escalation (DE) or dose expansion (DX) phases. Treatment continues until progressive disease (PD), unacceptable toxicities, consent withdrawal, or death. The primary objective is to assess safety, tolerability and determine the recommended phase 2 dose of this treatment combination. Secondary objectives include evaluation of pharmacokinetics, pharmacodynamics, biomarkers, and clinical activity per PCWG2/ RECIST 1.1. All subjects receive 160 mg enzalutamide once daily (QD) þ GSK2636771 starting at 300mg QD using a standard 3 þ 3 DE design with planned 100 mg incremental escalations or de-escalation. Results: In this ongoing study, 23 subjects received this treatment combination; 7 at 300 mg, 16 at 200 mg. Most AEs were Grade 1 or 2, with diarrhea the most common treatment-related AE (9/23; [39%]). Dose-limiting toxicities (DLTs) included Grade 3 hypocalcemia and reversible Grade 3 acute renal failure at 300mg and Grade 3 rash at 200mg. PK parameters suggested no drug-drug interaction between enzalutamide and GSK2636771. Among 13 evaluable patients at 200mg, 1 had a radiological partial response, and 2 had maximum PSA reductions of > 50%. Five subjects have been treated for 6 months. DE of 300 mg and DX of 200 mg cohorts are ongoing.

Volume 28 | Supplement 5 | September 2017

Conclusions: Our preliminary data indicate that GSK2636771 in combination with enzalutamide is largely well tolerated and confirm the clinical relevance of PI3Kb inhibition in PTEN-deficient mCRPC. GSK funds this study. Clinical trial identification: Protocol Number: 200331; NCT02215096; EudraCT No: 2013-005111-27 Legal entity responsible for the study: GlaxoSmithKline Funding: GlaxoSmithKline Disclosure: J. de Bono: Consulting or Advisory Role (Institution/Self) for the following companies: Astellas, AstraZeneca, Genentech, Genmab, GSK, Merck, Pfizer and Sanofi. Research support for the institution received from AstraZeneca, Genentech, GSK, Sanofi and Janssen. A. Aparicio: Self: GSK-Research/Advisor; Janssen/BMS/SanofiResearch; Genzyme-Honorarium Family Member: Boston Scientfic/Biosense WebsterResearch/Advisor/Speaker/Leadership/Travel; St. Jude-Advisor/Speaker/Honorarium/ Leadership; Hansen Medical-Research; Two Patents. S. Chowdhury: Self: GSK: Advisor/Speaker/Honorarium Sanofi: Research/Advisor/Speaker Clovis/Astellas/ Janssen: Advisor/Speaker Novartis: Speaker/Honorarium Bayer: Speaker Johnson and Johnson: Research. P. Twardowski: Self: Astellas, Medivation, Janssen: Speaker/ Honorarium Genentech: Advisor/Speaker Bayer, Sanofi: Speaker Roche: Advisor Exelxis: Stock Shareholder. N. Dawson: Self: Genentech, Astellas, Eisai, Janssen, Amgen, Sanofi, Bayer, Novartis: Speakers’ bureau. Abbvie, AstraZeneca: Advisor Honorarium received for speakers bureau and advisory board. Y. Zhou: Full or parttime employment by GSK. D. Fecteau: GSK: Full or part-time employment and stock holder. G. Ganji, J. Tolson, J. Medina, L. Yan: GSK: Full or part-time employment and stock shareholder. D.A. Smith: PAREXEL International: Full or part-time employment. GSK: Stock shareholder. All other authors have declared no conflicts of interest.

793PD

Phase 1 study of the PSMA-targeted small-molecule drug conjugate EC1169 in patients with metastatic castrate-resistant prostate cancer (mCRPC)

M. Morris1, N.J. Vogelzang2, O. Sartor3, A. Armour4, M. Groaning5, R. Messmann6, A. Robarts7, D.P. Petrylak8, A. Tolcher9, M.S. Gordon10, H. Babiker11 1 Genitourinary Oncology Service, Memorial Sloan-Kettering Cancer Center, New York, NY, USA, 2Medical Oncology, Comprehensive Cancer Centers of Nevada, Las Vegas, NV, USA, 3Oncology, Tulane University, New Orleans, LA, USA, 4Medical, Endocyte, Inc., Indianapolis, IN, USA, 5Clinical, Endocyte, Indianapolis, USA, 6Medical Oncology, Endocyte, West Lafayette, IN, USA, 7Genitourinary Oncology Service, Memorial Sloan Kettering Cancer Center, New York, NY, USA, 8Medical Oncology, Yale University School of Medicine Medical Oncology, New Haven, CT, USA, 9Medical Oncology, South Texas Accelerated Research Therapeutics (START), San Antonio, TX, USA, 10Medical Oncology, HonorHealth Research Institute, Scottsdale, AZ, USA, 11Medical Oncology, University of Arizona Comprehensive Cancer Center, Tucson, AZ, USA Background: Prostate-specific membrane antigen (PSMA) is highly expressed in prostate cancers, but not in most non-prostate normal tissues, making it a potential therapeutic target. EC1169, a PSMA-targeted conjugate of the microtubule inhibitor tubulysin B hydrazide is being studied in a two-part phase 1 dose escalation (A)/expansion (B) study in mCRPC. The utility of the PSMA-targeted companion imaging agent 99m Tc-EC0652 is also being evaluated as a patient selection tool. Part A has been completed. We now report the part B data on pts treated to date. Methods: EC1169 is administered as an IV bolus on days 1, 8 every 21 days. The RP2 dose of 6.5 mg/m2 was determined in Part A. Part B pts are treated at the RP2 dose and enrolled in 1 of 2 cohorts: mCRPC taxane naı¨ve (cohort 1, 45 pts) or taxane exposed (cohort 2, 40 pts). Prior to treatment, pts undergo a 99mTc-EC0652 SPECT/CT scan. The primary endpoint of Part B is median radiographic progression-free survival (rPFS). Other study evaluations are OS, PSA, and CTC-based biomarkers. Results: Thirty-four of a planned 85 pts in Part B have been treated (14 taxane naı¨ve, 20 taxane exposed). Median age is 70 (range: 49-84). Median number of cycles is 3 (range: 17). Twenty-six pts (76.5%) reported at least 1 treatment related AE. Most treatment related AEs (TRAEs) are Gr1 and 2; G3 treatment-related constipation occurred in 1 pt. No Grade 4 TRAEs have been reported. No dose reductions due to AEs have occurred. Six of twelve evaluable taxane-exposed pts in Part B had stable disease or better at their first post-baseline scan (9 wks). One pt currently beyond the 18-week scan has achieved durable resolution of his soft tissue disease. Imaging with 99mTc-EC0652 suggests excellent disease localization. Conclusions: The RP2 dose of EC1169 is 6.5 mg/m2 (D1, 8 every 21 days). EC1169 has been well tolerated in 34 Part B pts at the RP2 dose A PSMA-targeted therapeutic strategy appears viable. There is evidence of anti-tumor activity in both the taxane naı¨ve and taxane exposed pts. Clinical trial identification: NCT02202447 Legal entity responsible for the study: Endocyte, Inc. Funding: Endocyte, Inc Disclosure: A. Armour, M. Groaning: Employee of Endocyte, owns company stock. R. Messmann: Contractor for Endocyte, owns company stock. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx370 | 273

abstracts 794P

Annals of Oncology

EPI-506 (ralaniten acetate), a novel androgen receptor (AR) N-terminal domain (NTD) inhibitor, in men with metastatic castration-resistant prostate cancer (mCRPC): Phase 1 update on safety, tolerability, pharmacokinetics and efficacy

U. Vaishampayan1, R.B. Montgomery2, M.S. Gordon3, D.C. Smith4, K. Barber5, A. de Haas-Amatsaleh6, N. Thapar7, C. Chandhasin8, F. Perabo9, K.N. Chi10 1 Oncology, Karmanos Cancer Institute, Detroit, MI, USA, 2Genitourinary Medical Oncology, University of Washington, Seattle, USA, 3Medical Oncology, HonorHealth Research Institute, Scottsdale, AZ, USA, 4Medical Oncology, University of Michigan, Ann Arbor, MI, USA, 5Clinical Operations, ESSA Pharmaceuticals Corp., Houston, TX, USA, 6 Medical, ESSA Pharmaceuticals Corp., Houston, TX, USA, 7Clinical Pharmacology, ESSA Pharmaceuticals Corp., Houston, TX, USA, 8Medical & Scientiﬁc Affairs, ESSA Pharmaceuticals Corp., Houston, TX, USA, 9Medical, ESSA Pharmaceuticals Corp., Houston, TX, USA, 10Experimental Therapeutics, BC Cancer Agency, Vancouver, BC, Canada

McDonough Sarcoma oncogene (FMS) kinase, essential for the differentiation of osteoclasts, and is expected to be effective against bone metastases. In this expansion cohort of phase I trial, the safety and efficacy of TAS-115 were evaluated in CRPC pts with bone metastases. Methods: Eligible CRPC pts with bone metastases who were refractory to standard treatment including docetaxel, abiraterone and/or enzalutamide were enrolled. Two dose levels of TAS-115 (450 and 650 mg/body/day) were administered orally with a 5days-on/2-days-off schedule for up to 21 days per cycle in this expansion cohort. Efficacy was evaluated based on the RECIST ver 1.1 and bone scan response, defined as 30% decrease in bone scan index (BSI) calculated by quantitative software of V R (FUJIFILM RI Pharma, Japan; EXINI bone , Exini Diagnostics, BONENAVIV Sweden). Toxicities were evaluated based on the CTCAE ver 4.03. Results: As of Apr 2017, a total of 15 pts received TAS-115 (9 pts with 450 mg, and 6 pts with 650 mg). Bone scan response was reported in 4 of 9 pts (44.4%) who had baseline BSI 0.5%, which is equivalent to grade1 extent of disease. The best overall response per RECIST was stable disease in 7 of 15 pts (46.7%). These efficacies were observed regardless to dose levels. One patient had a long administration period exceeded to 15 months without disease progression, and another one patient experienced remarkable pain relief induced by bone metastases. TAS-115 had no effect on the PSA and ALP. The major (30%) adverse drug reactions (ADRs) were anorexia, fatigue, nausea, thrombocytopenia, rash, AST increased, anemia, vomiting and edema. The rate of grade 3 to all ADRs was 18.8%. These AEs were recovering by interruption of TAS115. Conclusions: Toxicities of TAS-115 were acceptable and manageable in CRPC pts, and preliminary anti-tumor activity, especially against bone metastases was recognized. A phase II trial for CRPC pts with bone metastases is ongoing. Clinical trial identification: JapicCTI-132333/163448 Legal entity responsible for the study: Taiho Phermaceutical CO., LTD. Funding: Taiho Phermaceutical CO., LTD. Disclosure: N. Matsubara: Honoraria: Taiho Pharmaceutical. Consulting or Advisory role: Janssen, AstraZeneca. Speakers’ bureau: Janssen, AstraZeneca, Sanofi. Research funding: Janssen, Bayer. Y. Naito: Speakers’ bureau: Eisai, Chugai, Taiho, Novartis, Eli Lilly, Meiji Seika, Bayer, Roche. Research Funding: Merck Serono, AstraZeneca, Eli Lilly, Nippon Kayaku. M. Sasaki: Speakers’ bureau: Taiho. Research funding: Eisai, Bayer. N. Yamamoto: Honoraria: AstraZeneca, Pfizer, Chugai, Bristol-Myers, Ono, Eli Lilly. Research funding: Chugai, Taiho, Eisai, Quintiles, Astellas, Novartis, Daiichi Sankyo, Eli Lilly, Boehringer Ingelheim, Takeda, Kyowa Hakko, Bayer, Pfizer. S. Takahashi: Honoraria: Eisai, Astellas, Taiho, Bayer, Daiichi, Pfizer, AstraZeneca, Merck, Sanofi, Novartis, Ono, Bristol, MSD. Consulting or Advisory role: Chugai, Astellas, Bayer, Pfizer. Research funding: Eisai, Chugai, MSD, Lilly, AstraZeneca, Novartis, Taiho, Bayer, Astellas, Ono, Daiichi. H. Uemura: Consulting or Advisory role: Janssen, Takeda. Speakers’ bureau: Janssen, Takeda, Astellas, Bayer, Sanofi, AstraZeneca, Fujifilm. Research funding: Astellas. T. Doi: Consulting or Advisory Role: Lilly, Chugai, Kyowa Hakko Kirin, Novartis, MSD, Daiichi Sankyo, Amgen. Research funding: Taiho, Merck, Astellas, Janssen, Takeda, Pfizer, Lilly, Sumitomo Group, Bayer, Chugai, Kyowa Hakko Kirin, Boehringer, Novartis, MSD, Daiichi Sankyo, Celgene. R

Background: EPI-506 (ralaniten acetate) is being studied in a Phase 1/2 study as a firstin-class transcription inhibitor of the AR NTD. Methods: Open-label, single-arm, Phase 1/2 study evaluating EPI-506 administered orally. The Phase 1 is a modified 3 þ 3 design to establish safety, tolerability, pharmacokinetics (PK), maximum-tolerated-dose (MTD) and the recommended phase 2 dose (RP2D) of EPI-506. Anti-tumor activity is evaluated by PSA and imaging. Inclusion criteria include: mCRPC with progression after > ¼1 line of hormonal therapy or chemotherapy, failure to treatment with enzalutamide and/or abiraterone. Results: Twenty-one patients (pts) have been enrolled in the dose escalation phase over 6 dose levels (80 - 2,400 mg). Median age was 72 (range 58-87). Prior treatments included enzalutamide only (N ¼ 9), abiraterone only (N ¼ 3) or both (N ¼ 9). Eight pts also had prior chemotherapy. Twelve pts discontinued due to disease progression and 2 pts due to adverse events (AEs): Grade 4 elevated amylase (probably related; at 640mg) and Grade 4 gastrointestinal bleeding (unrelated). Median exposure was 87 days at cut-off (range 21-418). Most frequently reported treatment emergent AEs were diarrhea (N ¼ 8), nausea (N ¼ 6), and pain in extremities (N ¼ 5). One possibly related Grade 3 AE (AST elevation) was observed at 1280 mg. PK data demonstrate a doseproportional profile for Cmax and AUC together with a positive food effect above 640 mg. Three of 21 evaluable pts demonstrated PSA declines ranging from 4 – 29%, and one pt with unchanged PSA at doses >1,280 mg. Three pts have had prolonged treatment (median of 286 days; range 219 – not reached), after intrapatient dose escalation. The study is currently enrolling pts with a total dose of 3,600 mg in both a QD and a BID dosing schedule. Conclusions: EPI-506 is well-tolerated with an acceptable safety profile. PK indicates dose-proportionality. PSA declines and stable disease have been observed at higher dose cohorts in this ongoing study. This study is the first to evaluate targeting the AR NTD, a region critical for transcriptional function of all known AR species. Clinical trial identification: NCT02606123 Legal entity responsible for the study: ESSA Pharmaceuticals Funding: ESSA Pharmaceuticals Disclosure: U. Vaishampayan, M.S. Gordon, D.C. Smith: ESSA Pharmaceuticals Corp. (Research funding) R.B. Montgomery, K.N. Chi: ESSA Pharmaceuticals Corp. (Scientific advisory board; Honoraria received; Research funding). K. Barber, F. Perabo, N. Thapar, C. Chandhasin: ESSA Pharmaceuticals Corp. (Employed, Ownership interest). A. de Haas-Amatsaleh: ESSA Pharmaceuticals Corp. (Consultant).

797P

Phase II trial of SM88 in non-metastatic biochemical recurrent prostate cancer

G. Del Priore1, W-T. Chen2, H. Dong2, S. Hoffman3, G. Sokol4 Tyme Inc, New York, NY, USA, 2School of Medicine, Stony Brook University, Stony Brook, NY, USA, 3Tyme Inc, New York, NY, USA, 4Tyme Inc, Uniformed Services University of the Health Sciences, Bethesda, MD, USA

1

796P

Phase I expansion cohort of TAS-115, a novel oral MET/VEGFR/FMS inhibitor, for castration-resistant prostate cancer patients (CRPC pts) with bone metastases 1

1

2

3

4

5

N. Matsubara , Y. Naito , M. Sasaki , N. Yamamoto , S. Takahashi , H. Uemura , T. Doi Department of Breast and Medical Oncology, National Cancer Center Hospital East, Kashiwa, Japan, 2Department of Medical Oncology, Kyoundo Hospital, Tokyo, Japan, 3 Department of Thoracic Oncology, National Cancer Center Hospital, Tokyo, Japan, 4 Department of Medical Oncology, Cancer Institute Hospital of JFCR, Tokyo, Japan, 5 Department of Urology/Renal Transplantation, Yokohama City University Medical Center, Yokohama, Japan, 6Department of Experimental Therapeutics, National Cancer Center Hospital East, Kashiwa, Japan

6

1

Background: TAS-115 is a novel small-molecule inhibitor of hepatocyte growth factor receptor (MET) and vascular endothelial growth factor receptor (VEGFR) with antitumor activity in preclinical models. It has also been shown that TAS-115 inhibits Feline

274 | Genitourinary tumours, prostate

Background: Despite toxicity and no clear clinical benefit, non-metastatic recurrent prostate cancer (nmPC) is typically treated with medical castration in North America. SM88 is a non-toxic novel combination therapy based on the Warburg effect, with activity in a variety of cancers including prostate (JCO 2017 e Abstract1). End of phase 1 results demonstrated stable or rising testosterone levels while achieving CTC (circulating tumor cells) benefit and no radiographic progression events (JCO 2017e Abstract2). We now report phase II data. Methods: Starting in Sept 2016, a prospective Phase Ib/II of SM88 (230mg po bid) enrolled recurrent nmPC with rising PSA (PCWG3 definition) and detectable CTCs, but no radiographically identified lesions. Results: 8 (of 34 planned) subjects have completed at least 1 cycle (median 5, range 17). Mean age was 69.7 (62-80); all had prior ADT after curative intent RT (50%) or surgery (50%); no patient is currently on ADT. Mean testosterone level (T) was 581.4 ng/

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology dL and rose or remained stable in the subjects except for one patient who entered the trial castrate (30% (n ¼ 4); at up to 6 cycles, no PSA progression (PCWG3) and no radiographic progression was reported (n ¼ 8). No subject required other toxic therapy (100% subsequent treatment free survival). Available preliminary neutrophile:lymphocyte ratio (N::L)(n ¼ 6) improved while urinary NTx, bone specific AlkPhos and LDH trends were essentially unchanged.

Table: 797P subject #

cycles completed

T ng/dL

CTCs baseline

Max Decrease

N:L Max Decrease

1 2 3 4 5 6

6 6 6 4 3 1

635.7 25% decrease in PSA occurred in 65% of pts and further decrease (>25%, compared to the nadir during AAþP treatment) has been seen in 26% pts during AAþD treatment.

Conclusions: D can induce further response during AA therapy by reversing glucocorticoid receptor activation or by superior activity of D administered even as a single agent. Our data supports that steroid switch may induce further PSA regression. Legal entity responsible for the study: Fruzsina Gyergyay Funding: None Disclosure: All authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx370 | 275

abstracts 800P

Annals of Oncology

Phase II study of prednisone-dexamethasone switch in metastatic castration resistant prostate cancer (mCRPC) patients treated with abiraterone and prednisone (AA1P)

R. Lozano Mejorada1, N. Romero Laorden2, A. Jayaram3, F. L opez4, M.I. Saez5, on9, J. Rogado10, R. Villatoro6, A. Montesa7, I. Moreno5, M. Ruiz Vico8, M. Garcıa Ferr Y. Cend on Fl orez11, P. Nombela Blanco2, L. Rivera2, G. Grau7, J.J. Cruz Hernandez1, D. Lorente Estelles12, G. Attard13, E. Castro Marcos2, D. Olmos Hidalgo14 1 Medical Oncology, Hospital Universitario de Salamanca, Salamanca, Spain, 2Prostate Cancer Clinical Research Unit, Spanish National Cancer Research Centre, Madrid, Spain, 3GU Unit, The Institute of Cancer Research (ICR), London, UK, 4Radiation Oncology, Hospital Universitario Ramon y Cajal, Madrid, Spain, 5Medical Oncology, H Universitario Virgen de la Victoria y Regional de M alaga, Malaga, Spain, 6Medical Oncology, Hospital Costa del Sol, M alaga, Spain, 7CNIO-IBIMA Genitourinary Research Unit, Hospitales Universitarios Virgen de la Victoria y Regional de M alaga., M alaga, alaga, Spain, 9Medical Spain, 8Medical Oncology, Hospital Universitario Carlos Haya, M 10 Oncology, Hospital Universitario Fundaci on Alcorc on, Madrid, Spain, Medical Oncology, Hospital Universitario de La Princesa, Madrid, Spain, 11Prostate Cancer Unit, CNIO- Spanish National Cancer Center, Madrid, Spain, 12Medical Oncology Department, Hospital Universitari i Polite`cnic La Fe, Valencia, Spain, 13Treatment Resistance, Division of Molecular Pathology, The Institute of Cancer Research (Sutton Site), Sutton, UK, 14Prostate Cancer Clinical Cancer Research Unit, Spanish National Cancer Research Centre (CNIO), Madrid, Spain Background: Abiraterone Acetate (AA) improves overall survival in mCRPC patients. It is administered with prednisone (P) to decrease the adverse events derived from CYP171A supression. In phase I/II of AA without steroids, dexamethasone (D) 0.5mg/ day was added after biochemical progression reaching a 25% of PSA decline. In a retrospective post-docetaxel cohort, Lorente et al (BJC,2014) reported that the switch induced durable biochemical responses in 40% of cases. Methods: This is a multicentre-prospective phase II study. Its primary aim was to evaluate the antitumour activity of the change of concomitant P 5mg/12h to D 0.5mg/ 24h daily in mCRPC patients with biochemical and/or limited radiological progression after 12 weeks of AAþP treatment. Biochemical response and progression free survival (bPFS) were evaluated by PCWG2 criteria. Radiological response and PFS (rPFS) were evaluated after 12 weeks by RECIST and PCWG2 criteria. Using a single-stage AHern Phase II design a minimum of 6 PSA responses >30% in 25 enrolled patients were required to accept the alternative hypothesis (a:5%, 1-b:80%). PTEN and TMPRSS2-ERG in archival tumour-biopsies, AR aberrations in ctDNA and AR-V7 in exosomal RNA were evaluated. The Kaplan-Meier curves were used to calculate survival outcomes. Results: 26 patients were included. Their clinical characteristics are shown in Table 1. No new safety concerns were observed with AAþD. A decline in PSA30% and 50% were observed in 12 (46%) and 8 (35%) patients, respectively; two radiological responses were observed; bPFS and rPFS after P to D switch were 4.2 months (CI 95% 2.2-6.2) and 11.8 months (CI 95% 6.9-16.8), respectively.

Table: 800P Characteristics

AA1D pre-CT (n 5 14) N %

AA1D post-CT N %

Age Median (range) Baseline PSA Median (range) Gleason 6-7 8-10 UK ECOG 0-1 2 Metastases Bone Nodes Visceral

72.9 60-85 26.6 4.5-367 5 36 8 57 1 7 13 80 1 7 12 86 8 57 1 7

72.9 66-78 39.9 6.9-1880 6 50 6 50 12 100 12 100 4 33 3 25

Conclusions: In selected clinically stable mCRPC patients the P to D switch as adjuvant of AA could be an acceptable and active therapeutic option. Biomarkers correlation with P to D switch benefit will be reported. Clinical trial identification: NCT02928432 Legal entity responsible for the study: Spanish National Cancer Research Centre (CNIO) Funding: Spanish National Cancer Research Centre (CNIO) Disclosure: D. Lorente Estelles: Speaker fees and advisory boards: Janssen. All other authors have declared no conflicts of interest.

801P

Effect of sequence on outcome of prostate cancer patients: retrospective study of a French cohort

R. Elaidi1, Y. Belhadj1, S. Oudard2 Medical Oncology - Hoˆpital Europe´en Georges Pompidou, ARTIC, Paris, France, 2 Immunothe´rapie et Traitement Antiangioge´nique en Pathologie Cance´rologique, Hopital European George Pompidou, Paris, France

1

276 | Genitourinary tumours, prostate

Background: several drugs are approved in prostate cancer (PC), both in localized and metastatic setting. Challenge of daily practice is the sequencing of available agents for optimal disease management. Trying to extract actionable information from the overall history of disease for each patient remains a difficult task but could provide new insights for better sequencing. This retrospective analysis aimed to follow-up patients included in the Rising-PSA phase 3 clinical trial (R-PSA-CP-03) until death or last contact. Methods: we retrospectively analyzed therapies received by pts included in R-PSA at the HEGP hospital (Paris, France). Drugs were coded in 8 categories: LH: LHRH modulators, AA: anti-androgens, AA2: new generation AA, DC: docetaxel, CZ: cabazitaxel, EX: blinded experimental drugs, P: therapeutic pause, PCB: placebo(experimental). Sequence rank, therapy duration and their interaction was estimated using both a conditional repeated events model (CREM) and a multi-state model (MSM) based on Markov process stratified on disease setting. Covariables included in the models were: age and Gleason score at inclusion time. Results: 152 pts included between 01/2003 and 09/2007 were followed > 10years. Metastatic progression: 70(46%). Death: 31(20%). Median age(y): 64(51-80)), Gleason 8: 47(31%). Median (range) number of sequences received: M0¼8(1-15) & M1¼6(119) including pauses. Number of times each therapy was used whatever the sequence (%M0/%M1): LH(48/10), AA(6/6), AA2(0/22), DC(10/10), CZ (0/10), EX(1/6), PCB(0/2), P(35/34). Upon CREM, the overall model fitted perfectly well the time on therapies and their sequence (robust estimation: p < 0,00001). Main effects were mostly related to docetaxel, cabazitaxel and experimental drugs in the metastatic setting. Effect of sequence was significant but no therapy x sequence interaction proved to be significant. Comparable results were obtained upon MSM and will be presented. Conclusions: to our knowledge, this is the first attempt to model the entire course of PC taking into account both therapies and sequence. Given the complexity of our model, these results should be validated with further studies and methods. Clinical trial identification: R-PSA-CP-03 Legal entity responsible for the study: ARTIC Funding: None Disclosure: All authors have declared no conflicts of interest.

802P

Patient preference between Cabazitaxel and Docetaxel for first-line chemotherapy in metastatic castrate-resistant prostate cancer (mCRPC): Results from the CABADOC randomized trial

K. Fizazi1, R. Delva2, G. Gravis3, G. Baciarello4, C. Theodore5, M. Gross-Goupil6, E. Bompas7, F. Joly Lobbedez8, Y. Tazi9, T. L’Haridon10, T. Nguyen Tan Hon11, P. Barthelemy12, S. Culine13, J-F. Berdah14, M. Deblock15, P. Beuzeboc16, A. Fle´chon17, C. Cheneau18, G. Martineau19, I. Borget20 1 Cancer Medicine, Gustave Roussy, Villejuif, France, 2Medical Oncology, Centre Paul Papin, Angers, France, 3Medical Oncology, Institute Paoli Calmettes, Marseille, France, 4 Cancer Medicine, Gustave Roussy Institut de Cance´rologie, Villejuif, France, 5Medical Oncology, Hopital Foch, Suresnes, France, 6Medical Oncology, CHU Bordeaux Hopital St. Andre´, Bordeaux, France, 7Medical Oncology, ICO Institut de Cancerologie de l’Ouest Rene´ Gauducheau, Saint-Herblain, France, 8Medical Oncology, Centre Francois Baclesse, Caen, France, 9Medical Oncology, Strasbourg Oncologie Libe´rale, Strasbourg, France, 10Medical Oncology, CHD Vendee - Hopital Les Oudairies, La Roche Sur Yon, France, 11Medical Oncology, CHU Besanc¸on, Hoˆpital Jean Minjoz, Besanc¸on, France, 12 Medical Oncology, C.H.U. Strasbourg-Nouvel Hopital Civil, Strasbourg, France, 13 Medical Oncology, Hoˆpital St. Louis, Paris, France, 14Medical Oncology, Hopital prive´ Hyeres, Hyeres, France, 15Medical Oncology, Institut de Cance´rologie de Lorraine - Alexis Vautrin, Vandoeuvre Les Nancy, France, 16Medical Oncology, Institut Curie, Paris, France, 17Oncology, Centre Le´on Be´rard, Lyon, France, 18Medical Oncology, CH Sud Bretagne, Lorient, France, 19Clinical reasearch, Gustave Roussy, Villejuif, France, 20 Statistics, Institut Gustave Roussy, Villejuif, France Background: Docetaxel and cabazitaxel are taxane chemotherapy approved in men with mCRPC after they demonstrated improved overall survival in first- and secondline, respectively. Recent data suggested similar efficacy when used in the first-line setting (Sartor O, ASCO 2016). These two taxanes have different safety profiles. Assessing patient preference between docetaxel and cabazitaxel would contribute to further differentiate between these two agents. Methods: The CABADOC study is a randomized trial with a cross-over design. Patients with mCRPC were randomized in a 1:1 ratio to receive either docetaxel 75mg/m2/q3w x 4 followed by cabazitaxel 25mg/m2/q3w x 4, or the reverse sequence. Randomisation was stratified based on prior abiraterone or enzalutamide. The primary endpoint was patient preference between taxanes, assessed in patients who had received at least one cycle of each taxane and who had not experienced a progression after the first taxane. Prescott’s test was used to analyze the primary endpoint taking into consideration the period effects. Results: From June 2014 to October 2016, 195 patients were randomized in 17 centers. The median age was 70 years and the median PSA was 49 ng/mL. Patients received 3.8 6 0.7 and 3.2 6 1.5 cycles of chemotherapy during the first and the second period, respectively. The eligible population for the primary endpoint comprised 150 patients (45 patients were ineligible for the primary endpoint as per protocol). Among them, 66 preferred cabazitaxel (44% IC95% ¼ [36-52]), 40 preferred docetaxel (27% IC95% ¼ [20-34]), and 44 expressed no preference between taxanes (29% IC95% ¼ [22-37]) (p ¼ 0.009). A greater proportion of patients preferred the first received taxane

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology (44%, IC95% ¼ [36-52]) versus the second taxane (27%, IC95% ¼ [20-34]), or had no preference (29% IC95% ¼ [22-37]). Less fatigue and improved quality of life were the two main reasons provided by patients for their choice. There were 3 toxicity-related deaths (1.5%). Pharmaco-economic analysis, toxicity, and quality of life data will be presented. Conclusions: A higher proportion of men with mCRPC who are candidates to receive a taxane prefer cabazitaxel over docetaxel. Clinical trial identification: NCT02044354 Legal entity responsible for the study: Gustave Roussy Funding: Sanofi Disclosure: K. Fizazi: Advisory boards/honorarium for Amgen, Astellas, AstraZeneca, Bayer, Clovis, Curevac, Essa, Genentech, Janssen, Orion, Sanofi. G. Gravis: Travels supported by Astellas, Janssen and Sanofi. M. Gross-Goupil: Advisory boards/honorarium for Amgen, Astellas, Janssen, MSD, Sanofi. A. Fle´chon: Honorarium from Astellas, Bayer, Janssen, Sanofi Transportation supported by Pfizer, Sanofi, Astellas, Janssen, MSD, AstraZeneca, Roche, Ipsen, Novartis. All other authors have declared no conflicts of interest.

803P

Indirect comparison of abiraterone acetate and docetaxel for treatment of metastatic “hormone-sensitive” prostate cancer

Conclusions: Our analyses suggest that ADTþAAþP has greater reduction in risk of progression and death vs ADTþDOC. In absence of head-to-head trials, indirect comparisons based on Bayesian NMA can provide useful insights to clinicians and reimbursement decision makers on the relative efficacy of treatment options for men with mHSPC. Clinical trial identification: NCT01715285 (LATITUDE), NCT00309985 (CHAARTED), NCT00104715 (GETUG-AFU 15) Legal entity responsible for the study: Janssen Global Services, LLC Funding: Janssen Global Services, LLC Disclosure: F. Saad: Honoraria, Consulting/Advisory Role and Research Funding: Janssen, Astellas, Sanofi, Bayer T. Li, T. Ito, J. Diels, P. De Porre: Employment at Janssen Research and Development Stock Ownership of Johnson & Johnson. S. Van Sanden: Employment at Janssen Research and Development Stock Ownership of Johnson & Johnson. J. Roı¨z: Employment: Evidera Stock ownership: Evidera. S. Abogunrin: Employment: Evidera Research Funding: Janssen, Novartis, GSK, Shire, Sanofi, Amgen, Pfizer, Bayer, Ipsen, Clovis, AstraZeneca, Boehringer Ingelheim. K. Fizazi: Honoraria, Consulting/Advisory Role: Astellas, Bayer, Clovis, Curevac, Genentech, Janssen, Orion, Sanofi. All other authors have declared no conflicts of interest.

804P

S. Feyerabend1, F. Saad2, T. Li3, T. Ito4, J. Diels5, S. Van Sanden5, P. De Porre6, J. Roı¨z7, S. Abogunrin8, K. Fizazi9 1 Urologic Oncology, Studienpraxis Urologie, Nu¨rtingen, Germany, 2Surgery, Centre Hospitalier de l‘Universite´ de Montre´al/CRCHUM, Montre´al, QC, Canada, 3Global Market Access Oncology, Janssen Global Services, Raritan, NJ, USA, 4Health Economics & Market Access EMEA, Janssen, High Wycombe, UK, 5Health Economics & Market Access EMEA, Janssen, Beerse, Belgium, 6Clinical Oncology, Janssen Research & Development BE, Beerse, Belgium, 7Modeling & Simulation, Evidera, London, UK, 8Outcomes Research, Evidera, London, UK, 9Cancer Medicine, Institut Gustave Roussy, University of Paris Sud, Villejuif, France Background: Androgen deprivation therapy (ADT) with or without chemotherapy (docetaxel [DOC]) is recommended in the clinical guidelines as the mainstay of management for metastatic “hormone-sensitive” prostate cancer (mHSPC). The LATITUDE trial demonstrated the efficacy of abiraterone acetate in combination with prednisone and ADT (ADTþAAþP) vs ADT in newly diagnosed mHSPC pts with high-risk disease (NDx HRD). We performed an indirect comparison to determine the relative efficacy of AA vs DOC in mHSPC. Methods: We conducted a systematic literature review of randomized controlled trials (RCTs) of treatments for mHSPC. To increase comparability of results across studies, the population of interest from LATITUDE and DOC studies was restricted to men with NDx HRD and/or high volume disease (NDx HVD). Two RCTs (CHAARTED, GETUG-AFU 15), both evaluating ADT vs DOCþADT, met the inclusion criteria. Fixed effects Bayesian network meta-analyses (NMAs) were performed to estimate the relative treatment effects for ADTþAAþP vs DOCþADT on overall survival (OS) and radiographic progression-free survival (rPFS). The HVD subgroup of LATITUDE was used in the main analysis. The LATITUDE ITT population (NDx HRD) was included as a sensitivity analysis. As STAMPEDE did not report an NDx HVD/HRD subgroup, its M1 population was included in a sensitivity analysis. Results: The results for HRD/HVD suggested improvement with ADTþAAþP vs DOCþADT in OS (HR 0.84) and in rPFS (HR 0.73), with Bayesian probabilities (P) for ADTþAAþP 86.8% (OS) and 99.2% (rPFS) more effective. Main results were consistent with all sensitivity analysis results (Table).

Practice patterns in metastatic castration-resistant prostate cancer (mCRPC): Evidence from the veterans health administration

A. Halwani1, Z. Burningham2, K.M. Rasmussen2, V. Patil2, S. Narayanan3, S-W. Lin3, S. Carroll3, L-I. Hsu3, J. Graff4, R. Dreicer5, S. Gupta1, C. Low1, B.C. Sauer2 1 Division of Hematology, Huntsman Cancer Institute, Salt Lake City, UT, USA, 2Division of Epidemiology, University of Utah, Salt Lake City, UT, USA, 3Oncology, Genentech, Inc., South San Francisco, CA, USA, 4Division of Oncology, Oregon Health & Science University- Knight Cancer Center, Portland, OR, USA, 5Division of Hematology/ Oncology, University of Virginia, Charlottesville, VA, USA Background: Practice patterns for metastatic castration-resistant prostate cancer (mCRPC) have evolved over the last decade due to introduction of agents such as abiraterone and enzalutamide. This study examines mCRPC treatment practices over a 10year period (which includes the time periods before and after the introduction of novel oral anti-androgens) for the first 3 lines of therapy in the largest nationwide integrated health system in the United States, the Veterans Health Administration. Methods: By linking patient information from the Veterans Affairs (VA) Clinical Cancer Registry to clinical notes, laboratory, procedure and imaging data from the VA Corporate Data Warehouse (CDW), we identified patients who were diagnosed with prostate cancer at the VA and ultimately developed mCRPC, defined as radiological evidence of metastasis and evidence of rising PSA levels concomitant with surgical (bilateral orchiectomy) or medical castrate testosterone levels ( 50 ng/dL within the last 3 months or ongoing treatment with androgen deprivation). Therapies used to treat mCRPC were extracted from CDW pharmacy dispensation records (docetaxel, abiraterone, enzalutamide, cabazitaxel, and others). Results: From 2006 to 2015, 120,374 patients were diagnosed with prostate cancer, of whom 3,637 developed mCRPC. Median age at initial prostate diagnosis was 68 years (range, 41-94), average BMI was 26.5 (range, 9-59), average CCI score was 1.5 (range, 0-12) and average PSA was 45.5 ng/mL. Practices for the first 3 lines of treatment from 2006 to 2010 and 2011 to 2016 are summarized in Table 1. Patients diagnosed with mCRPC between 2006 and 2010 were more likely to receive cytotoxic therapy than patients diagnosed between 2011 and 2016 (37% vs 22%). Compared with the cohort diagnosed between 2006 and 2010, the later cohort was more likely to receive treatment

Table: 803P ADT1AA1P vs ADT LATITUDE HVD & HRD Main analysis OS X rPFS X Sensitivity analysis OS OS X rPFS

ADT1DOC vs ADT CHAARTED

GETUG- AFU 15

STAMPEDE

HVD

HVD

M1

HRD (ITT)

X X

Volume 28 | Supplement 5 | September 2017

ADT1AA1P vs ADT1DOC

X

X X

X X

X X X

X

HR [95%-CrI]

PAA>DOC

0.84 [0.63, 1.14] 0.73 [0.57, 0.94]

86.8% 99.2%

0.92 [0.69, 1.23] 0.79 [0.61, 1.03] 0.80 [0.63, 1.02]

72.2% 96.0% 96.6%

doi:10.1093/annonc/mdx370 | 277

abstracts

Annals of Oncology

Table: 804P ENTIRE COHORTa (20062016) N 5 3,637

mCRPC Diagnosis (2006-2010) N 5 1,118

mCRPC Diagnosis (20112016) N 5 2,519

1L, % pts 2L, % pts 3L, % pts

DOC 27%, AA 22%, ENZ 6% AA 13%, ENZ 10%, DOC 4% ENZ 6%, DOC 4%, AA 3%

AA 29%, DOC 22%, ENZ 9% AA 15%, ENZ 14%, DOC 6% ENZ 8%, DOC 5%, AA 3%

No treatment, % pts Top 3 most common treatment sequences from 1L to 2L (% of pts)

1L 43%, 2L 68%, 3L 86% DOC-AA (11%) AA-ENZ (8%) AA-DOC (4%)

DOC 37%, AA 4%, ENZ 1% AA 7%, MIT 5%, CAB 3% AA 3%, ENZ 2%, DOC 2% 1L 56%, 2L 80%, 3L 90% DOC-AA (7%) DOC-MIT (5%) DOC-CAB (3%)

1L 38%, 2L 63%, 3L 81% DOC-AA (12%) AA-ENZ (11%) AA-DOC (5%)

AA, abiraterone acetate; CAB, cabazitaxel; DOC, docetaxel; ENZ, enzalutamide; MIT, mitoxantrone a 1 line of therapy (20% vs 37%). For patients diagnosed between 2011 and 2016, the most common therapies were as follows: 1L, abiraterone (29%); 2L, abiraterone (15%) and enzalutamide (14%); and 3L, enzalutamide (8%). Conclusions: Our study is the first to describe adoption of non-chemotherapeutic treatments in a nationwide cohort of patients with mCRPC treated in the largest integrated healthcare system in the United States. Further research will focus on understanding clinical outcomes associated with this shift in practice patterns. Legal entity responsible for the study: Ahmad Halwani, MD Funding: Genentech, Inc. Disclosure: A. Halwani: Research funding: Kyowa Hikko Kirin, BMS, Seattle Genetics, Roche/Genentech, Genentech, Takeda, Pharmacyclics, Amgen, Abbvie, Immune Design, Miragen, Seattle Genetics. S. Narayanan: Stock ownership and employee of Roche/Genentech Inc. S-W. Lin: Stock ownership and employee of Roche/Genentech Inc. S. Carroll: Stock ownership and employee of by Genentech Inc. L-I. Hsu: Stock owner and employee of Genentech Inc. J. Graff: Honorarium from Astellas, research funding from Sanofi, Merck, Bristol Myers Squibb, Astellas, Janssen, travel funds from Sanofi, Merck, and Clovis. R. Dreicer: Consulting relationships with: Orion, Medivation, Astellas, Asana, Genentech, Roche, AstraZeneca, Exelexis, Eisai, Bristol Myers Squibb, Sanofi-Genzyme, EMD Serono. S. Gupta: Spouse: Stock ownership in Salarius Pharmaceuticals. B.C. Sauer: Research funding from Genentech Inc. and Amgen. All other authors have declared no conflicts of interest.

805P

Efficacy and safety of first-line combined androgen blockade in advanced prostate cancer: A meta-analysis

Y. Sun, Y. Yang, R. Chen, L. Zhao, F. Liu, S. Ren, H. Wang, X. Lu, X. Gao, C. Xu Department of Urology, Shanghai Changhai Hospital, Shanghai, China Background: Combined androgen blockade (CAB) is one of the therapies for advanced prostate cancer. In this analysis, we compared the efficacy and safety of first-line CAB with castration monotherapy in advanced prostate cancer patients. Methods: The meta-analysis was PRISMA compliant and was registered in PROSPERO (CRD42016054301). We searched PubMed, EMBASE, Cochrane and Scholar for randomized controlled trials (RCTs) published through 12th December 2016 and compared efficacy and safety of first line CAB vs. luteinizing hormone releasing hormone agonists (LHRHa) monotherapy/orchiectomy in advanced prostate cancer. Overall survival (OS) and progression free survival (PFS) were the primary outcomes (fixed/ random effects model). Safety was the secondary outcome. Sub-group analyses included: i) Eastern vs. Western patients; ii) non-steroidal anti-androgen (NSAA) vs. steroidal anti-androgen (SAA). Studies with reported HR/presenting median survival and JADAD score >2 were included. Results: We identified 16 studies (6084 patients; West-12; East-4) for inclusion. CAB treatment significantly improved the OS (14 RCTS; HR 0.90; 95% CI 0.84 to 0.97, P ¼ 0.003) and PFS (13 RCTs; HR 0.89; 95% CI 0.80 to 1.00, P ¼ 0.04) in advanced prostate cancer patients, compared with monotherapy. No significant difference in OS

278 | Genitourinary tumours, prostate

(P ¼ 0.71) and PFS (P ¼ 0.49) was observed between Western vs. Eastern patients. CAB with NSAA significantly improved OS (HR 0.88; 95% CI 0.82 to 0.95, P ¼ 0.0009) and PFS (HR 0.85; 95% CI 0.73 to 0.98, P ¼ 0.007); whereas, CAB with SAA reported similar OS (HR 1.03; 95% CI 0.86 to 1.25, P ¼ 0.74) or PFS (HR 1.01; 95% CI 0.87 to 1.17, P ¼ 0.74) compared with castration monotherapy. Incidence of grade 3 or 4 AEs was not significantly different between CAB and castration monotherapy (P ¼ 0.1083). Conclusions: First-line CAB therapy significantly improved OS and PFS in advanced prostate cancer patients, with no significant difference in grade 3 or 4 AEs compared with castration monotherapy. Legal entity responsible for the study: Shanghai Changhai Hospital. 168 Changhai Road, Shanghai 200433, China Funding: AstraZeneca Disclosure: All authors have declared no conflicts of interest. 806P

Outcomes of prechemotherapy (pCRx) abiraterone acetate (AA) or enzalutamide (E) for metastatic castration-resistant prostate cancer (mCRPC) after ADT 1 Docetaxel (D) or ADT alone for metastatic hormone sensitive prostate cancer (mHSPC) in a multi-institution hospital-based registry

E. Francini1, S. Yip2, N.S. Ahmed2, H. Li3, L. Ardolino4, C.P. Evan1, M. Kaymakcalan1, G.K. Shaw1, P. Kantoff5, M.E. Taplin1, N. Alimohamed2, A.M. Joshua4, D.Y.C. Heng2, C.J. Sweeney1 1 Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute, Boston, MA, USA, 2Genitourinary Oncology, Tom Baker Cancer Centre, Calgary, AB, Canada, 3 Departments Oncology and Community Health Sciences, University of Calgary, Calgary, AB, Canada, 4Medical Oncology, Saint Vincent’s Hospital, Sydney, Australia, 5 Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, USA Background: The E3805: CHAARTED trial noted that the addition of D to ADT was associated with a hazard ratio (HR) of 0.56 (95% confidence interval [CI], 0.44 to 0.70; P < 0.001) for time to CRPC and resulted in a prolongation of overall survival (OS). Therefore, we postulated that pCRx AA or E had greater activity after ADTþD compared to after ADT alone. Methods: A cohort of mCRPC patients (pts) treated with pCRx AA or E for mCRPC between 2014 and 2017 was identified from three hospitals’ IRB approved databases. Patients were grouped by use of D for mHSPC. This time frame was chosen as ADTþD became a valid therapeutic option for mHSPC in 2014 and thus time to pCRx and follow-up were short. The endpoints included OS (time to death from all causes) from ADT start, time to AA/E start from ADT start, and OS from AA/E start. Survival outcomes were analyzed by Kaplan-Meier method. Results: Of the 102 identified, 50 (49%) had previously received ADT alone, while 52 (51%) had ADTþD. No statistically significant difference in OS from ADT start or from AA/E start was observed between the 2 cohorts (Table 1). Notably, survival in both groups from ADT start was shorter than commonly reported. Yet, deaths in the ADTþD group were 12 vs. 21 in the ADT alone, after a median follow-up of 24.4 and 29.8 months, respectively.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology

Table: 806P pCRx AA/E

ADT ADTþD

N (%)

50 (49%) 52 (51%)

N Deaths (median follow-up mo)

Median OS from ADT start (95% CI mo)

P-value

Time to AA/E start (95% CI mo)

P-value

Median OS from AA/E start (95% CI mo)

P-value

21 (29.8) 12 (24.4)

33.5 (22.4 – NR) NR (NR-NR)

0.2047

11.0 (8.5 – 13.7) 12.8 (11.1 – 15.7)

0.7265

17.3 (13.7 – NR) NR (13.1 – NR)

0.6514

Conclusions: In a cohort of ADT/ADTþD treated mCRPC pts with short times to pCRx AA/E and follow-up, the efficacy of AA/E is similar regardless of previous use of D. It is possible that the pts selected for ADTþD had poorer prognostic factors and yet still did at least as well with AA/E and deaths were lesser. Larger sample sizes, longer follow-up, and better characterization of patient and tumor factors are needed to assess the efficacy of different sequences. Legal entity responsible for the study: Edoardo Francini Funding: None Disclosure: P. Kantoff: SAB/consulting: Astellas, Bayer, Bellicum, BIND Biosciences Inc, BN Immunotherapeutics, DRGT, Ipsen Pharmaceuticals, Janssen, Metamark, MTG Therapeutics, New England Research Institutes, Omnitura, OncoCellMDX, Progenity, Sanofi, Tarveda Therapeutics. M.E. Taplin: Research funding and advisory board for Janssen and Medivation. N. Alimohamed: Consulting or advisory role for: Bayer, Sanofi, Pfizer, Astellas, Merck Inc., Novartis Pharma, Roche-Peru. D.Y.C. Heng: Consulting or advisory role: Pfizer, Novartis, Bristol-Myers Squibb, Janssen, Astellas Pharma. C.J. Sweeney: Consultant with compensation for: Astellas, Bayer, Genentech, Janssen, Pfizer, Sanofi. Research funding from: Astellas, Janssen, Sotio, Sanofi. All other authors have declared no conflicts of interest.

807P

First interim results of the radium-223 (Ra-223) REASSURE observational study: Analysis of patient (Pt) characteristics and safety by use of abiraterone and/or enzalutamide (Abi/Enza)

L.C. Harshman1, J. Logue2, C.N. Sternberg3, S. Sundar4, D. Schrijvers5, M. Schostak6, J. Sylvester7, S. George8, M. Tucci9, C. Mantz10, P. Borrega11, P. Ziem12, R. Concepcion13, K. Miller14, O. Sartor15, J. Kalinovsky16, Y. De Sanctis17, B. Tombal18 1 Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute, Boston, MA, USA, 2Oncology, The Christie NHS Foundation Trust, Manchester, UK, 3Department of Oncology, San Camillo and Forlanini Hospitals, Rome, Italy, 4Oncology, Nottingham University Hospitals NHS Trust-City Hospital Campus, Nottingham, UK, 5Department of Medical Oncology, Ziekenhuis Netwerk Antwerpen (ZNA) Middelheim, Antwerp,

Belgium, 6Department of Urology, Universit€ atsklinik fu¨r Urologie und Kinderurologie, Magdeburg, Germany, 7Oncology, 21st Century Oncology, Lakewood Ranch, FL, USA, 8 Dept of Medicine, Roswell Park Cancer Institute, Buffalo, NY, USA, 9Oncology, Azienda Ospedaliera Universitaria San Luigi Gonzaga, Orbassano, Italy, 10Oncology, 21st antara Hospital, Century Oncology, Fort Myers, FL, USA, 11Oncology, San Pedro de Alc C aceres, Spain, 12Klinik fu¨r Urologie, Dietrich-Bonhoeffer-Klinikum Neubrandenburg, 13 Neubrandenburg, Germany, Oncology, Urology Associates, Nashville, TN, USA, 14 Oncology, Benjamin Franklin Medical Center, Berlin, Germany, 15Oncology, Tulane University, New Orleans, LA, USA, 16Oncology, Bayer, Bratislava, Slovak Republic, 17 Oncology, Bayer, Whippany, NJ, USA, 18Urology, Cliniques Universitaires St. Luc, Brussels, Belgium Background: Ra-223 prolongs survival with a favorable safety profile in metastatic castration-resistant prostate cancer (mCRPC). The pivotal phase 3 ALSYMPCA trial had a relatively short 3-year follow-up and was conducted before availability of 2nd generation hormonal agents. The REASSURE study was designed to assess long-term safety (7 years follow-up) and conducted in an era when pts had access to other effective 1st line agents such as abi/enza. Methods: REASSURE is a global, prospective, single-arm, observational study that enrolled pts with mCRPC with bone metastases planned to start Ra-223. Treatment decision was made independently before enrollment. We undertook a planned interim descriptive analysis of safety and drug completion based on prior or concomitant abi/ enza use. Results: REASSURE enrolled 1106 pts in N. America and Europe from Sep 2014 to Sep 2016. The interim analysis included 583 pts who received 1 Ra-223 dose (Table; median 7 months observation). Prior abi/enza use was reported in 168 (29%) and concomitant in 153 (26%) pts. Treatment-related adverse events (TRAEs) occurred in 37%: prior abi/enza 45%, no prior abi/enza 34%; concomitant abi/enza 29%, no concomitant abi/enza 40%. TRAEs were most often gastrointestinal or hematological, with permanent discontinuation of Ra-223 in 6%: prior abi/enza 8%, no prior abi/ enza 5%; concomitant abi/enza 5%, no concomitant abi/enza 7%. Serious TRAEs (mostly hematologic) occurred in 4.5% leading to permanent Ra-223 discontinuation in 1.5%.

Table: 807P Baseline characteristics and treatment completion by prior or concomitant* abi/enza

ECOG 0–1, n (%) No. of metastases**, n (%) 20 Superscan ALP (U/L), median 8, and unknown in 19%, 49%, 23%, and 9%, respectively. At enrollment; the median PSA was 12.71 ng/ml. CGA showed cognitive impairment in 61%, physical impairment in 54%, depression in 16% and impairment of instrumental activities of daily living in 13%. The median time to 90% PSA decline was 7 weeks. 79% of patients had 80% DHT decline by 9 months. At the time of analysis, all patients had PSA decline of > 90% without radiographic evidence of disease progression. Baseline CGA did not correlate with efficacy (P-values >0.1). Common Grade 1 AEs included gynecomastia (26%), fatigue (35%), hot flashes (22%) and paresthesia (13%). None had Grade 3 or 4 AEs. Three men withdrew from the study due to treatment-related AEs (Grade 2 fatigue and paresthesia). Another three patients withdrew due to unrelated issues. Conclusions: Enz with Dut/Fin appears to be safe and efficacious for older patients with M0 and M1 HNSPCa. Future research will report effects of treatment on CGA domains. Clinical trial identification: NCT02213107 Legal entity responsible for the study: University of Rochester Funding: Astellas and Pfizer Inc Disclosure: All authors have declared no conflicts of interest.

809P

Cabazitaxel plus prednisone and prophylaxis of neutropenia complications in the treatment of metastatic castration-resistant prostate cancer after failure to docetaxel: A multicenter, noncomparative, open-label, phase IV study

F.C. Maluf1, D. Cubero2, F.A.M. de Oliveira3, P.E.R. Liedke4, L. Brust5, C.G. Inoc^encio6, F.S.M. Monteiro7 1 Clinical Oncology Department, Hospital Beneﬁcencia Portuguesa of Sao Paulo, Sao Paulo, Brazil, 2Hematology and Oncology, Faculty of Medicine of ABC, Santo Andre´,

280 | Genitourinary tumours, prostate

Annals of Oncology Brazil, 3Clinical Oncology Department, Hospital Erasto Gaertner, Curitiba, Brazil, Oncology Clinical Search Unit, Hospital de Clınicas de Porto Alegre, Porto Alegre, Brazil, 5Oncology Department, Hospital Bruno Born, Lageado, Brazil, 6Oncology and Hematology, “Dr. Ademar Lopes” Search Center, Cuiaba, Brazil, 7Clinical Oncology, Hospital Santa Lucia, Brasilia, Brazil 4

Background: This study aimed to evaluate the effectiveness of granulocyte colonystimulating factor (G-CSF) and ciprofloxacin in the prophylaxis of hematological complications in mCRPC patients treated with cabazitaxel after docetaxel failure and at risk for neutropenia. Methods: Phase IV, non-randomized, open-label, single-arm interventional study, with men aged 65 years (or < 65 years and 25% irradiated bone marrow), presenting mCPRC after docetaxel failure, ECOG status 1, life expectancy >12 weeks, that provided informed consent. Cabazitaxel 25 mg/m2 was given with prednisone on day1 every 21 days. G-CSF was administered on days 2 to 8 of each cycle or until ANC >2,000/mm3 and ciprofloxacin 1000mg on days 5 to 12. Primary endpoint was the rate of neutropenia grade 3 during the first cycle; secondary endpoints were the rate of neutropenia grade 3, febrile neutropenia, diarrhea grade 3, PSA response and quality of life (FACT-P) during treatment. Statistical significance was set at 0.05 and 95% confidence intervals were determined. Results: 46 patients with median age 71.5 years (mean: 71.8 years) and 69.0 months on median since diagnosis (mean: 75.2 months) of prostatic cancer were included. Among the 45 treated patients, exposed to a median of 9.0 cycles (mean: 9.5 cycles) during 210 days, 40.0% (95% CI, 25.7%-54.3%) presented one episode of neutropenia grade 3 during the first cycle. During treatment, 42.2% patients presented at least one neutropenia grade 3; febrile neutropenia occurred in one patient (2.2%) as well as diarrhea grade 3. Twenty-nine patients (64.4%) achieved PSA response and 77.2% improved FACT-P score in at least one visit. Three patients (6.7%) had a serious TEAEs leading to death (none related to treatment), and 13.3% had 7 TEAEs leading to treatment discontinuation (3 related to treatment). Conclusions: Prophylactic G-CSF and ciprofloxacin was effective in the prevention of neutropenia grade 3 and other hematological complications during the mCRPC treatment with cabazitaxel 25 mg/m2 in patients who were at risk for neutropenia. Clinical trial identification: NCT01649635 - Global Statitic Report: January 26th 2017. Legal entity responsible for the study: Sanofi Funding: Sanofi Disclosure: F.C. Maluf: Speaker at Sanofi-Aventis, Janssen, Bayer, AstraZeneca, Astellas, Ferring, Ache. Clinical trial investigator at Sanofi-Aventis, GSK, Janssen, Astellas, Bayer. Member of advisory board at Sanofi-Aventis, Janssen, Bayer, Astellas. F.A.M. de Oliveira: Speaker at Janssen. P.E.R. Liedke: Grants from Roche, Pfizer, Sanofi. L. Brust: Advisory Board Boheringer, Janssen, Agendia. F.S.M. Monteiro: Speaker at Sanofi, BMS, Pfizer, Janssen, Astellas e Merck Serono. Advisory Board at Janssen. Chair LACOG GU (Latin America Cooperative Oncology Group - Genito– Urinary Tumors). All other authors have declared no conflicts of interest.

810P

Assessment of health-related quality of life (HRQL) in PROSELICA: A Phase 3 trial assessing cabazitaxel 20 mg/m2 (C20) vs 25 mg/m2 (C25) in post-docetaxel (D) patients (pts) with metastatic castration-resistant prostate cancer (mCRPC)

M. Eisenberger1, A-C. Hardy-Bessard2, C.S. Kim3, L. Ge´czi4, D. Ford5, L. Mourey6, o10, G. Barnes11, H. Wang12, W. Zhang12, J. Carles7, P. Parente8, A. Font9, G. Kacs A. Ozatilgan13, J. de Bono14 1 Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins School of Medicine, Baltimore, MD, USA, 2CARIO, Centre Armoricain d’Oncologie, Ple´rin, France, 3 Department of Urology, Asan Medical Center, Seoul, Republic of Korea, 4Department of Oncological Internal Medicine, National Institute of Oncology, Budapest, Hungary, 5 Cancer Centre Queen Elizabeth Hospital, City Hospital, Birmingham, UK, 6IUCT-O, Institut Claudius Regaud, Toulouse, France, 7Vall d’Hebron Institute of Oncology, Vall d’Hebron University Hospital, Barcelona, Spain, 8Eastern Health Clinical School, Monash a d’Oncologia, Hospital Universitari Univeristy, Box Hill, Australia, 9Institut Catal Germans Trias i Pujol, Badalona, Spain, 10Amethyst Cluj, Iuliu Hatieganu Medical 11 University, Cluj-Napoca, Romania, Department of Biostatistics, Sanoﬁ, Cambridge, MA, USA, 12Department of Biostatistics, Sanoﬁ, Bridgewater, NJ, USA, 13Medical Affairs, Sanoﬁ, Cambridge, MA, USA, 14Institute of Cancer Research, Royal Marsden NHS Foundation Trust, Sutton, UK Background: PROSELICA (NCT01308580) assessed effect of C20 vs C25 on overall survival in a non-inferiority study of pts with mCRPC. Primary analyses included assessment of HRQL in the overall population. Post-hoc subgroup analyses investigated changes in HRQL in pts receiving C20 vs C25 according to median treatment cycles received (6). Methods: Functional Assessment of Cancer Therapy Prostate (FACT-P) was used to assess HRQL. The least square means of change in FACT-P total score (TS) from baseline (BL) was assessed via a mixed-effect model for repeated measurements and differences were compared for C20 vs C25 in pts receiving > 6 or 6 treatment cycles. Results: Overall change in FACT-P TS from BL to Cycle 10 was not significantly different for C20 vs C25 (C20 n ¼ 137: 0.02 [95% confidence interval [CI] -2.57, 2.61]; C25 n ¼ 141: 1.33 [95% CI -1.26, 3.93]; p ¼ 0.369). For evaluable pts who received > 6

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology cycles, change in FACT-P TS from BL to Cycle 10 favored C25 but not C20 (C25 n ¼ 140: 3.06 [95% CI 0.25, 5.86], p ¼ 0.033; C20 n ¼ 137: 2.67 [95% CI -0.17, 5.51], p ¼ 0.065). Difference in change was not significant for C20 vs C25 (-0.39 [95% CI: 3.66, 2.88], p ¼ 0.816). For evaluable pts who received 6 cycles, change in FACT-P TS from BL to Cycle 6 favored pts receiving C25 (C25 n ¼ 49: -4.61 [95% CI: -8.27, -0.95], p ¼ 0.014; C20 n ¼ 39: -6.58 [95% CI: -10.46, -2.69], p < 0.001) but the difference between the treatment arms was not significant (-1.96 [95% CI: -6.8, 2.87], p ¼ 0.426). Increasing cycles, BL ECOG performance score (0–1 vs 2) and receiving > 6 cycles significantly improved FACT-P TS change from BL (p < 0.001). Difference in treatment dose (C20 vs C25) did not have a significant effect on FACT-P TS change from BL (p ¼ 0.354). Conclusions: In the overall population, HRQL did not differ significantly from BL to Cycle 10 for C20 vs C25. Additionally, there were no significant differences between the two treatment arms (C20 vs C25) in either subgroup (> 6 or 6 cycles). A significant change in HRQL from BL to Cycle 10 was observed in patients who received > 6 cycles of C25. Funding: Sanofi. Clinical trial identification: NCT01308580 Legal entity responsible for the study: Sanofi Funding: Sanofi Disclosure: D. Ford: Honoraria from Astellas, Janssen and Sanofi. L. Mourey: Personal fees and non-financial support from Sanofi, Astellas, Janssen, Pfizer and Novartis, personal fees from Ipsen and Sandoz, non-financial support from Roche, grant from GSK. J. Carles: Consultancy role for Johnson & Johnson, Bayer, Astellas, BMS, Pfizer and Sanofi. G. Kacs o: Advisory boards and/or conferences for Amgen, Astellas, AstraZeneca, Bayer, Ipsen-Beaufour, Janssen, Novartis, Pfizer and Sanofi, clinical trials for Astellas, Ipsen-Beaufour, Janssen and Sanofi. G. Barnes: Consultant for Sanofi. H. Wang: Employee of Sanofi and own Sanofi stock. Previously employed by Merck & Co. W. Zhang: Consultant for Sanofi. A. Ozatilgan: Employee of Sanofi. J. de Bono: Grants and personal fees from Sanofi, AstraZeneca, Genentech and Genmab, personal fees from Pfizer, Merck, Merck Serono, Daiichi-Sankyo, grant from Taiho. All other authors have declared no conflicts of interest.

811P

Cabazitaxel followed by androgen deprivation therapy (ADT) significantly improves time to progression in patients with newly diagnosed metastatic hormone sensitive prostate cancer (mHSPC): A randomized, open label, phase III, multicenter trial

O. Andre´n1, A. Widmark2, A. Falt3, E. Ulvskog4, S. Davidsson1, C. Thellenberg Karlsson2, M. Hjalm-Eriksson5 1 Department of Urology, Faculty of Medicine and Health, Orebro University, Orebro, Sweden, 2Department of Radiation Sciences, Oncology, Umea University, Umea, Sweden, 3Department of Epidemiology and Biostatistics, Faculty of Medicin and Health, Orebro University, Orebro, Sweden, 4Department od Oncology, Orebro University, Orebro, Sweden, 5Department of Oncology, Karolinsk Institutet, Stockholm, Sweden Background: Patients with newly diagnosed mHSPC have a poor prognosis with a 3-year overall survival (OS) rate of 50%. Recently, combination of docetaxel (75mg/m2 every 6 weeks for 6 cycles) with ADT has become a new standard for such patients, based on results of 2 large phase 3 trials showing a significant OS benefit. In these trials, docetaxel was initiated within 3 months after ADT start. Timing of ADT and chemotherapy (CT) is controversial. In breast cancer, endocrine therapy is always started after CT, the rational being that ADT will turn clones of tumor cells in to a stage of dormancy where CT is less effective. Methods: This phase 3 trial randomized newly diagnosed mHSPC patients to receive cabazitaxel (CABA), 25 mg/m2 every 3 weeks for 10 cycles, followed by ADT (immediately after last CABA cycle) versus ADT alone. Primary end-point was OS. Secondary end-point was progression free survival. The study planned to include 400 patients but was closed prematurely due to low inclusion rate. A total 31 patients with newly diagnosed mHSPC were included and here we present the results. Results: Median follow up was 31 month. Of the CABA treated patients, 66.8% got six cycles or more and 46.7% completed all 10 courses. Median OS was 32,5 months with CABA followed by ADT and 29,5 months with ADT alone (HR 1.43, 95% CI 0.38-5.38). Median progression free survival was significantly longer in CABA treated patients (29 vs 12 months, HR 3.96 (95% CI 1.49-10.49). Main grade 3 toxicities were neutropenia (66%). Conclusions: In conclusion, results from this prematurely terminated trial suggest that CABA followed by ADT is effective in newly diagnosed mHSPC and shows a manageable toxicity. These results have to be validated in larger randomized trials. Clinical trial identification: NCT01978873 Legal entity responsible for the study: Department of Urology, Orebro University Hospital, Sweden Funding: Sanofi aventis Disclosure: All authors have declared no conflicts of interest.

812P

Health-related quality of life (HRQL) in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC) treated with cabazitaxel (CBZ) in a prospective observational study (CAPRISTANA)

A. Pichler1, G. Barnes2, J. Katolicka3, H. Korunkova4, A. Tomova5, M. Ghosn6, F. El Karak6, I. Koroleva7, J. Makdessi8, A. Ozatilgan9, S. Hitier10, J. Carles11 1 Department of Hematology and Oncology, Landeskrankenhaus LKH Leoben, Leoben, Austria, 2Department of Biostatistics, Sanoﬁ, Cambridge, MA, USA, 3Department of Oncology, St. Ann’s University Hospital, Brno, Czech Republic, 4Department of Oncology and Radiotherapy, University Hospital, Plzen, Czech Republic, 5Department of Oncology, Complex Oncology Center, Plovdiv, Bulgaria, 6Department of Hematology and Oncology, Hotel Dieu de France Hospital, Beirut, Lebanon, 7Department of Medicine, Medical University “Reaviz”, Samara, Russian Federation, 8Department of Medicine, St. George Hospital, Beirut, Lebanon, 9Global Medical Affairs Oncology, Sanoﬁ, Boston, USA, 10Department of Biostatistics, Sanoﬁ, Chilly-Mazarin, France, 11Vall d’Hebron University Hospital, Vall d’Hebron Institute of Oncology, Barcelona, Spain Background: Cabazitaxel (CBZ) is a taxane approved for mCRPC treatment post docetaxel based on Phase 3 clinical trial data. Observational studies are in progress to gather information on real-world treatment patterns, safety, efficacy and HRQL effect of CBZ outside of clinical trials. Methods: The prospective, observational study CAPRISTANA evaluated the routine clinical use of CBZ (25 mg/m2 every 3 weeks plus prednisone 10 mg/day) in pts with mCRPC previously treated with docetaxel. HRQL was assessed using Functional Assessment of Cancer Therapy - Prostate (FACT-P) version 4 and EQ-5D-3L (including VAS - visual analogue scale) questionnaires at baseline and every two cycles until CBZ discontinuation. Results: A total of 192 pts were treated in 55 centers across 6 countries (Apr 2012–Jun 2016); 161 and 157 pts were evaluable for FACT-P and EQ-5D, respectively. Pts received 6 (median) cycles of CBZ (range 1–24); 53.6% achieved disease control with CBZ. The main reason for CBZ treatment discontinuation was disease progression (58.3%). No new safety signals were identified. In the overall FACT-P score analysis, HRQL improvement during CBZ treatment was recorded in 31.8%, no change in HRQL in 40.4%, and deterioration was recorded in 27.8% of pts. The highest rate of improvement was observed for the Prostate-Specific Concerns subscale (49.3%) and Pain Control subscale (54.2%). The highest rate of deterioration was recorded for the Functional Well-Being subscale (40.9%). Mean FACT-P score and EQ-5D health utility index and VAS scores did not show statistically significant changes during CBZ treatment. Conclusions: In this real-world study investigating HRQL associated with the use of CBZ in pts with mCRPC, no significant changes were observed in mean on-treatment FACT-P score and EQ-5D scores. However, in contrast to observations in prospective clinical studies, pts had improvement in the Pain Control FACT-P subscale. These results suggest that, in addition to the previously demonstrated effectiveness, CBZ treatment may help pts to achieve better pain control. Legal entity responsible for the study: Sanofi Funding: Sanofi Disclosure: G. Barnes: Employee of Sanofi. M. Ghosn: Advisory boards for Sanofi, Astellas and Janssen. I. Koroleva: Research funding and speakers’ bureau for AstraZeneca and Teva, travel reimbursement from MSD and Eisai. A. Ozatilgan: Employee of Sanofi. S. Hitier: Employee of Sanofi. J. Carles: Consulting/advisory role to Johnson&Johnson, Bayer, Astellas, BMS, Pfizer and Sanofi. All other authors have declared no conflicts of interest.

813P

Real-world use of docetaxel for metastatic castration-resistant prostate cancer in China: Results from a large observational study

D. He1, Z. Sun2, J. Guo3, Z. Zhang4, Y. Shan5, L. Ma6, H. Li7, J. Jin8, Y. Huang9, J. Xiao10, Q. Wei11, D. Ye12 1 Department of Urology, The First Afﬁliated Hospital of Xi’an Jiaotong University, Xi’an, China, 2Department of Urology, Huadong Hospital Afﬁliated to Fudan University, Shanghai, China, 3Department of Urology, Zhongshan Hospital, Fudan University, Shanghai, China, 4Department of Urology, Sir Run Run Shaw Hospital Zhejiang University School of Medicine, Hangzhou, China, 5Department of Surgery/Urology, The Second Afﬁliated Hospital of Soochow University, Suzhou, China, 6Department of Urology, The Third Hospital of Peking University, Beijing, China, 7Department of Urology, Peking Union Medical College Hospital, Beijing, China, 8Department of Urology, Peking University First Hospital, Beijing, China, 9Department of Urology, Renji Hospital Shanghai Jiaotong University School of Medicine, Shanghai, China, 10 Department of Urology, The Second Afﬁliated Hospital of Zhejiang University School of Medicine, Hangzhou, China, 11Department of Urology, West China Hospital, Sichuan University, Chengdu, China, 12Department of Urology, Fudan University Shanghai Cancer Center, Shanghai, China Background: This study investigated real-world patterns of docetaxel use for metastatic castration-resistant prostate cancer (mCRPC) in China.

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx370 | 281

abstracts

Annals of Oncology

Table: 813P Pattern of use of docetaxel in Chinese patients with mCRPC

All patients After failure of 1st-line hormonal therapy After failure of 2nd-line hormonal therapy After failure of 3rd-line hormonal therapy After failure of estramustine therapy Other

n (%)

403 (100) 170 (42.2) 125 (31.0) 51 (12.7) 46 (11.4) 11 (2.7)

Median overall survival, months (95% CI)

PSA response rate, % (n/na)

22.4 (20.4, 25.8) 22.5 (19.2, 29.5)b 23.3 (18.1, 26.5)b 22.4 (19.0, 36.5) 20.2 (16.6, 27.7) 28.6 (17.5, not evaluable)

70.9 (168/237) 73.6 (64/87) 67.1 (55/82) 65.4 (17/26) 69.7 (23/33) 100.0 (9/9)

a

Denominator is the number of patients in each category who had PSA 20 ng/ml at baseline; p ¼ 0.781 for median overall survival with initiation of docetaxel following failure of 1st- and 2nd-line hormonal therapy. mCRPC, metastatic castration-resistant prostate cancer; PSA, prostate speciﬁc antigen.

b

Methods: A prospective, multi-centre, observational study of Chinese adults (18 years) with histologically confirmed metastatic prostate adenocarcinoma who received 1 dose of docetaxel following hormonal therapy failure (disease progression and serum testosterone 20), PSA value at diagnosis and previous to CT, PSA response previous to CT (>25%, 25-50%, 50-100% or no response), time from diagnosis to first docetaxel treatment (>or< 50days). Progression free (PFS) and overall survival (OS) were the endpoints analyzed by logrank test. Results: Median PFS was 17m and median OS has not been reached (80% of patients alive at 20 months). Median follow-up: 16.6m. Median age 64,3y (range: 46-80), median PSA at diagnosis 691ng/mL (range:15235-1), median PSA previous to CT 214ng/ mL (range:5060-0), PSA responses to ADT previous to CT was 25% in 35 pts (46%), 25-50% in 14 pts (18.4%), less than 50% in 13 (17.2%), no response in 14 (18.4%), Gleason 6-7: 19 (25%), 8-10: 53 (69,7%), UK 3 (5.3%). Median time from diagnosis to docetaxel 45.3 d (range 0-167). PFS nor OS was related to age, PSA at diagnosis, PSA response prior to docetaxel or Gleason. Time from diagnosis to docetaxel (p ¼ 0.04) (median 21 vs 15m; HR:2.2)) and Gleason (median not reached vs 15m; HR: 3.3) were statistically significant factors for PFS. Presence of visceral metastasis (p ¼ 0.08) (20m vs median not reached;HR: 3.8) and time from diagnosis to docetaxel (p ¼ 0.02) (median not reached vs 24m; HR:4.1) were significative factors for OS. Conclusions: A time from diagnosis to docetaxel start longer than 50 days is associated with lower PFS and OS in mþHSPC patients treated with ADT þ docetaxel. Gleason 8 score correlates with shorter PFS and the presence of visceral metastases with a lower OS. Legal entity responsible for the study: Dr. Miguel A. Climent Duran Funding: None Disclosure: All authors have declared no conflicts of interest.

815P

Prognostic value of systemic inflammatory biomarkers in patients with mCRPC treated with abiraterone in pre-docetaxel setting

R. Ratta1, E. Verzoni1, F. Pantano2, P. Grassi1, D. De Lisi2, A. Onorato2, M. Prisciandaro1, R. Montone3, F. de Braud1, D. Santini2, G. Procopio1 1 Medical Oncology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 2 Medical Oncology, Campus Bio-Medico University, Rome, Italy, 3Trial Center, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy Background: Systemic inflammatory biomarkers have shown a prognostic impact in several solid tumors. The aim of this study was to examine the prognostic role of baseline neutrophil-to-lymphocyte-ratio (NLR), platelet-to-lymphocyte-ratio (PLR) and lymphocyte-to-monocyte-ratio (LMR) and NLR, PLR and LMR changes at 1, 2 and 3 months in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC) treated with Abiraterone Acetate (AA) in pre-docetaxel setting. Methods: We retrospectively included mCRPC pts treated with AA at two Italian hospitals from November 2012 to April 2017. NLR, PLR and LMR were evaluated at baseline and after 1, 2 and 3 months of treatment. The impact of NLR, PLR and LMR on progression-free survival (PFS) was evaluated by Cox regression analyses both in univariate and multivariate fashion. Other clinico-pathological factors, such as PSA baseline level, Time to CRPC, Gleason Score, Presence of Visceral Metastases and Bone Metastases Burden were included. Results: Fifty mCRPC pts treated with AA were evaluated. At univariate analysis, elevated baseline NLR and PLR were significantly associated with shorter median PFS (p ¼ 0.01, hazard ratio [HR]¼1.224 and p ¼ 0.0001, HR ¼ 1.013 respectively); after 1

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology month of treatment, NLR and PLR were significantly predictors of worst PFS (p ¼ 0.03, HR ¼ 1.320 and p ¼ 0.02, HR ¼ 1.012 respectively). After 2 and 3 months of treatment, only high PLR was associated with poor prognosis (p ¼ 0.01, HR ¼ 1.012 at month 2; p ¼ 0.009, HR ¼ 1.009 at month 3 respectively). LMR didn’t show any prognostic relevance. At multivariate analysis, only baseline PLR was independently associated with PFS (p ¼ 0.006, HR ¼ 1.013). Conclusions: High baseline and early-assessed NLR and PLR during treatment with AA are associated with shorter PFS in mCRPC pts. PLR more than NLR may be considered as an early and easy-to-perform prognostic marker in this setting. Legal entity responsible for the study: Fondazione IRCCS Istituto Nazionale dei Tumori of Milan Funding: None Disclosure: E. Verzoni: Advisory boards: Jannsen. G. Procopio: Advisory board: Astellas, Bayer, Janssen and Roche. All other authors have declared no conflicts of interest.

816P

Table: 817P Distribution and 3-year cancer-specific survival rates for different scores in the evaluated cohort

68Ga-PSMA-PET/CT as a changing practice tool in biochemically recurrent prostate cancer

P.M. Domingues1, T. Abreu2, M.B. Siqueira3, G. Marinho1, D. Rosa1, H. Cottas1, L.H. Araujo1, D. Herchenhorn2, F.A. Peixoto1 1 Medical Oncology, Instituto COI - Clinicas Oncologicas Integradas, Rio De Janeiro, Brazil, 2Medical Oncology, Oncologia DOR, Rio De Janeiro, Brazil, 3Medical Oncology, Oncoclinica, Rio De Janeiro, Brazil Background: PSMA-PET/CT have demonstrated interesting results in the detection of loco-regional and distant disease in prostate cancer patients with biochemical relapse. Even with low levels of PSA, PSMA imaging is able to identify metastatic lesions, being a possible tool for tailoring treatment decisions. This study aims to describe the use of PSMA-PET/CT in the daily practice and its clinical impact in the management of prostate cancer patients who have rising PSA after curative treatment. Methods: We performed a retrospective analysis in 29 localized prostate cancer patients of three private Brazilian cancer institutions who underwent PSMA-PET/CT for rising PSA after treatment with curative intent. The clinical impact of PSMA-PET/CT was evaluated by whether the assistant physician changed or not the treatment strategy based solely on PSMA results. In addition, modifications related to local (salvage radiotherapy [SRT], salvage lymphadenectomy [SL]) and systemic (antiandrogen deprivation therapy [ADT], chemotherapy [chemo]) treatment were described. Results: In total, 29 patients were enrolled. Twenty-seven (93%) had undergone radical prostatectomy, and 2 (7%) radiotherapy as the local curative treatment. Sixteen cases (55%) had not received any radiotherapy previously. The mean Gleason score, PSA level and PSADT at time of the examination were 8, 4.2 (0,05-41) ng/ml and 4.4 (0.427) months, respectively. PSMA-PET/CT detected at least one suspicious lesion for prostate cancer in 21/29 (58%) patients. Overall, 15/29 (51%) patients had their treatment strategy changed due to results in PSMA imaging. In only 3/29 (10%) the modifications were related exclusively to systemic protocols (1 avoided ADT, 1 added ADT and, 1 added chemo). Whereas in the 12/29 (41%) remaining cases, treatment strategy change involved local treatment. Of these 12 with a local treatment change, 7 added (6 SL, 1 SRT) and 5 avoided (5 SRT) local therapies. Conclusions: Half of the patients with biochemical relapse that underwent PSMAPET/CT had their treatment protocol changed, most changes related to local treatment. Although the role of PSMA imaging is not clearly defined, PSMA-PET/CT has been used as a practice changing tool in the daily practice. Legal entity responsible for the study: Instituto COI Funding: None Disclosure: All authors have declared no conflicts of interest.

817P

Results: A total of 8727 patients with treatment-naı¨ve advanced prostate cancer and complete baseline data were identified in the period from 2010-2013. The following factors were associated with better cancer-specific survival (isolated regional nodal disease, lower PSA level and lower grade group) (P < 0.0001). Based on the results of the multivariate analysis, the prostascore was described. A prostascore point was given for each of: PSA level 60 and grade group 4 or 5. The site/distribution of extra-prostatic disease were given the following points: 0 for isolated regional lymph nodes (N1), 1 for non regional lymph deposits (M1a), 2 for bone deposits (M1b) and 3 for other sites (M1c). A total prostascore was then calculated for each case in the cohort (which may range from 0 to 5). After assignment of a prostascore for each patient, cancer-specific and overall survivals were compared according to the score. Pair wise comparisons between all different scores were conducted. For both cancer-specific and overall survival assessment according to the prostascore model, P values for pair wise comparisons among different score points were significant (P1 Median (range) Previous VEGFR Tyrosine Kinase Inhibitor

Male ¼ 26 (86.7%) 58.4 years (25-81) 27 (90%) Papillary (P) 17 Chromophobe (Chr) 6 Other: unclassiﬁed 3, translocation 2, sarcomatoid (sarc) 1, mucinous tubular/spindle cell 1 Good/Intermediate/Poor 2/20/8 2/ 23/5 3 7 20 2 (0-5) 26

Median PFS was 8.6 months (mos) (95%CI: 6.1-14.7), and median OS was 22.7 mos (95%CI:10.8-NR), median follow up 10.6 months (95%CI: 7.1-14.1). There were no significant differences detected between patients with papillary versus non-papillary histologies with respect to PFS or OS. At last follow up, 13 patients remain on treatment with median time on therapy for all patients of 15.0 months. Of the 28 patients with measurable disease, there were 4 confirmed PRs (2 P, 1 Chr, 1 unclassifed) for a 14% ORR. For the entire cohort, 20 of 30 (66.7%) with stable disease, and 6 of 30 with progressive disease (20%), for a disease control rate of 24 of 30 (80%). Of 21 patients who started C at 60 mg/d, 12 (57%) required dose reduction due to toxicity. Multiple patients required treatment breaks but none discontinued therapy due to toxicity. Conclusions: In this retrospective study, C produced a clinically meaningful benefit in patients with metastatic vhRCC, the majority of whom had PD on prior VEGFR-TKIs. Prospective trials of C in vhRCC are warranted and planned. Legal entity responsible for the study: Matthew T Campbell Funding: None Disclosure: M.T. Campbell: Serve on advisory boards for Eisai and AstraZeneca. M.A. Bilen: Advisory board for Exelixis. N. Tannir: Served as a consultant and has served on advisory board for Exelixis. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx371 | 323

abstracts 913P

Avelumab in patients with metastatic adrenocortical carcinoma (mACC): Results from the JAVELIN solid tumor trial

C. Le Tourneau1, C. Zarwan2, C. Hoimes3, D.J. Wong4, S. Bauer5, M. Wermke6, H.J. Grote7, A. von Heydebreck8, K. Chin9, J. Gulley10 1 Department of Medical Oncology, Institut Curie, Paris, France, 2Medical Oncology, Lahey Hospital and Medical Center, Burlington, MA, USA, 3Hematology and Oncology, Case Western Reserve University and University Hospitals Seidman Cancer Center, Cleveland, OH, USA, 4Department of Medicine, University of California, Los Angeles Medical Center, Los Angeles, CA, USA, 5Department of Medical Oncology, University Hospital Essen Westdeutsches Tumorzentrum, Essen, Germany, 6Early Clinical Trial Unit, Universit€atsklinikum Dresden, Dresden, Germany, 7Oncology, Merck KGaA, Darmstadt, Germany, 8Biostatistics, Merck KGaA, Darmstadt, Germany, 9Immune-Oncology, EMD Serono, Inc, Billerica, MA, USA, 10Genitourinary Malignancies Branch, National Cancer Institute, Bethesda, MD, USA Background: Avelumab is a human anti–PD-L1 IgG1 antibody that has shown promising clinical activity in multiple tumor types, and is approved in the US for the treatment of metastatic Merkel cell carcinoma. Here, we report an updated analysis of avelumab in patients (pts) with mACC, representing the largest prospective monotherapy study performed to date in this rare cancer with limited therapeutic options. Methods: In a phase 1b cohort (NCT01772004), pts with mACC and prior platinumbased therapy received avelumab at 10 mg/kg IV Q2W until progression, unacceptable toxicity, or withdrawal. Prior and ongoing treatment with mitotane was permitted. Tumors were assessed every 6 wks (RECIST v1.1). Endpoints included safety (NCICTCAE v4.0), best overall response, objective response rate (ORR), progression-free survival (PFS), and overall survival (OS). Results: As of Dec 31, 2016, 50 pts from 6 countries received avelumab for a median of 3.4 mos (0.5–24.8). Median follow-up was 16.5 mos (11.7-27.6); 5 pts (10.0%) remained on treatment. Median age was 50 y (range 21–71) and median time since diagnosis of metastatic disease was 14.5 mos. 24 pts (48.0%) had received 2 prior lines of treatment for advanced disease (median 1, range 0–6). 41 pts (82.0%) had a treatmentrelated adverse event (TRAE) of any grade; the most common (>15%) were nausea (20.0%) and fatigue (18.0%). 8 pts (16.0%) had a grade 3 TRAE, of which only increased ALT (4.0%) occurred in > 1 pt. 12 pts (24.0%) had an immune-related AE of any grade. Confirmed ORR was 6.0% (3 partial responses; 95% CI 1.3–16.5); response was ongoing in 1 pt at data cutoff. 21 pts (42.0%) had stable disease as best response (disease control rate 48.0%). Median PFS was 2.6 mos (95% CI 1.4–4.0). Median OS was 10.6 mos (95% CI 7.4–not estimable) and the 12-mo OS rate was 47.0% (95% CI 31.8–60.9). Responses occurred in 2 pts with PD-L1þ tumors and 1 PD-L1 (5% tumor cell cutoff). In PD-L1 þ (n ¼ 12) vs PD-L1 (n ¼ 30) subgroups, median PFS was 5.5 vs 1.7 mos (HR 0.66; 95% CI 0.3–1.4) and median OS was 14.4 vs 11.5 mos (HR 0.82; 95% CI 0.3–2.2), respectively. Conclusions: Avelumab had a manageable safety profile and demonstrated clinical activity in pts with platinum-treated mACC. Clinical trial identification: NCT: NCT01772004 Protocol: EMR 100070-001 Legal entity responsible for the study: Pfizer Inc., New York, NY, USA; Merck KGaA, Darmstadt, Germany. Funding: Funding was provided by Pfizer Inc., New York, NY, USA and Merck KGaA, Darmstadt, Germany. Disclosure: C. Le Tourneau: Provided a consulting role for MSD, Bristol-Myers Squib, Novartis and AstraZaneca and received honoraria from Merck Serono and AstraZaneca. C. Zarwan: Provided an advisory role for Revere Pharmaceutics and consulting for Perceptive Informatics. C. Hoimes: Provided an advisory role for Seattle Genetics and Eisai, and participated in speaker’s bureau’s for Bristol-Myers Squib and Genentech. D.J. Wong: Received research funding from Armo Biosciences, BioMed Valley Discoveries, Roche-Genentech, Merck, EMD Serono, Bristol-Meyers Squibb, KURA Oncology, AstraZeneca. Provided an advisory role for Bristol-Meyers Squibb. S. Bauer: Research support: Novartis, Blueprint Medicines, Ariad; Consultant: GSK, Novartis, Pfizer, Bayer, Fresenius, Lilly, Blueprint Medicines, Deciphera; Honoraria (CME): Pharmamar, GSK, Pfizer, Bayer; Travel support: Pharmamar, Bayer. M. Wermke: Received research funding from Novartis, Pfizer, Roche, Novartis, Roche, Boehringer Ingelheim and Celgene. Provided an advisory role for Roche, Novartis, Bristol-Myers Squib, AstraZeneca and received honoraria from Roche, Novartis, Boehringer Ingelheim. H.J. Grote: Employee of Merck KGaA, Darmstadt, Germany. A. von Heydebreck: Employee of Merck KGaA, Darmstadt, Germany and holds Merck KGaA, Darmstadt, Germany stock. K. Chin: Employee of EMD Serono Inc. All other authors have declared no conflicts of interest.

Annals of Oncology

914P

Do patients (pts) with advanced nonseminomatous germ cell tumors (aNSGCT) and unfavorable time to normalization (TTN) of tumor markers benefit with prolongation of 1-st line chemotherapy (ChT)?

A. Tryakin1, M. Fedyanin1, J. Sergeev2, A. Bulanov1, B. Akhmedov3, T. Zakharova4, V. Matveev5, I. Fainstein2, A. Garin1, S. Tjulandin1 1 Clinical Pharmacology and Chemotherapy, Russian Oncology Research Center Named After Blokhin N.N., Moscow, Russian Federation, 2Radiosurgery, Russian Oncology Research Center Named After Blokhin N.N., Moscow, Russian Federation, 3Thoracic Surgery, Russian Oncology Research Center Named After Blokhin N.N., Moscow, Russian Federation, 4Pathology, Russian Oncology Research Center Named After Blokhin N.N., Moscow, Russian Federation, 5Urology, Russian Oncology Research Center Named After Blokhin N.N., Moscow, Russian Federation Background: Three-four cycles of BEP are commonly recognized as a standard 1-st line ChT in aNSGCT. However, there were no trials studying optimal cycles numbers in this setting, especially in case of unfavorable TTN of tumor markers. We performed retrospective single center analysis to evaluate if pts with unfavorable TTN of tumor markers may benefit with prolongation of induction ChT. Methods: Inclusion criteria were as follows: (1) ChT-naı¨ve aNSGCT pts treated with etoposide- and cisplatin-based chemotherapy; (2) AFP and hCG levels available at days 0 and 18-22 of cycle 1 to calculate TTN (Fizazi K., JCO 2004). Pts who received less than “standard” number of cycles (3xBEP or 4xEP for IGCCCG good risk, 4xBEP for intermediate and poor risk) for any reason were excluded from the analysis. Coxregression multivariate analysis was also performed. TTN was calculated by K.Fizazi’s methodic (JCO 2004). Results: From 1987 to 2011 952 pts with aNSGCT received 1-st line ChT. 584 pts matched the inclusion criteria. Unfavorable TTN had 24 (11%), 61 (41%) and 122 (84%) of pts with good, intermediate and poor IGCCCG risk, respectively. More than standard number of cycles received 199 pts. Prolongation of ChT did not result in significant improvement of OS in any IGCCCG prognostic groups irrespectively of TTN (Table).

Table: 914P IGCCCG risk group Good

TTN/# of cycles

favorable/standard favorable/>standard unfavorable/standard unfavorable/> standard Intermediate favorable/standard favorable/>standard unfavorable/standard unfavorable/> standard Poor favorable/standard favorable/>standard unfavorable/standard unfavorable/> standard

N pts

5-y OS, %

p (HR, 95%CI)

176 42 15 9 127 20 32 29 12 10 23 89

92% 98% 93% 78% 88% 80% 81% 86% 91% 100% 65% 72%

0,44 (HR 0,63, 0,23-1,89) 0,65 (HR 1,48, 0.24-10.48) 0,14 (HR 2,1, 0,73-9,38) 0,66 (HR 0,78, 0.25-2.41) 0,17 (HR 0,14, 0.01-2.27) 0,55 (HR 0,79, 0.33-1.79)

Conclusions: Prolongation of 1-st line ChT beyond standard number of cycles does not improve outcome of pts with unfavorable TTN of tumor markers. Legal entity responsible for the study: Alexey Tryakin Funding: None Disclosure: All authors have declared no conflicts of interest.

915P

The prognostic role of neutrophil-to-lymphocyte ratio (NLR) in patients with metastatic germ cell tumors

D. Ribnikar1, P.L. Bedard1, R.J. Hamilton2, M. Jewett2, P. Warde3, P. Chung3, L. Anson Cartwright3, A. Templeton4, E. Amir1, A.R. Hansen1, J. Lewin1 1 Division of Medical Oncology and Hematology, Princess Margaret Cancer Center, Toronto, ON, Canada, 2Department of Surgery, Princess Margaret Cancer Center, Toronto, ON, Canada, 3Department of Radiation Oncology, Princess Margaret Cancer Center, Toronto, ON, Canada, 4Department of Oncology, St. Claraspital Basel, Basel, Switzerland Background: NLR is a robust prognostic factor in many solid tumors. Limited data exist about its role in patients with metastatic germ cell tumors (GCTs).

324 | Genitourinary tumours, non-prostate

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Methods: We utilized a single institution database of patients diagnosed with metastatic GCTs between January 1990 and December 2013 who were treated with chemotherapy at Princess Margaret Cancer Centre. The peripheral blood count prior to first line chemotherapy was used to calculate the derived NLR (absolute neutrophil count divided by the total white blood cell count minus the absolute neutrophil count). Predictive accuracy was assessed as the association between NLR and overall survival and was evaluated using a Cox proportional hazard model adjusted for the International Germ Cell Cancer Collaborative Group (IGCCCG) risk classification. Discriminatory accuracy was evaluated by determining the area under the receiver operating characteristic curve (AUROC) for survival at 5 years. The optimal cut-off for NLR selection was chosen based on a highest AUROC. Results: In total, 475 patients were identified of which NLR data were available from 354 (75%) patients. Among these, 63% were good risk, 23% intermediate risk and 15% poor risk. The 5-year survival for good, intermediate and poor risk groups was 96.3%, 92.4% and 62.9%, while 10-year survival was 94.8%, 92.4% and 62.9%, respectively. Over the whole cohort, a NLR cut-off of 2.5 provided the best discriminatory accuracy with an AUROC of 0.70 (95% CI 0.59-0.75, p < 0.001). In a univariable analysis, NLR >2.5 was associated with a hazard ratio (HR) of 3.91 (95% CI 2.01-7.60, p < 0.001) which persisted after adjustment for IGCCCG risk group (HR 2.33, 95% CI 1.14-4.76, p ¼ 0.02). Among patients with IGCCCG high risk, 5-year survival was 87.5% if NLR 2.5, whilst if NLR > 2.5, 5-year survival was only 51.3%. Conclusions: A high NLR is associated with an adverse survival in patients with metastatic GCTs undergoing first line chemotherapy and provides moderate discriminatory accuracy in this setting. The utility of NLR appears particularly marked in patients with IGCCCG high risk disease. Legal entity responsible for the study: Senior authors, Dr. Jeremy Lewin and Dr. Eitan Amir Funding: None Disclosure: All authors have declared no conflicts of interest.

916P

Biological assessment of viable germ cell tumor (VT) in patients (pts) with seminoma (S) and non-seminoma (S) using miR371

L. Nappi1, M. Thi1, L. Fazli1, K. Chi2, B. Eigl2, C. Nichols3, M. Gleave1, C.K. Kollmannsberger2 1 Urologic Sciences, Vancouver Prostate Centre, Vancouver, BC, Canada, 2Medicine, British Columbia Cancer Agency, Vancouver, BC, Canada, 3Precision Genomics Cancer Research Clinic, Intermountain Health Care, Murray, UT, USA Background: The pathological constitution of residual nodes after chemotherapy or of the borderline enlarged retroperitoneal (RP) lymph nodes in clinical stage I (CSI) germ cell tumor (GCT) pts on surveillance is challenging, especially in tumor markers (TM) negative pts. Currently, accurate assessment requires pathological confirmation with RPLND or clinical follow-up to establish a pattern of growth. A plasma-based approach to identify patients with VT would be uniquely valuable. Methods: Formalin-fixed paraffin embedded (FFPE) and plasma from GCTs patients were used for miRNAs extraction. Non-cancer FFPE testicular tissue and plasma from healthy volunteers were used as negative controls. miR371 expression was detected by RT-PCR and relative expression calculated by the 2-DDCt method. miR-93-5p was used as positive internal control. Results were analyzed for associations with clinicopathologic features using Fisher’s exact test. Results: miR371 was over-expressed in all the primary testicular (n ¼ 4) and mediastinal (n ¼ 3) samples while it was undetectable in the atrophic testis (n ¼ 1) and mediastinal or gonadal teratoma (n ¼ 2), confirming the applicability of the technique to the FFPE samples. 21 metastatic samples were analyzed: 2 lung, 1 brain, 17 lymph nodes and 1 IVC tumor thrombus. The samples were collected prior to (n ¼ 2) or after (n ¼ 12) chemotherapy, while 7 pts were treated only with surgery. miR371 was undetectable in any samples (0/9) with no VT on pathological examination and overexpressed in 11/12 (91.6%) of those with viable GCT (OR 145.7; p < 0.0001). 90% of pts with positive miR-371 had negative TM (100% of S and 75% of NS) while no pts with negative miR-371 had positive TM. Plasma miR-371 also showed high correlation with VT (Table).

Table: 916P Pts

Initial stage

Stage at the suspicious relapse

Histology

miR71

1 2 3 4

I I I I

IIA IIA IIA IIA

S S S NS

þ – – þ

Volume 28 | Supplement 5 | September 2017

Evidence of VT

þ (RPLND) (negative PET scan) (regressing CT scan) þ (progressing CT scan)

Conclusions: Elevated plasma levels of miR-371 correlate with the presence of active germ cell malignancy. These encouraging findings suggest that plasma miR371 levels may lead to biological rather than radiographic assessment of the presence of active GCT in patients with S and NS. Legal entity responsible for the study: Lucia Nappi Funding: None Disclosure: All authors have declared no conflicts of interest.

917P

A single centre retrospective review of testosterone deficiency in germ cell cancer patients

L.S. Gibbs, G.L. Gullick, E. Allison, S. Brand, A. Addeo, A. Kuchel, J. Braybrooke Bristol Cancer Institute, University Hospitals Bristol NHS Trust Bristol Haematology and Oncology Centre, Bristol, UK Background: Testosterone deficiency syndrome (TDS) is frequently described in men treated for germ cell cancer with rates quoted between 11 and 38% (Huddart et al 2005). Observational studies show that TDS reduces quality of life and carries cardiovascular, metabolic and bone health risks. At our institute we observed that men with symptoms of TDS and ‘low normal’ testosterone (T) (8.6 – 12 nmol/L) were not reliably recognised. Methods: We collected retrospective data from all germ cell cancer referrals to the Bristol Cancer Institute from 2011 – 2016. We documented age, treatment, at least one random T level within a year of diagnosis (grouped into < 8, 8 – 12 and > 12 nmol/L), details of symptoms and treatment of TDS. Results: Data was collected on 462 patients (36 excluded with non germ cell diagnoses and 26 excluded due to T never being measured). Median age was 36 years (range 17 – 89) with 85% of patients aged under 50. 58% of men had seminoma, 32% nonseminoma and 10% combined germ cell cancer. 41% of all patients had a T level < 12 nmol/L at first measurement (32% of 20 – 29 year olds, 42% of 30 – 49 and 58% of 50 – 59 year olds) and 16% had T < 8 nmol/L. T therapy was prescribed in 19% of patients. Men receiving adjuvant carboplatin had the highest rate of T therapy (23%) compared with patients on surveillance (18%) and BEP or EP chemotherapy (14%). Conclusions: In this retrospective series 41% of patients had at least one total T value median exposure) vs low exposure ( median exposure). Maximum plasma concentration (Cmax) and area under the curve over the dosing interval (AUCtau) at steady state (SS) were the exposure metrics. Progression-free survival (PFS) was used as efficacy endpoint. Hazard ratios (HRs) and 95% confidence intervals (CI) for the low and high niraparib exposure groups in each cohort were provided, with exposure group as the independent variable and PFS as the dependent variable. Exposure and safety data of gBRCAmut and non-gBRCAmut cohorts were combined to evaluate exposure-safety relationships using logistic regression. Results: In the gBRCAmut cohort, the HR was 0.91 (95% CI, 0.54-1.52) for niraparib exposure as measured by the SS AUCtau for pts in the high exposure vs. the low exposure group. In the gBRCAmut cohort, the HR was 0.70 (95% CI, 0.49-0.99). Logistic regression analysis did not show any significant relationship between the incidence of thrombocytopenia or anemia Grade 3 and the SS Cmax or AUC increase. Conclusions: Observed exposure-response relationships support the selection of 300 mg as the starting dose in both gBRCAmut and non-gBRCAmut pt populations. A trend towards increased efficacy associated with increased exposure was observed in the non-gBRCAmut cohort. Clinical trial identification: NCT01847274 Legal entity responsible for the study: TESARO, Inc. Funding: TESARO, Inc. Disclosure: J. Wang, Z-Y. Zhang, X. Wang, S. Agarwal: Employment: Tesaro Stock: Tesaro. M.R. Mirza: Consulting or Advisory Role: Roche, Clovis Oncology, AstraZeneca, Tesaro. L. Gilbert: Honoraria: AstraZeneca, Advaxis Consulting/Advisory Role: AstraZeneca, Advaxis. A.V. Tinker: Consulting/Advisory: AstraZeneca. J.S. Berek: Consulting/Advisory: Atara Biotherapeutics. H.S. Pentikis, V. Kansra: Advisory board or board of directors: Tesaro Consulting: Tesaro. B. Benigno: Honoraria: AstraZeneca, Insys Therapeutics Research funding: Tesaro. S.J. Hazard: Employment: Tesaro, Genentech/Roche Stock: Tesaro, Genentech/Roche Travel, Accommodations, Expenses: Tesaro, Genentech/Roche. B.J. Rimel: Consulting/Advisory Role: AstraZeneca, Tesaro, Genentech/Roche Honoraria: Genentech. U.A. Matulonis: Consulting/Advisory: Merck KGaA, AstraZeneca, Immunogen, Tesaro, Genentech. All other authors have declared no conflicts of interest.

934PD

Safety and Efficacy of Niraparib in Elderly Patients (Pts) with Recurrent Ovarian Cancer (OC)

M. Fabbro1, K.N. Moore2, A. Dørum3, A.V. Tinker4, S. Mahner5, I. Bover6, S. Banerjee7, G. Tognon8, F. Goffin9, R. Shapira-Frommer10, R.M. Wenham11, K. Hellman12, D.M. Provencher13, P. Harter14, I. Palacio Vazquez15, P. Follana16, M.J. Pineda17, M.R. Mirza18, S.J. Hazard19, U.A. Matulonis20 1 Oncology, Groupe d’Investigateurs Nationaux pour l’Etude des Cancers Ovariens (GINECO) and Institut du Cancer de Montpellier, Montpellier, France, 2Obstetrics and Gynecology, Stephenson Cancer Center, University of Oklahoma HSC; Sarah Cannon Research Institute (Nashville, TN), Oklahoma City, OK, USA, 3Gynecologic Oncology, The Nordic Society of Gynecological Oncology (NSGO) and Radiumhospitalet, Oslo University Hospital, Oslo, Norway, 4Systemic Therapy, BC Cancer Agency Vancouver, akologische Vancouver, BC, Canada, 5Gynecologic Oncology, Arbeitsgemeinschaft Gyn€ Onkologie (AGO) and University of Munich, Munich, Germany, 6Medical Oncology, Grupo Espa~ nol de Investigaci on en C ancer de Ovario (GEICO) and Hospital Son Ll atzer, Palma De Mallorca, Spain, 7Gynaecology Unit, The National Cancer Research Institute (NCRI) and The Royal Marsden NHS Foundation Trust, London, UK, 8Obstetrics and Gynecology, ASST Spedali Civili di Brescia, Brescia, Italy, 9Gynecologic Oncology, Belgium and Luxembourg Gynaecological Oncology Group (BGOG) and University of Lie`ge, CHU de Lie`ge, Site Hoˆpital de la Citadelle, Liege, Belgium, 10Oncology, Israeli Society of Gynecologic Oncology (ISGO) and Sheba Medical Centre, Ramat Gan, Israel, 11 Gynecologic Oncology, H. Lee Mofﬁt Cancer Center & Research Institute, Tampa, FL, USA, 12Gynaecologic Oncology, The Nordic Society of Gynecological Oncology (NSGO) and Karolinska University Hospital, Stockholm, Sweden, 13Obstetrics and Gynecology, Centre Hospitalier de L’Universite de Montreal (CHUM), Montreal, QC, Canada, 14 Gynecology & Gynecologic Oncology, Arbeitsgemeinschaft Gyn€ akologische Onkologie nol de (AGO) and Kliniken Essen-Mitte, Essen, Germany, 15Oncology, Grupo Espa~ Investigacion en C ancer de Ovario (GEICO) and Hospital Central de Asturias, Oviedo, Spain, 16Medical Oncology, Groupe d’Investigateurs Nationaux pour l’Etude des Cancers Ovariens (GINECO) and Centre Antoine Lacassagne, Nice, France, 17Gynecologic Oncology, Northwestern University, Chicago, IL, USA, 18Oncology, Nordic Society of Gynecologic Oncology (NSGO) and Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark, 19Clinical Science, TESARO, Inc., Waltham, MA, USA, 20 Gynecologic Oncology, Dana Farber Cancer Institute, Boston, MA, USA Background: Niraparib (ZejulaTM) is a selective poly(ADP-ribose) polymerase 1/2 inhibitor (PARPi) approved for maintenance therapy in adults with recurrent OC who are in response to platinum-based therapy. Here, we report safety and efficacy of niraparib in the subgroup of pts from the ENGOT-OV16/NOVA trial who were aged 65 years (y). Methods: Pts were assigned to one of two independent cohorts based on germline BRCA mutation (gBRCAmut) status and randomized 2:1 within each cohort to receive either niraparib (300 mg) or placebo once daily. Pts were stratified by age (12. Pts were treated with a median of 6 cycles (range: 1-12) of T-PLD. 47 (51.6%) pts received T-PLD as 3rd line of chemotherapy (range: 2-8). The toxicity profile in the PFI subgroups was not different from that of the overall population. The number of pts with grade 3/4 hematological toxicities in the PFI 6-12 and PFI>12 cohorts was: neutropenia 29.6%/17.6%, febrile neutropenia 4.4%/3.3%, thrombocytopenia 7.7%/12.1%, anemia 5.5%/2.2%. Grade 3 hand and foot syndrome (1 pt) and mucositis (1 pt) were observed in the PFI 6-12 group. Increases in transaminases (grade 3/4) were experienced by 11 pts (10/1) in the PFI 6-12 group and by 5 pts (4/1) in the PFI>12 group. 3 patients in the PFI 6-12 group and 3 in the PFI>12 group discontinued treatment because of toxicities, 6 and 2 due to premature death. Partial and complete responses were achieved in 43 pts (PFI 6-12: n ¼ 26; PFI>12: n ¼ 17, p ¼ 0.82). Median PFS after T-PLD was 5.9 months (95%CI 4.9-6.7) in the PFI 6-12 group and 5.8 months (95%CI 3.7-8.5, logrank p ¼ 0.37) in the PFI>12 group. OS data were not mature at the time of this analysis. Conclusions: The safety profile of T-PLD when used in real-life management of nonselected OC pts is similar to that observed in clinical trials. T-PLD remains a valuable option to pts with both partially and fully platinum-sensitive disease. Clinical trial identification: NCT02163720 May 28, 2014 Legal entity responsible for the study: ARCAGY-GINECO Funding: PharmaMar Disclosure: L. Gladieff: PharmaMar fees in 2015 and 2016. All other authors have declared no conflicts of interest.

967P

An observational, multicenter, prospective study of trabectedin plus pegylated liposomal doxorubicin (PLD) in patients with platinumsensitive recurrent ovarian cancer (PSROC)

D. Reichert1, P. Wimberger2, U. Ringsdorf3, M. Krohn4, I. Runnebaum5 Gemeinschaftspraxis fu¨r Onkologie, Medizinische Studiengesellschaft NORD-WEST GmbH, Westerstede, Germany, 2Gynecology and Obstetrics, Carl-Gustav-Carus akologie und Geburtshilfe, University Dresden TU Dresden, Dresden, Germany, 3Gyn€ akologie und Lahn-Dill-Kliniken GmbH, Wetzlar, Germany, 4Frauenklinik - Gyn€ gyn€akologische Onkologie, Geburtshilfe und Perinatologie, Agaplesion 5 Diakonieklinikum Rotenburg GmbH, Rotenburg, Germany, Medical Oncology, Klinikum der Friedrich-Schiller-Universit€ at Jena - Klinik fu¨r Frauenheilkunde, Jena, Germany

1

Background: The OVA-YOND prospective non-interventional phase IV study evaluated trabectedin plus PLD in real-life clinical practice to assess the toxicity and efficacy of the combination when given in accordance with the marketing authorization to women with PSROC. Methods: Data from patients treated with PLD 30 mg/m2 and immediately followed by trabectedin 1.1 mg/m2 3-h i.v. infusion every 3 weeks have been collected. Results: From 02/2013 to 12/2016, 77 enrolled patients from 31 sites across Germany who received at least one cycle of trabectedin plus PLD were evaluated. All patients had a platinum-sensitive relapse with a median platinum-free interval of 12 months (range: 686 months). Median age of patients was 66 years (range: 40-78) and 80.5% had ECOG performance status 0/1. Serous carcinoma was the most prevalent histological type (n ¼ 54; 70.1%), tumor was localized at the ovary in 88.3% of patients and 38 patients (49.4%) were diagnosed with FIGO IIIC stage. Median number of trabectedin plus PLD cycles received per patient was 6, with 39 patients (50.6%) receiving 6 cycles and up to a maximum of 21 cycles. Median treatment duration was 4.24 months, mostly on an outpatient basis (66.7-100% of cases). Five patients (6.5%) had a complete response and 19 patients (24.7%) achieved a partial response for an ORR of 31.2% with a median duration of 6.25 months. Additionally, 16 patients (20.8%) had disease stabilization for a disease control rate of 51.9%. With 64 PFS events recorded, median PFS was 6.3 months (CI95%: 5.1-7.3) and median OS was 16.4 months (CI95%: 11.3-19.3). A total of 278 trabectedinrelated adverse events (TRAE) occurred in 57.1% of the patients who recovered in 70.9% cases. Most common grade 3/4 TRAE were leukopenia (18.2% of patients), neutropenia (15.6%), thrombocytopenia (9.1%), ALT (7.8%) and AST (6.5%) increase, and nausea/ vomiting (5.2% each). No grade 5 or unexpected TRAE occurred. Conclusions: Trabectedin plus PLD confer clinically meaningful benefit to patients with PSROC, being either comparable or better to those previously observed in selected population from clinical trials or other real-life studies, and with a manageable safety profile. Clinical trial identification: NCT01869400; OVA-YOND Legal entity responsible for the study: PharmaMar Funding: PharmaMar Disclosure: P. Wimberger: Honoraria for scientific talks from Pharma Mar. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

968P

PRO-002, a phase Ib dose-escalation study of NUC-1031 with carboplatin for recurrent ovarian cancer

S.P. Blagden1, A. Sukumaran2, L. Spiers1, V.K. Woodcock1, A. Lipplaa1, S. Nicum3, C. Gnanaranjan4, D.J. Harrison5, E. Ghazaly6 1 Early Phase Clinical Trials Unit, Churchill Hospital University of Oxford, Oxford, UK, 2 Surgery and Cancer, Imperial College, London, UK, 3Oncology, Churchill Hospital University of Oxford, Oxford, UK, 4Centre for Haemato-Oncology, Barts Cancer InstituteQueen Mary University of London, London, UK, 5School of Medicine, University of St Andrews, Fife, UK, 6Centre for Haemato-Oncology, Barts Cancer Institute, London, UK Background: NUC-1031 (Acelarin) is a first-in-class nucleotide analogue that overcomes key cancer cell resistance mechanisms to generate high intracellular levels of dFdCTP. In the PRO-001 study, single agent NUC-1031 achieved impressive clinical activity across multiple solid tumors. In this PRO-002 study, NUC-1031 was combined with carboplatin in 25 pts with recurrent ovarian cancer (OC). Methods: NUC-1031 was administered in a dose-escalation schedule as a 30-min infusion on days 1 & 8 with carboplatin on day 1, every 3 weeks for 6 cycles. Four dose cohorts of NUC-1031 (500, 625 and 750 mg/m2) with carboplatin (AUC4 or 5) were studied. The primary endpoint was to determine RP2D. Secondary endpoints included safety, RECIST response, PFS and PK/PD. Results: 25 pts (median age 64 yrs) participated, having received a median of 3 (range 2-6) prior lines of therapy. All had received prior platinum regimens (10 had prior carboplatin þ gemcitabine). 23 pts were response-evaluable, of whom 7 were platinum refractory, 10 were platinum resistant, 4 were partially platinum sensitive and 2 were platinum sensitive. Strong efficacy signals were achieved with 1 unconfirmed CR at the end of treatment (4%), 8 PRs (35%, 4 confirmed), 12 SDs (52%, 10 confirmed) and 2 PD (9%). Median PFS was 7.4 months (range 1.2-11.2). The combination regimen was well tolerated. 5 DLTs occurred in 4 pts: 2 Grade (G) 4 thrombocytopenia; 2 G3 fatigue and 1 G4 neutropenia. No DLTs occurred in the NUC-1031 500 mg/m2 þ carboplatin AUC5 group. 7 pts reported treatment-related SAEs. The most common SAE was thrombocytopenia, reported in 3 pts. NUC-1031 was stable in plasma (apparent t1/2¼3.8 h). Combination with carboplatin rapidly generated very high intracellular dFdCTP levels (Cmax¼12.1 lM/mg, TP/500 mg/m2 and Tmax¼30 min) that were maintained for 24 h. Conclusions: NUC-1031 combined with carboplatin is well tolerated and effective in recurrent platinum resistant and sensitive OC. dFdCTP levels were increased 25% by the addition of carboplatin. The RP2D was 500 mg/m2 NUC-1031 on days 1 & 8 with AUC5 carboplatin day 1, q21d. The efficacy and synergy of this schedule and the ability to deliver carboplatin at AUC5 makes this an attractive therapeutic combination. Clinical trial identification: NCT02303912 Legal entity responsible for the study: Nucana Biomed Funding: Nucana Biomed Disclosure: S.P. Blagden: Acted on advisory boards for Novartis and Clovis Oncology and has directorship of RNA Guardian Ltd. S. Nicum: Consultant/Advisory Boards: Roche, AstraZeneca, Abbvie, Tesaro, Clovis Speaker: AstraZeneca, Roche Clinical Trials Sponsored: AstraZeneca. D.J. Harrison: DJH has a research consultancy and research sponsored by Nucana Biomed. Family member has stock options with Nucana Biomed. All other authors have declared no conflicts of interest.

969P

Phase 1/2 trials of peptides cocktail vaccine for resistant cervical and ovarian cancer: Qol analysis

S. Takeuchi, M. Kagabu, T. Nagasawa, H. Omi, Y. Nitta, H. Itamochi, T. Sugiyama Obstetrics and Gynecology, Iwate Medical University School of Medicine, Morioka, Japan Background: We conducted phase 2(P2) studies of peptides vaccine (PV) immunotherapy for cervical (CC) and ovarian cancer (OC) using HLA-restricted tumor specific peptides and VEGF receptor 1 (R1)and 2(R2). All HLA0201(A02) and 2402 (A24) restricted peptides used were found from human CC and OC, respectively in GMP grade. The P1 showed feasible (presented in ECCO2013), and further P2 had completed in 66 accruals and the results showed efficacy (ASCO2015 5567). Approval of IRB had obtained. This time, QOL study using GOG FACT-O had analyzed and survival data were up-dated. Methods: Heavily treated CC and OC with A24 or A02 within ECOG PS 2 were candidate. Fully written-IC had obtained. PV cocktails were as follows: for OC of A24 comprised FOXM1, MELK, HJURP, VEGFR1(R1) and R2. As for OC A02, PV comprised HIG2, R1 and R2. For CC of A24, it comprised of FOXM1, MELK and HJURP, and CC with A02, it comprised of URLC10 and HIG2. Each peptide was mixed at a dose of 1mg with adjuvant, MONTANIDE in total 1ml. Vaccination schedule included 12 subcutaneous weekly injections followed with additional 8 administrations (adm.) in twoweek intervals and more 8 adm. in 1 month (m) intervals were performed. At pre- and every 8th post-adm., QOL sheet was obtained. Results: PV was well tolerated with no major adverse events (AEs), and during first 8 adm. AEs had caused by prior-chemotherapy/radiotherapy (after 1 month (m) washout) had been improved. As for QOL, physical-, functional- social-, and mental QOL were getting preferable during first 16 adm., however, during longer observation beyond 3 m they were deteriorated gradually due to disease progression. The final mOS of CC and OV was 4.9 m þ (0.6-76.6mþ) and 7.2m(1.1-59.1m), respectively. As for

doi:10.1093/annonc/mdx372 | 345

abstracts

Annals of Oncology

prognostic factors, significant benefits were seen among lower c-reactive protein (CRP), higher levels of peripheral lymphocytes count, younger age, and smaller residual tumors at baseline. They were strongly rerated to good QOL in p < 0.003 or more by Wilcoxon signed rank test. Conclusions: This PV immunotherapy had efficacy not only in safeness, response, and overall survival, but also in maintenance of QOL in this cohort of patients Further PV trials for other cohorts such as adjuvant therapy would be warranted. Clinical trial identification: UMIN:000003860, 000003862, 000003902, 000003902 Legal entity responsible for the study: Captivation Funding: None Disclosure: All authors have declared no conflicts of interest.

970P

A phase 1 study of selinexor (S) in combination with paclitaxel (P) and carboplatin (C) in patients (pts) with advanced ovarian (OC) or endometrial cancers (EC)

V. Makker, N. Boucicaut, K.A. Cadoo, R. Grisham, D.M. Hyman, R. O’Cearbhail, A. Snyder Charen, W. Tew, M. Martin, C. Aghajanian Medicine, Memorial Cloan Kettering Cancer Center, New York, NY, USA Background: Selinexor (S) is an oral, first in class, inhibitor of exportin 1 (XPO1). In a Phase II clinical trial of pts with relapsed ovary cancer (OC) andendometrial cancer (EC), single agent S, demonstrated anti-cancer activity. In addition, clinical exploratory analysis has demonstrated S target engagement and a relationship between baseline circulating tumor cells and duration of response. Here we report results of a phase 1 study evaluating safety/tolerability of S combined with C and P in pts with advanced OC and EC or carcinosarcomas. Methods: Patients (Pts) were enrolled using 3 þ 3 dose escalation design for each regimen (reg). All pts with OC received 1 prior platinum (plt) therapy. Ptswith EC could be chemotherapy naı¨ve or have received 1 prior plt therapy. Pts were enrolled to 1 of 4 regimens regardless of disease type as described in Table. Response was evaluated Q9 weeks (RECIST 1.1). Results: 16 pts (12 EC, 3 OC, 1 endometrial carcinoma) were enrolled. 1 drug related DLT of G3 syncope occurred on Reg 2. Most common G2 AEs were hyperglycemia (43.8%), leukopenia (43.8%), anemia (31.3%). Most common Grade 3 and 4 AEs were anemia (62.5%), neutropenia (37.5%), lymphopenia (43.8%), neutropenia (31.3%) thrombocytopenia (12.5%). 50% of evaluable pts on Reg 1 and 2 were dose reduced due to S toxicity. One dose reduction of S on Reg 3. There were no dose reductions on Reg 4. 13 pts were evauble for efficacy: 2 CRs, 10 PRs, and 1 SD. Time on study ranged from 2–10.8 mos with 3 pts still on study. Conclusions: Selinexor in combination with carboplatin and paclitacel (CP) chemotherapy in advanced OC, EC, and carcinosarcomas was well tolerated. The RP2Ds have been established at 30mg/m2twice weekly of S and 60 mg flat dose weekly in combination with CP chemotherapy. Given encouraging response, expansion cohorts for Regs 3/4 are planned. Frequent molecular alterations seen in the EC pts included: TP53, PIK3CA, and KRAS. Evaluation of S target engagement/correlatives of response will be discussed.

Table: 970P Regimen #

N

Regimen details

1

4

2

6

3

3

4

3

C AUC5 (day 1), P 175 mg/m2 (day 1) and S 30 mg/m2 (days 1, 4, 8, 11, 15, 18) C AUC5 (day 1), P 80 mg/m2 (days 1, 8, 15) and S 30 mg/m2 (days 1, 4, 8, 11, 15, 18) C AUC5 (day 1), P 80 mg/m2 (days 1, 8, 15) and S 60 mg (days 1, 8, 15) C AUC5 (day 1), P 175 mg/m2 (day 1) and S 60 mg (days 1, 8, 15)

Clinical trial identification: NCT02269293 Legal entity responsible for the study: Memorial Sloan Kettering Cancer Center Funding: Karyopharm Pharmaceuticals Disclosure: All authors have declared no conflicts of interest.

346 | Gynaecological cancers

971P

Pazopanib and oral cyclophosphamide in women with platinum resistant epithelial ovarian cancers

S. Gulia1, S. Gupta2, J. Ghosh2, J. Bajpai2, A. Maheshwari2, R. Kerkar3, S. Thumkur4, K. Deodhar5, M. Thakur6 1 Medical Oncology, Tata Memorial Centre, Mumbai, India, 2Medical Oncology, Tata Memorial Hospital Centre, Mumbai, India, 3Surgical Oncology, Tata Memorial Centre, Mumbai, India, 4Gynaec Ongology, Tata Memorial Centre, Mumbai, India, 5Pathology, Tata Memorial Centre, Mumbai, India, 6Radiology, Tata Memorial centre, Mumbai, India Background: Women with recurrent, multiply treated epithelial ovarian cancer (EOC) have unfavorable prognosis with limited treatment options after failure of platinum based regimens. Antiangiogenic therapies have shown some efficacy in these patients. We report here a retrospective analysis of women with recurrent, platinum resistant EOC treated with an oral regimen of anti-angiogenic agent Pazopanib and Cyclophosphamide. Methods: Women with histologically proven recurrent platinum-resistant EOC were treated with tablets pazopanib (600mg p.o. daily in two divided doses, 400 mg and 200 mg) and cyclophosphamide (50 mg p.o. daily for 14 days every 21 days) until disease progression or unacceptable toxicity. Response was evaluated radiologically every 12 weeks. Results: Eighteen patients (16 platinum resistant and 2 platinum refractory) were treated between April 2014 and April 2017 with a mean age of 50 (38-60) years and median 4 (2- 8) previous lines of chemotherapy, including three patients with progressive disease on bevacizumab. Patients received a median of 6 (2-28) cycles of pazopanib and cyclophosphamide with partial response in 8 (44%) patients (including 1 of 3 prior bevacizumab treated patients), stable disease in 5 (28%) and disease progression in 5 (28%) patients, as best response. The median progression-free survival was 5.0 months. Common adverse events were fatigue (50%), diarrhea (50%), elevated liver enzymes (43%), mucositis (61%), myelosuppression (28%), Skin toxicity (33%), hypertension (6%) and hair depigmentation (6%). Dose reduction due to toxicity was required in 11 (61%) patients and no patient stopped treatment due to toxicity. Conclusions: Pazopanib and oral cyclophosphamide is a well-tolerated regimen with clinically relevant benefit in platinum resistant, epithelial ovarian cancer patients. Clinical trial identification: This is a retrospective analysis of Platinum resistant epithelial ovarian cancer patients treated at our instittute. The approval for doing this analysis was taken from Insitute s ethics committee. Legal entity responsible for the study: Institutional Review Board Institutional Review Board, Tata Memorial Hospital, Mumbai, India Funding: None Disclosure: All authors have declared no conflicts of interest.

972P

Apatinib as a salvage treatment in gynecologic cancer patients failed from two or more lines of prior chemotherapy

C. Xie, X. Jin Radiotherapy and Chemotherapy, The 1st Afﬁliated Hospital of Wenhzou Medical University, Wenhou, China Background: Apatinib is an oral inhibitor of the vascular endothelial growth factor receptor (VEGFR)-2. There is currently no standard treatment for patients with gynecologic cancer who failed from 2 lines of chemotherapy. The purpose of this study is to evaluate the benefits and adverse events of apatinib in the treatment of patients with advanced cervical and ovarian cancer who failed from 2 lines of chemotherapy. Methods: Patients with advanced cervical and ovarian cancer received at least two lines of prior chemotherapy before being treated with apatinib were retrospectively reviewed between April 2015 and January 2017. All included patients received continuous apatinib treatment until disease progression, death, or intolerable toxicity. Prognosis and toxicities were evaluated by the Kaplan-Meier method and according to NCI-CTC 3.0, respectively. Results: Twenty-six patients were eligible (cervical cancer, n ¼ 12 (46.2%); ovarian cancer, n ¼ 14 (53.8%)). After apatinib dose adjustment, 14 patients (53.8%) received 500 mg/day, 8 received 250 mg/day, 3 received 425 mg/day, and one received 675 mg/ day. The median progression-free survival (PFS) of cervical and ovarian cancer was 8 months (95% confidence interval (CI): 3.83-12.17) and 4 months (95%CI:1.57-6.44), respectively. The objective response rates in cervical cancer and ovarian cancer were 50% (n ¼ 6/12) and 50% (n ¼ 7/14), respectively. The disease control rates were 100% (n ¼ 12/12) for cervical cancer and 71.4% (n ¼ 10/14) for ovarian cancer. No complete response was observed. The toxicities associated with apatinib were generally acceptable: eight patients (30.8%) developed grade 3/4 toxicity. The most common adverse events were hypertension (n ¼ 17; 65.4%), hand-foot syndrome (n ¼ 24, 92.3%), and mouth mucositis (n ¼ 20, 76.9%). Conclusions: Apatinib monotherapy could be a promising and tolerable treatment for patients with advanced/recurrent cervical and ovarian cancer who failed from two or more lines of chemotherapy. Legal entity responsible for the study: Congying Xie Funding: None Disclosure: All authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology

973P

A GINECO phase II study of Navitoclax (ABT 263) in women with platinum resistant/refractory recurrent ovarian cancer (ROC)

P.E. Brachet1, M. Fabbro2, A. Leary3, J. Medioni4, P. Follana5, A. Lesoin6, J-S. Frenel7, S. Abadie Lacourtoisie8, A. Floquet9, L. Gladieff10, B. You11, C. Gavoille12, E. Kalbacher13, M. Briand14, P-A. Just15, C. Blanc-Fournier16, A. Leconte17, J. Lequesne17, L. Poulain14, F. Joly Lobbedez1 1 Medical Oncology & Clinical Research & INSERM U1086 “ANTICIPE” (Interdisciplinary Research Unit for Cancers Prevention and Treatment, Axis BioTICLA “Biology and Innovative Therapeutics for Ovarian Cancers”), UNICANCER, Centre Franc¸ois Baclesse & Normandie Univ, UNICAEN, Caen, France, 2Medical Oncology, UNICANCER, ICM Regional Cancer Institute of Montpellier, Montpellier, France, 3Medical Oncology, UNICANCER, Institut Gustave Roussy, Villejuif, France, 4Medical Oncology, AP-HP, Hopital European George Pompidou, Paris, France, 5Medical Oncology, UNICANCER, Centre Antoine Lacassagne, Nice, France, 6Medical Oncology, UNICANCER, Centre Oscar Lambret, Lille, France, 7Medical Oncology, UNICANCER, ICO Institut de Cancerologie de l’Ouest Rene´ Gauducheau, Saint-Herblain, France, 8Medical Oncology, UNICANCER, Institut de Cance´rologie de l’Ouest site Paul Papin, Angers, France, 9 Medical Oncology, UNICANCER, Institute Bergonie´, Bordeaux, France, 10Medical Oncology, UNICANCER, Institut Claudius Regaud- IUCT-Oncopole, Toulouse, France, 11 Medical Oncology, Centre Hospitalier Lyon Sud, Pierre Be´nite, France, 12Medical Oncology, UNICANCER, Institut de Cance´rologie de Lorraine - Alexis Vautrin, Vandoeuvre Les Nancy, France, 13Medical Oncology, CHU Besanc¸on, Hoˆpital Jean Minjoz, Besanc¸on, France, 14INSERM U1086 “ANTICIPE” (Interdisciplinary Research Unit for Cancers Prevention and Treatment, Axis BioTICLA “Biology and Innovative Therapeutics for Ovarian Cancers”), UNICANCER, Centre Franc¸ois Baclesse & Normandie Univ, UNICAEN, Caen, France, 15Service de Pathologie, APHP, site Cochin, Paris, France, 16 Department of Anatomo-Pathology & INSERM U1086 “ANTICIPE” (Interdisciplinary Research Unit for Cancers Prevention and Treatment, Axis BioTICLA “Biology and Innovative Therapeutics for Ovarian Cancers”), UNICANCER, Centre Franc¸ois Baclesse & Normandie Univ, UNICAEN, Caen, France, 17Department of Clinical Research, UNICANCER, Centre Franc¸ois Baclesse, Caen, France Background: Among ovarian cancer patients with early relapse after platinum chemotherapy, there is no convincing active treatment. In preclinical studies, we previously demonstrated promising activity of Navitoclax (ABT-263), an anti-apoptotic inhibitor of Bcl-2 family, in ROC tumors, suggesting a potential action in platinum resistant patients. In this prospective multicentric phase II study, we evaluated the efficacy of Navitoclax monotherapy in heavily pretreated ROC patients. Methods: This study included high grade serous patients with platinum resistance. Navitoclax was orally administered at 150 mg/day during a lead in period (7 to 14 days) and then increased to 250 mg in the absence of dose-limiting thrombocytopenia (50%, and no adjuvant RT or CT were identified as adversely impacting RFS, DFS and OS. Conclusions: In FIGO stage III endometrial cancer patients, use of both CT and RT is associated with improved RFS, DFS and OS and therefore should be recommended in all eligible patients after resection. 5Y RFS, DFS and OS are similar across stages IIIA to IIIC2. Risk factors including age, high grade and deep myometrial invasion are independent predictors of survival. Legal entity responsible for the study: BC Cancer Agency Funding: None Disclosure: J.J. Ko: Received honorarium from Janssen, Astellas, Bristol-Myers Squib and Merck. Participated in the advisory board for Bristol-Myers Squib, Merck and AstraZeneca. A. Kumar: Participated in advisory boards for Celgene, Roche-Peru, and AMtene. All other authors have declared no conflicts of interest.

977P

OS (%) (p < 0.00001)

Is chemotherapy worthwhile in patients with FIGO stage 1B, lymph nodes negative, grade 3 endometrial cancer?

C. Fontanella1, S. Lepori1, A. Barcellini1, G. Maltese1, C. Andreetta2, E. Tripodi1, F. Martinelli1, A. Cerrotta1, G. Bogani1, A. Ditto1, M. Signorelli1, C. Scaffa1, C. Sacco3, F. Raspagliesi1, D. Lorusso1 1 Gynecologic Oncology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 2 Department of Oncology, University Hospital of Udine, Udine, Italy, 3Department of Oncology, AOU Santa Maria della Misericordia, Udine, Italy Background: FIGO stage 1b endometrial cancer represents a major treatment challenge and standard of care is still unclear. Methods: From March 1996 to March 2016, we retrospectively collected patients diagnosed with endometrial cancer stage Ib (invasion 50% of the myometrium, 2009 FIGO staging), lymph nodes negative after lymphadenectomy, and grade 3. We performed descriptive analysis and Kaplan Meier test using SPSS 20.0. Results: Overall, 39 consecutive patients have been collected (28 at the National Cancer Institute of Milan and 11 at the University Hospital of Udine). Median age was 65.8 years (range 35.6-84.9). Endometrioid adenocarcinoma was diagnosed in 32 patients (82.1%), 4 serous adenocarcinoma (10.3%), 2 papillary serous adenocarcinoma (5.1%), and 1 clear cell adenocarcinoma (2.6%). Taking into account only endometrioid histotype, 23 patients received adjuvant radiotherapy (RT): 13 patients (40.6%) received brachyRT, 7 patients (21.8%) received external RT, 3 patients (9.4%) received both; 13 patients underwent platinum-based adjuvant chemotherapy (CT): 7 patients only CT, 2 patients external RT followed by CT and 2 patients brachyRT followed by

348 | Gynaecological cancers

CT. After a median follow up of 45.8 months (range 27.3 -236.8), median disease-free survival was 23.3 months (range 4.7-157.4); 7 patients (21.9%) experienced disease relapse and 5 patients (15.6%) died due to endometrial cancer. Relapse rate was 21.7% in patients who received RT versus 22.2% who did not. To note, relapse rate was only 9.1% in patients who received CT versus 28.6% in patients who did not. Conclusions: According to our study, patients with stage 1b, node negative, grade 3 endometrioid endometrial cancer seems to derive a great benefit from adjuvant chemotherapy. This data needs to be further investigated in a large prospective clinical trial. Legal entity responsible for the study: Department of Gynecologic Oncology, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy Funding: None Disclosure: All authors have declared no conflicts of interest.

978P

Significance of MSH2 promoter methylation in endometrial cancer with MSH2 deficiency

J. Haraga1, T. Nagasaka2, K. Nakamura1, T. Haruma1, T. Nishida1, A. Nyuya2, K. Yasui2, T. Fujiwara2, A. Goel3, H. Masuyama1 1 Department of Obstetrics and Gynecology, Okayama University Hospital, Okayama, Japan, 2Department of Gastroenterological Surgery, Okayama University Hospital, Okayama, Japan, 3Center for Gastrointestinal Research, Center for Translational Genomics and Oncology, Baylor Scott & White Research Institute, Charles A. Sammons Cancer Center, Baylor University Medical Center, Dallas, TX, USA Background: Inactivation of MSH2 was frequently observed in endometrial cancer (EC) with microsatellite instability (MSI) or mismatch repair complex deficiency (dMMR). With respect to MSH2 deficiency (dMSH2), most of dMSH2 were caused by germline mutations in the MSH2 gene or EpCAM deletions. Meanwhile, heritable germline epimutations in MSH2 reported in a few Lynch syndrome families that lacked germline mutations in the MSH2 gene. We previously provided evidence for frequent MSH2 hypermethylation in Lynch syndrome colorectal tumors with dMSH2 and MSH2 methylation may serve as the “second hit” at the wild-type allele. Herein, we examined precise epigenetic alteration in MSH2 promoter and tried to reveal associations to family history of Lynch syndrome related tumors. Methods: We analyzed MSH2 promoter methylation status, as well as MLH1 methylation status, and expression status of the mismatch repair proteins (MLH1, MSH2, PMS2, and MSH6) by immunohistochemistry in 326 EC patients. DNA was extracted from formalin-fixed, paraffin-embedded tissue, and analyzed MSI status by four mononucleotide markers and both MLH1 and MSH2 promoter methylation status by a fluorescent quantitative bisulfite PCR assay. Results: MSI or dMMR was observed in 82 (25.2%) or 89 ECs (27.3%), respectively. ECs with dMSH2 were observed in 18 (5.5%) of 326 ECs (20.2% of dMMR). MSH2 promoter methylation was detected in 8 tumors (2.5% in 319 tumors excluding not available 7 ECs), and significantly correlated with dMSH2 (P ¼ 0.0072, Fisher’s exact probability test). Then, we also examined the family history of first-degree relatives. In this cohort, although patients with dMMR were significantly associated with family history of Lynch syndrome related tumor (P ¼ 0.0312), patients with this family history are more frequently observed in patients with dMSH2 (P ¼ 0.0052). Interestingly, patients with MSH2 promoter methylation were strongly associated with the family history of Lynch syndrome related tumor (P ¼ 0.0053), though patients with MLH1 methylation were not (P ¼ 0.8345). Conclusions: MSH2 methylation significantly correlated with ECs with dMSH2 and may have strong relation with family history of Lynch syndrome related tumor, suggesting it takes a role as “second hit” to the MSH2 gene. Legal entity responsible for the study: Takeshi Nagasaka Funding: None Disclosure: All authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 979P

Achievement of complete response (CR) in metastatic or recurrent cervical cancer (MRCC): Does it matter?

J.F. Grau Be´jar1, V. Rodriguez Freixinos2, L. Fari~ nas Madrid2, G. Villacampa Javierre3, A. Gil-Moreno4, R. Verges5, M.A. Pe´rez Benavente4, A. Garcıa6, R. Dienstmann7, A. Oaknin1 1 Medical Oncology, Vall d’Hebron University Hospital, Barcelona, Spain, 2Oncologıa Me´dica, Hospital Vall dHebron, Barcelona, Spain, 3Biostatistics, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 4Gynaecology, Vall d’Hebron University Hospital, Barcelona, Spain, 5Radiation Oncology, Vall d’Hebron University Hospital, Barcelona, Spain, 6Pathology, Vall d’Hebron University Hospital, Barcelona, Spain, 7Grupo de Oncology Data Science (ODysSey), Vall d’Hebron Institute of Oncology, Barcelona, Spain Background: MRCC is a devastating disease with poor long-term outcomes. Bevacizumab (BEV) added to chemotherapy (CT) improves significantly overall survival (OS) in MRCC patients (pts). Aim: to characterize clinic-pathologic features associated to CR and its impact on pts outcome. Methods: Single-institution chart review of MRCC pts who were treated with 1st line CT between 2005 and 2016. CR was defined by Response Evaluation Criteria in Solid Tumors (RECIST v1). The prognostic and predictive value of clinic-pathologic features, was evaluated. Results: Seventy-two pts (62% squamous; 30% adenocarcinoma; 8% others); with median age of 48 years (28-77) were selected. Forty-five pts (62%) had prior CT-radiation; 55 pts (79%) had recurrent/persistent disease (27 pts > 12 months (m) disease free interval) and 15 pts (21%) were stage IVb (90% visceral involvement). Moore risk distribution: 7/44/21 pts were high/medium/low risk, respectively. Eleven pts (15%) received BEV þ CT; 57 pts (79%) platinum-based-CT (PCT) (54% Cisplatin; 26% Carboplatin) and 4 (6%) non-PCT. After a median follow-up of 33 m, ORR 51%, median OS 13 m (9.5-NA) and median PFS 6 m (4.6 -7.7) were observed for overall population. Moore criteria correlated with prognosis (high-risk pts had significantly worse OS (HR ¼ 0.04, p 0.2 all comparisons). Higher ORR was observed among low and intermediate risk pts (51%, 67%; p ¼ 0.006). CRs occurred in 13/71 (18%) evaluable pts (BEV group 2/11; non-BEV 11/60, p ¼ 1). Clinic-pathologic features, including Moore criteria, did not correlate with CR in univariate analysis. Median time to CR was 3.5 m (3 –NA) and median duration of CR was 7 m (4.3–NA). Five pts (7%) had CR in the irradiated field. CR significantly impacted on PFS (9.7 m vs 4.7 m nonCR, p ¼ 0.002) and OS (31 m vs 9.5 m non-CR, p ¼ 0.001). Eight pts discontinued treatment due to toxicity. Conclusions: CR is a meaningful surrogate marker for improved PFS and OS in MRCC pts treated with 1st line CT, but no predictive features have been identified in our cohort. Moore prognostic score was validated in real-word practice but its capability guiding therapy needs further evaluation. Legal entity responsible for the study: Vall dHebron Institute of Oncology Funding: None Disclosure: A. Oaknin: Consulting or advisory role for PharmaMar, Clovis, Roche, AstraZeneca. All other authors have declared no conflicts of interest. 980P

Factors Negatively Affecting Voluntary Cervical Cancer Screening Among Educated Indians Above Poverty Line

A. Shukla, R.M. Dokala, J.R. Philomen Research & Consultancy Unit, Oncofocus Solutions, Bangalore, India Background: Cervical cancer is the second highest cause of cancer mortality in Indian women with 67,477 reported deaths in 2012 and 83,370 estimated deaths in 2020. Pap Smear, an affordable screening test, has shown to reduce mortality by 50 - 80% in various developed countries. However, low cervical cancer screening rate (3.1%) in India has resulted in about 70% of cases being diagnosed at an advanced stage (stage III or IV). The aim of this study is to understand reasons behind lack of voluntary testing among those educated Indians who are above poverty line and who are aware of cervical cancer being a preventable disease. Methods: We designed a two-part web-based questionnaire containing 18 questions (90% multiple choice questions). While the first part was designed to capture demographic attributes of the participants, the second part aimed to understand reasons behind low screening levels. The study was distributed between 1st of January 2017 to 30th of April 2017 through social media. Results: We received a total of 212 responses. After excluding participants who are not currently residing in India or who did not complete the survey, we had 167 evaluable responses. Notably, about 50% (n ¼ 84) of valid participants were aware of cervical cancer, indicating a decent level of awareness among the evaluable population. Among respondents who were aware of cervical cancer, 75% (n ¼ 63) were aware that cervical cancer is preventable by regular screening. However, only 22% of 63 respondents (n ¼ 14) underwent or took their family members for cervical cancer screening. Out of the 49 participants who did not get tested, despite being aware that cervical cancer is preventable, 57% (n ¼ 28) stated time, 18% (n ¼ 9) stated lack of access, and 4% (n ¼ 2) mentioned affordability as a constraint. The remaining 21% gave other reasons most of which are related to the belief that they or their family members have low probability of falling victim to cervical cancer. Conclusions: Time constraint emerged as the predominant reason for low cervical cancer screening levels among educated Indians who are above poverty line. We propose a

Volume 28 | Supplement 5 | September 2017

proactive approach wherein stakeholders organise well-advertised, easily-accessible screening camps in the residential areas during weekends. Legal entity responsible for the study: Oncofocus Solutions Funding: Oncofocus Solutions Disclosure: A. Shukla, R.M. Dokala, J.R. Philomen: Acts as an advisory firm offering research and consulting services and regularly interact with health care professionals to understand their evolving views. This research work has no commercial interest and is part of our social outreach initiative.

981P

Reactive stroma mediates CD81 T cell spatial distribution and function in ovarian cancer

Y. Wang1, L. Ryner1, A.R. Udyavar2, M. Desbois1, C. Kozlowski3, C-W. Chang4, Y. Guan1, S. Lu1, H. Koeppen5, J. Ziai5, R. Bourgon2, P.S. Hegde1 1 Oncology Biomarker, Genentech Inc. - Roche - USA, South San Francisco, CA, USA, 2 Bioinformatics, Genentech Inc. - Roche - USA, South San Francisco, CA, USA, 3Digital Patholody, Genentech Inc. - Roche - USA, South San Francisco, CA, USA, 4Biostatistics, Genentech Inc. - Roche - USA, South San Francisco, CA, USA, 5Pathology, Genentech Inc. - Roche - USA, South San Francisco, CA, USA Background: Close proximity between cytotoxic T cells and tumor cells is key to effective immunotherapy. Ovarian cancer exhibits diverse immune phenotypes with distinct prevalence and spatial localization of CD8þ T cells. This study is aimed to characterize the molecular mechanisms orchestrating the localization and function of CD8þ T Cells in ovarian cancer. Methods: CD8 IHC and RNAseq were performed on 277 ovarian tumor tissues from ICON7 phase 3 trial. CD8 T-cells in tumor vs. stromal area was assessed by digital pathology. A Random Forest regression model was constructed to identify molecular features associated with enumeration or spatial localization of CD8þ T cells. In situ validation was performed by MHCI IHC and FAP RNAish. Functional role of ovarian fibroblasts was characterized by ex vivo T cell function assays. Results: We identified three main immune phenotypes, including T-cell infiltrated, T-cell excluded, and immune desert. The immune phenotypes are highly associated with prognosis and the molecular subtypes of ovarian cancer. The T-cell infiltrated phenotype is denoted by high expression of T-effector signatures and antigen presentation machinery. The T-cell excluded phenotype showed similar expression of T-effector signatures as the T-cell infiltrated phenotype, however, most of the CD8þ T-cells were excluded from the tumor bed. The T-cell excluded phenotype showed high expression of the reactive stroma signatures (i.e. FAP), and low expression of class I antigen presentation genes. Lastly, the immune desert phenotype featured low prevalence of CD8þ T-cells, and high expression of neuroendocrine and metabolic pathways. In situ analysis confirmed the two key molecular features associated with the T-cell excluded phenotype: 1) loss of the MHC I expression in the tumor compartment, and 2) high FAP expression in CAFs. Co-culturing of ovarian fibroblast cells with T-cells resulted in reduced T-cell activation and proliferation. Conclusions: Our study uncovered key molecular mechanisms mediating the interplay between CD8þ T cell localization and function in ovarian cancer. Our findings underscore the potential of targeting reactive stroma as a novel therapeutic strategy to optimize immunotherapy for ovarian cancer patients. Legal entity responsible for the study: Genentech Funding: Genentech Disclosure: All authors have declared no conflicts of interest.

982P

Phase II study of the safety and efficacy of oral capecitabine in patients with platinum-pretreated advanced or recurrent cervical carcinoma

S. Lepori, C. Fontanella, G. Maltese, E. Tripodi, F. Martinelli, G. Bogani, A. Ditto, M. Signorelli, C. Scaffa, F. Raspagliesi, D. Lorusso Gynecologic Oncology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy Background: Cervical cancer is underrepresented in the gynecological clinical research. The objective of this observational study was to evaluate the activity and the safety of capecitabine in patients with platinum-pretreated recurrent cervical carcinoma. Methods: In this phase II study we enrolled patients with advanced or recurrent cervical carcinoma pretreated with platinum-based therapy. All patients signed an informed consent and were treated at the Gynecological Units of the IRCCS National Cancer Institute of Milan (Italy). All patients received a starting dose of oral capecitabine 1250 mg/m2 twice a day continuously from day 1 to day 14 every 21 days, dose reduction to 1000 mg/m2 twice a day was permitted due to adverse events (AE). We used Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 to evaluate response to therapy and Common Terminology Criteria for Adverse Events (CTCAE) version 4.0 to evaluate adverse events. We performed descriptive analysis and Kaplan Meier test using SPSS 20.0. Results: From December 2013 to January 2017, we enrolled 20 patients with advanced or recurrence cervical carcinoma, already exposed to platinum, to received oral capecitabine. All patients receive a combination of carboplatin plus paclitaxel as first-line therapy for advanced/recurrent disease. Median age at the first capecitabine administration was 56.9 years (range from 27 to 82 years). After three cycles of oral capecitabine the clinical benefit rate (CBR) was 60.0% (5.0% of CR, 30.0% of PR and 25.0% of SD). No grade 3 or worse adverse events were reported. CBR was 88.8% in adenocarcinomas

doi:10.1093/annonc/mdx372 | 349

abstracts versus 36.4% in squamous cell carcinomas (P ¼ 0.067). The most frequent grade 1 or 2 adverse events were fatigue (50%), hand-foot syndrome (38.9%) and diarrhea (22.2%). Conclusions: Our study suggests that oral capecitabine should be considered an active and safe treatment in patients with platinum-pretreated advanced or recurrent cervical carcinoma. A greater activity has been documented in patients with adenocarcinomas compared with squamous cell carcinomas. Legal entity responsible for the study: Not applicable Funding: None Disclosure: All authors have declared no conflicts of interest. 983P

Investigation of the clinicopathological features of vulva cancer: a retrospective survey of the JGOG Net Work study

S. Nishio1, T. Shibata2, S. Yamaguchi3, H. Kanao4, A. Kojima5, M. Takekuma6, A. Tozawa7, H. Tokunaga8, E. Miyagi9, H. Kato10, K. Kurihara11, T. Yanase12, K. Ushijima1, M. Mikami13, T. Sugiyama14 1 Obstetrics and Gynecology, Kurume University, Kurume, Japan, 2Molecular Life Sciences, Tokai University, Isehara, Japan, 3Gynecology, Hyogo Cancer Center, Akashi, Japan, 4Gynecology, Cancer Institute Hospital, Tokyo, Japan, 5Obstetrics and Gynecology, Iwate Medical University, Morioka, Japan, 6Gynecologic Oncology, Shizuoka Cancer Center, Shizuoka, Japan, 7Obstetrics and Gynecology, St. Marianna University, Kawasaki, Japan, 8Obstetrics and Gynecology, Tohoku University Hospital, Sendai, Japan, 9Obstetrics and Gynecology, Yokohama City University Hospital, Yokohama, Japan, 10Gynecology, Kanagawa Cancer Center, Yokohama, Japan, 11 Gynecology, Saitama Cancer Center, Saitama, Japan, 12Gynecology, Niigata Cancer Center Hospital, Niigata, Japan, 13Obstetrics and Gynecology, Tokai Univeresity Hospital, Isehara, Japan, 14Department of Obstetrics and Gynecology, Iwate Medical University School of Medicine, Morioka, Japan Background: Vulvar cancer is a rare malignancy in women. During the past 30 years, large surveys of vulva cancer have not been performed in Japan. We therefore conducted a multicenter study to clarify the clinicopathological features of vulva cancer in Japan (UMIN000017080). Methods: In this multicenter retrospective cohort study, the clinical data of patients with vulva cancer were surveyed. The medical records of patients with vulvar cancer patients treated between 2001 and 2010 were retrospectively reviewed after obtaining approval from the Institutional Review Board of each institution. Survival analysis was performed using Kaplan-Meier curves. The effects of the clinical factors on OS were investigated using a Cox regression model. Results: A total of 1082 patients treated in 108 centers were studied. The median age was 72 years (range, 20 to 96). The disease stage was stage I in 415 patients (38.3%), stage II in 249 (23%), stage III in 255 (23.6%), and stage IV in 163 (15.1%) (FIGO 2009). The diagnosis was squamous cell carcinoma in 779 patients (72%), Paget’s disease in 158 (14.6%), adenocarcinoma in 63 (5.8%), and others in 82 (7.6%). Positive lymph nodes were found in 237 patients (21.9%). The median tumor diameter was 35 mm (range, 1 to 180). The 5-year overall survival was 86% in stage I, 74.7% in stage II, 48.2% in stage III, and 39.3% in stage IV (P < 0.001), and that according to histology was 63.9% in squamous cell carcinoma, 57.1% in adenocarcinoma, 79.7% in Paget’s disease, and 85.4% in others. The hazard ratio was 0.51 in patients with a histology of Paget’s disease or others (vs. squamous cell carcinoma or adenocarcinoma; P ¼ 0.001; 95% CI, 0.35-0.75), 2.14 in patients with a number of positive lymph nodes 2 or more (vs. 0 or 1; P < 0.001; 95% CI, 1.50-43.05), 2.10 in patients with a tumor diameter of 35mm (vs. < 35mm; P ¼ 0.001; 95% CI, 1.36-3.25). Conclusions: Treatment outcomes in Japanese patients with vulvar cancer were similar to those reported previously. However, squamous-cell carcinoma, adenocarcinoma, positive lymph nodes, and bulky tumors were associated with poor outcomes. Multidisciplinary treatment might be required in patients with these characteristics. Clinical trial identification: Registry Name: UMIN Clinical Trials Registry Registration Number: UMIN000017080 Legal entity responsible for the study: No Funding: None Disclosure: All authors have declared no conflicts of interest.

984TiP

Japan CHARLOTTE: Characterizing the cross-sectional approach to ovarian cancer: Genetic testing of BRCA

T. Sugiyama1, D. Aoki2, T. Enomoto3, N. Takeshima4, Y. Watanabe5, K. Fujikawa6, T. Jinnai6, J. Kigawa7 1 Department of Obstetrics and Gynecology, Iwate Medical University School of Medicine, Morioka, Japan, 2Department of Obstetrics and Gynecology, Keio University School of Medicine, Tokyo, Japan, 3Department of Obstetrics and Gynecology, Niigata University School of Medicine, Niigata, Japan, 4Department of Gynecology, Cancer Institute Hospital, Tokyo, Japan, 5Department of Obstetrics and Gynecology, Tohoku Medical and Pharmaceutical University, Sendai, Japan, 6Department of Medical, AstraZeneca K.K., Osaka, Japan, 7Department of Obstetrics and Gynecology, Matsue City Hospital, Matsue, Japan

Annals of Oncology BRCA2, are associated with the majority of HBOC cases. Information on the frequency of gBRCA1 and gBRCA2 (gBRCA1/2) mutations in Japanese patients with ovarian cancer is limited. In a recent study of 95 unselected Japanese ovarian cancer patients, 12 (12.6%) had gBRCA1/2 mutations. There is a need for further examination of the prevalence of gBRCA1/2 mutations in a large number of Japanese patients. The Japan CHARLOTTE study is the first epidemiological survey in a large number of Japanese ovarian cancer patients. The primary objective is to examine the frequency of gBRCA1/2 mutations among newly diagnosed ovarian cancer patients in Japan. The secondary objectives are to examine the frequency of gBRCA1/2 mutations in subpopulations (e.g. histological subtype, family history), and to evaluate patient satisfaction with the explanation of BRCA genetic testing. Trial design: Japanese women aged 20 years with newly diagnosed, histologically confirmed epithelial ovarian cancer, primary peritoneal cancer or fallopian tube cancer, who provide written consent for study participation within 60 days of diagnosis, are eligible. Patients who have an acute or chronic disease or mental illness that could affect the study results as judged by the attending physician will be excluded. The following data will be collected from the medical records: demographics, past medical history and medication use, reproductive history and menopausal status; diagnosis and treatment of the gynecological cancer including staging based on the FIGO stage; and CA-125 levels. A detailed family history of cancer will be obtained by interview at consultation and a blood sample will be taken for BRCA genetic testing. Pathological slides will be sent to a central laboratory to confirm the pathological diagnosis. A questionnaire will be administered to assess patient satisfaction with the explanation of BRCA genetic testing. It is anticipated that 600 patients will be enrolled, with the involvement of around 50 institutions. Clinical trial identification: UMIN000025597 Legal entity responsible for the study: AstraZeneca K.K. Funding: AstraZeneca K.K. Disclosure: T. Sugiyama: Lecture’s fee from Chugai Pharmaceutical Company. D. Aoki: Lecture’s Fee from AstraZeneca K.K., Consulting fee/honorarium from AstraZeneca K.K. and FALCO biosystems Ltd. T. Enomoto: Lecture’s Fee from AstraZeneca, Chugai, Mochida, and Taisho-Toyama. K. Fujikawa, T. Jinnai: Employee of AstraZeneca K.K. All other authors have declared no conflicts of interest.

985TiP

IMagyn050 / GOG3015 / ENGOT-ov39: A randomized, double-blind, phase III study of atezolizumab vs placebo combined with chemotherapy 1 bevacizumab in stage III-IV ovarian, fallopian tube & peritoneal cancers (OC)

K.N. Moore1, A. Okamoto2, F. Wu3, Y.G. Lin4, S. Pignata5 Department of Obstetrics and Gynecology, University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA, 2OB/GYN, The Jikei University School of Medicine, Tokyo, Japan, 3Product Development Biometrics, Roche, Beijing, China, 4Product Development Oncology, Genentech, Inc., South San Francisco, CA, USA, 5Urology and Gynecology, Istituto Nazionale Tumori “Fondazione G. Pascale” IRCCS, Naples, Italy

1

Background: Despite surgical advances and initial responses to chemo, OC has the highest mortality among gynecologic cancers, underscoring the need to improve longterm outcomes over standard platinum/taxane regimens. OC is a VEGF-driven tumor, susceptible to both the anti-angiogenic and immunomodulatory activities of bev. The immune response may also be enhanced by blocking the PD-L1/PD-1 and PD-L1/B7.1 pathways with atezo (anti–PD-L1), further potentiating anti-cancer T-cell activity. Atezo has shown safe and durable clinical benefit as a mono- and combinatorial therapy in human cancers, including OC. Here we present the first study in OC assessing dual immunomodulation with bev þ atezo with standard chemo. Trial design: IMagyn050 (NCT03038100) will enroll 1300 newly diagnosed stage III-IV epithelial OC patients (pts) globally. Pts must have ECOG PS 2, evaluable tissue for PDL1 testing (VENTANA SP142 assay) and will either have gross residual disease postoperatively (primary surgery group) or receive neoadjuvant (neo) therapy followed by surgery then adjuvant (adj) therapy (neo group). A concurrent phase (IV AUC6 carboplatin [Cb] þ 175 mg/m2 paclitaxel [pac] þ 15 mg/kg bev þ 1200 mg atezo/PL 6 cycles [C]) will be followed by maintenance bev þ atezo/PL on 21-d C for a total of 22 C atezo/PL (Table), with pts treated until completion of maintenance therapy, toxicity or recurrence, whichever occurs first. The neo group will undergo neo therapy followed by interval surgery between C3-4 then adj therapy followed by maintenance bev þ atezo/PL. Stratification factors are stage and/or residual disease, ECOG PS, PD-L1 status and treatment approach. Co-primary endpoints PFS and OS will be assessed in all and PD-L1positive ( 1%) pts. Safety, efficacy, PROs and translational data will also be evaluated.

Table: 985TiP Treatment Groups (1:1 randomization) Primary surgery Neo

Concurrent

Maintenance

C1: Cb þ pac þ atezo/PL C2-6: C7 onward: bev þ Cb þ pac þ bev þ atezo/PL atezo/PL C1-2 and 5-6: Cb þ pac þ bev þ C7 onward: bev þ atezo/PL C3-4: Cb þ pac þ atezo/PL atezo/PL

Background: Approximately 5–10% of breast and ovarian cancers are inherited, a condition known as hereditary breast and ovarian cancer (HBOC). Two genes, BRCA1 and

350 | Gynaecological cancers

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Clinical trial identification: NCT03038100 Legal entity responsible for the study: F. Hoffman-La Roche Ltd. Funding: F. Hoffman-La Roche Ltd. Disclosure: K.N. Moore: Serve on advisory boards for AstraZeneca, Advaxis, Clovis, Tesaro, Immunogen, Genentech/Roche, VBL Therapeutics; Serve on steering committees for Advaxis, Clovis And Immunogen. F. Wu: Roche employee. Y.G. Lin: Genentech-Roche employee. S. Pignata: Honorarium by Roche. All other authors have declared no conflicts of interest.

986TiP

A randomized, double-blind, placebo-controlled multicenter phase 3 trial of niraparib maintenance treatment in patients with advanced ovarian cancer following frontline chemotherapy

A. Gonzalez Martin1, L.A. Rojas2, P.S. Braly3, J. Barter4, D.M. O’Malley5, A.M. Oza6, A.F. Haggerty7, C. Vulsteke8, D.M. Provencher9, W. Graybill10, Y. Li11, I.A. Malinowska12, M.R. Mirza13, I. Vergote14, B. Pothuri15, B.J. Monk16 1 Servicio de Oncologia Medica, Grupo Espa~ nol de Investigaci on en C ancer de Ovario (GEICO) & MD Anderson Cancer Center Center Espana, Madrid, Spain, 2Gynecologic 3 Oncology, Avera Cancer Institute, Sioux Falls, SD, USA, Gynecologic Oncology, Women’s Cancer Care, Covington, LA, USA, 4Gynecologic Oncology, Holy Cross Hospital, Silver Spring, MD, USA, 5Gynecologic Oncology, The Ohio State University Medical Center, Columbus, OH, USA, 6Division of Medical Oncology and Hematology, Princess Margaret Cancer Centre, Toronto, ON, Canada, 7Gynecologic Oncology, The University of Pennsylvania, Philadelphia, PA, USA, 8Molecular Imaging, Pathology, Radiotherapy & Oncology, Center for Oncological Research, Antwerp University, Antwerp, Belgium, 9Obstetrics and Gynecology, Centre Hospitalier de L’Universite de Montreal (CHUM), Montreal, QC, Canada, 10Gynecologic Oncology, Medical University of South Carolina, Charleston, SC, USA, 11Biostatistics, Tesaro, Inc., Waltham, MA, USA, 12 Clinical Science, Tesaro, Inc., Waltham, MA, USA, 13Oncology, Nordic Society of Gynecologic Oncology (NSGO) and Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark, 14Obstetrics and Gynaecology, Belgian Gynecologic Oncology Group (BGOG) and University of Leuven, Leuven Cancer Institute, Leuven, Belgium, 15 Obstetrics & Gynecology, NYU Langone Medical Center, New York, NY, USA, 16 Obstetrics and Gynecology, Arizona Oncology (US Oncology Network), University of Arizona College of Medicine, Creighton University School of Medicine at St. Joseph’s Hospital, Phoenix, AZ, USA Background: Niraparib (ZEJULATM) is a selective poly (ADP-ribose) polymerase (PARP) 1/2 inhibitor. In preclinical studies, niraparib concentrated in the tumor vs plasma, delivering 90% durable PARP 1/2 inhibition and a persistent antitumor effect. In the ENGOT-OV16/NOVA trial, niraparib demonstrated clinical efficacy in patients with recurrent ovarian cancer (OC) following complete response (CR) or partial response (PR) to platinum-based chemotherapy (PBC) regardless of BRCA mutation or homologous recombination deficiency (HRD) status. Trial design: The primary objective of the ongoing ENGOT-OV26/PRIMA trial is efficacy (measured as progression-free survival) of niraparib vs placebo in advanced OC patients with CR or PR following frontline PBC. Secondary objectives include overall survival, patient-reported outcomes (PROs), time to first subsequent therapy, time to progression on the next anticancer therapy, and safety and tolerability of niraparib. Target enrollment is 330 patients with stage III or IV OC with PR or CR after PBC. Eligibility criteria include all patients with stage IV disease and patients with stage III disease who were treated with neoadjuvant chemotherapy followed by interval debulking surgery or who have either inoperable disease or visible residual disease after primary debulking surgery. Patients are stratified based on HRD status (HRD positive, including the known deleterious BRCA mutations gBRCAmut or sBRCAmut/HRD negative), neoadjuvant chemotherapy (yes/no), and best response to PBC (CR/PR). Patients are randomized 2:1 to oral niraparib 300 mg or matched placebo once daily in 28-day cycles. PRO data will be collected. Clinical trial identification: NCT02655016 Legal entity responsible for the study: Tesaro, Inc. Funding: Tesaro, Inc. Disclosure: A. Gonzalez Martin: Consulting or Advisory Role: Roche, PharmaMar, AstraZeneca; Speakers’ Bureau: Roche, PharmaMar, AstraZeneca; Travel, Accommodations, Expenses: Roche, PharmaMar, AstraZeneca. L.A. Rojas: Research funding: Pfizer. P.S. Braly: Speaker’s Bureau: Clovis Oncology, Roche, Myriad Genetics; Research Funding: AstraZeneca, Tesaro, Janssen, PharmaMar. J. Barter: Speakers’ Bureau: Baxter. D.M. O’Malley: Consulting or Advisory: Janssen Oncology, Eisai, AstraZeneca, Clovis Oncology, Genentech/Roche, Amgen, Tesaro, Novocure. A.M. Oza: Consulting or Advisory Role: Amgen, Verastem, Clovis Oncology, Immunovaccine; Travel, Accomodations, Expenses: AstraZeneca Honoraria: WebRx. C. Vulsteke: Consulting or Advisory Role: Novartis, LEO Pharma, Roche; Travel, Accommodations, Expenses: Novartis, Pfizer, Roche. D.M. Provencher: Consulting or Advisory Role: AstraZeneca; Speakers’ Bureau: AstraZeneca. W. Graybill: Consulting or Advisory Role: Tesaro. Y. Li: Employment: Tesaro; Stock and other ownership interests: Tesaro. I.A. Malinowska: Employment: Tesaro; Stock and Other Ownership Interests: Tesaro. M.R. Mirza: Consulting or Advisory Role: Clovis Oncology, AstraZeneca, Tesaro. I. Vergote: Travel, Accomodations, Expenses: GCI Health, Roche,

Volume 28 | Supplement 5 | September 2017

Oasmia Pharmaceutical AB, PharmaMar, AstraZeneca. B. Pothuri: Research funding: Tesaro, Caris Life Sciences, Celgene, Clovis Oncology. B.J. Monk: Consulting/ Advisory: GSK, Merck, Tesaro, Roche/GNE, AZ, Gradalis, Advaxis, Verastem, Cerulean, Amgen, Vermillion, Immunogen, Bayer, NuCana BioMed, Insys Therapeutics, Clovis Onc, Oxigene, Pfizer; Speakers Bureau: Roche/GNE, AZ, Janssen, Myriad Genetics. All other authors have declared no conflicts of interest.

987TiP

OReO/ENGOT Ov-38: A phase IIIb trial of olaparib maintenance retreatment in patients with epithelial ovarian cancer

E. Pujade-Lauraine1, N. Colombo2, R. Glasspool3, B. Asselain4, T. Huzarski5, J. Korach6, F. Marme7, M.R. Mirza8, A. Redondo9, G. Scambia10, C. Blakeley11, A. Milner11, F. Selle12, I. Vergote13 1 Universite´ Paris Descartes, AP-HP and GINECO, Paris, France, 2University of MilanBicocca, Istituto Europeo Oncologia and MANGO, Milan, Italy, 3Beatson West of Scotland Cancer Centre and NCRI and SGCTG, Glasgow, UK, 4Institut Curie and GINECO, Paris, France, 5Pomeranian Medical University, Szczecin, Poland, 6Sheba Medical Centre, Tel Aviv University and ISGO, Tel Hashomer, Israel, 7University Hospital Heidelberg and AGO, Heidelberg, Germany, 8Rigshospitalet-Copenhagen University Hospital, Copenhagen, Denmark, 9Hospital Universitario La Paz and GEICO, Madrid, a Cattolica del Sacro Cuore di Roma, Rome, Italy, 11AstraZeneca, Spain, 10Universit Cambridge, UK, 12Hoˆpital Tenon and GINECO, Paris, France, 13University Hospital Leuven and BGOG, Leuven, Belgium Background: A significant progression-free survival (PFS) improvement in all-comer patients (pts) with platinum-sensitive, relapsed (PSR), high-grade serous ovarian cancer supported the approval of olaparib (LynparzaTM) capsules as maintenance monotherapy (Ledermann et al NEJM 2012); these results were confirmed using the new olaparib tablet formulation in pts with a germline BRCA1/2 mutation (BRCAm) (Pujade-Lauraine et al SGO 2017). Pts with epithelial ovarian cancer (EOC) often retain platinum sensitivity despite progression on PARP inhibitor (PARPi) maintenance therapy. After relapse, it is unknown whether such pts could derive clinical benefit from olaparib retreatment. The OReO/ENGOT Ov-38 trial (NCT03106987; D0816C00014) will evaluate efficacy and safety of maintenance retreatment with olaparib tablets in allcomer pts with EOC. Trial design: OReO/ENGOT Ov-38 is a randomized, placebo-controlled multicentre trial of olaparib maintenance retreatment in pts with non-mucinous EOC, and a complete or partial response to their most recent platinum-based chemotherapy. Eligibility requires prior receipt of maintenance PARPi therapy, or prior participation by pts in a study with a PARPi experimental arm. Randomization 2:1 to olaparib (300 mg twice daily tablets) or matching placebo will be split across two cohorts (approximately 416 pts): pts with a known BRCAm in cohort one; BRCA wild-type pts in cohort two. Pts with a known BRCAm (germline or somatic) must provide a tumour sample (archival or fresh) and a blood sample for post-analysis BRCA testing. Primary endpoint is investigator-assessed PFS (RECIST v1.1). Cohorts are independently powered at 80% to detect a PFS benefit (assuming a hazard ratio [olaparib:placebo] of 0.5 [cohort one] or 0.65 [cohort two]) at the two-sided 5% level. Table lists the secondary endpoints, including outcome measures for health-related quality of life. Enrolment began in Q2 2017.

Table: 987TiP Secondary endpoints OS TTP by GCIG criteria TDT TFST TSST PRO measures for HRQoL: – Change from baseline in the TOI of the FACT-O – Proportion of pts with an improved PRO score – Best overall response (improved, no change, worsened, other) – PRO deterioration-free survival Safety FACT-O, Functional Assessment of Cancer Therapy – Ovarian; GCIG, Gynecologic Cancer InterGroup; HRQoL, health-related quality of life; OS, overall survival; PRO, Patient Reported Outcome; TDT, time from randomization to study treatment discontinuation, or death; TFST, time from randomization to ﬁrst subsequent treatment commencement, or death; TOI, Trial Outcome Index; TSST, time from randomization to second subsequent treatment commencement, or death; TTP, time to progression

doi:10.1093/annonc/mdx372 | 351

abstracts Clinical trial identification: NCT03106987 Legal entity responsible for the study: AstraZeneca and ENGOT Funding: AstraZeneca Disclosure: E. Pujade-Lauraine: Received advisory board membership and honoraria from AstraZeneca and Pfizer, and advisory board membership, honoraria and speakers’ bureau membership from Roche. N. Colombo: Received a grant from AstraZeneca, and personal fees from AstraZeneca, Roche, Pharmamar, Clovis, Pfizer and Tesaro. R. Glasspool: AstraZeneca: travel, registration, and accommodation for a noncompensated advisory board in June 2016, and for ASCO 2017. Advisory boards for Clovis and Tesaro (2016) and a speaker at a ROCHE meeting in March 2017. F. Marme: Received honoraria for Scientific advisory boards etc from: Roche, AstraZeneca, Novartis, Pfizer, PharmaMar, Eisai, Genmoic Health. M.R. Mirza: Board of Directors: Karyopharm., Metamark Genetics, Sera Prognostics Inc. Consultant/Advisory: Advaxis, AstraZeneca, Boehringer Ingelheim, Cerulean, Clovis, Genmab, Karyopharm, Novocure, Pfizer, Roche, Tesaro Study grants: AstraZeneca, Boehringer Ingelheim, Clovis, Pfizer, Roche, Tesaro. A. Redondo: AstraZeneca: Honoraria and travel expenses. C. Blakeley, A. Milner: Employed as a contractor for AstraZeneca, but does not own stock. All other authors have declared no conflicts of interest.

988TiP

ARIEL4: An international, randomised phase 3 study of the PARP inhibitor rucaparib vs chemotherapy for the treatment of BRCAmutated, relapsed, high-grade ovarian cancer

R.S. Kristeleit1, D. Lorusso2, A. Oaknin3, T. Safra4, E.M. Swisher5, I.M. Bondarenko6, oka10, L.S. Viola11, C. Tankersley12, T. Huzarski7, J. Klat8, V. Moiseyenko9, R.L. P L. Maloney12, S. Goble13, C. Unger14, A. Dowson12, H. Giordano12, A.M. Oza15 1 Oncology, University College London Cancer Institute, London, UK, 2Gynecologic Oncology, MITO and Unit a di Ginecologia Oncologica, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy, 3Medical Oncology Department, Vall d’Hebron University Hospital, Vall d’Hebron Institute of Oncology (VHIO), Barcelona, Spain, 4 Oncology and Gynecology Service, Sackler School of Medicine, Tel Aviv University & Tel Aviv Sourasky Medical Center, Tel Aviv, Israel, 5Department of Obstetrics and Gynecology, University of Washington, Seattle, WA, USA, 6Oncology, Dnipropetrovsk Medical Academy, City Multiple-Discipline Clinical Hospital, Dnepropetrovsk, Ukraine, 7 Department of Medicine, Private Health Care Innovative Medicine, Grzepnica, Poland, 8 Obstetrics and Gynecology, University Hospital Ostrava, Ostrava, Czech Republic, 9 Oncology, NN. Petrov Research Institute of Oncology Cancer Center, Saint Petersburg, Russian Federation, 10Gynaecologic Oncology, Debrecen University Clinical Center, Debrecen, Hungary, 11Oncology, Pontiﬁcal Catholic University of Rio Grande do Sul, Porto Alegre, Brazil, 12Clinical Science, Clovis Oncology, Inc., Boulder, CO, USA, 13 Biostatistics, Clovis Oncology, Inc., Boulder, CO, USA, 14Clinical Operations, Clovis Oncology, Inc., Boulder, CO, USA, 15Division of Medical Oncology and Hematology, Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada Background: Approximately 18% of patients (pts) with high-grade epithelial ovarian cancer (OC) harbour a deleterious germline BRCA1 or BRCA2 (BRCA1/2) mutation, and 7% harbour a somatic BRCA1/2 mutation (Pennington et al. Clin Cancer Res. 2014;20:764-75). The poly(ADP-ribose) polymerase (PARP) inhibitor rucaparib is approved in the United States for the treatment of pts with deleterious BRCA mutation (germline and/or somatic) associated advanced OC who have been treated with 2 chemotherapies. Data comparing PARP inhibitors to standard of care (SOC) treatment for relapsed OC are limited. Randomised studies are needed to assess the benefit-risk profile of PARP inhibitors vs SOC as treatment for BRCA1/2-mutated, relapsed, high-grade OC. Trial design: ARIEL4 (EudraCT 2016-000816-14; NCT02855944) is evaluating rucaparib vs SOC chemotherapy as treatment for pts (n345) with relapsed, high-grade OC (regardless of histology) and a deleterious germline or somatic BRCA1/2 mutation who received 2 prior chemotherapy regimens. Pts stratified by progression-free interval after their most recent platinum regimen will be randomised 2:1 to receive rucaparib (600 mg BID) (n230) or chemotherapy (n115). Pts with platinum-resistant (progressive disease [PD] 1 to < 6 mo after last platinum) or partially platinumsensitive disease (PD 6 to < 12 mo after last platinum) will receive rucaparib or weekly paclitaxel; pts with platinum-sensitive disease (PD 12 mo after last platinum) will receive rucaparib or platinum-based therapy (single-agent or doublet, per investigator discretion). Pts receiving chemotherapy have the option to cross over to rucaparib upon radiographic disease progression. The primary endpoint is investigatorassessed progression-free survival (RECIST version 1.1). Secondary endpoints include overall survival, objective response rate, RECIST/CA-125 response, duration of response, and patient-reported outcomes. Safety will be summarised descriptively using standard adverse event reporting. Clinical trial identification: EudraCT 2016-000816-14; NCT02855944 Legal entity responsible for the study: Clovis Oncology, Inc. Funding: Clovis Oncology, Inc. Disclosure: R.S. Kristeleit: Consulting or advisory role: Clovis Oncology, Roche/ Genentech, Sotio, Lytix Biopharma, Medivation Travel, Accommodations, Expenses: Clovis Oncology, Basilea Honoraria: Clovis Oncology, Roche/Genentech, AstraZeneca. D. Lorusso: Consulting or Advisory Role: AstraZeneca, Clovis Oncology, Roche, PharmaMar Travel, Accommodations, Expenses: Roche, PharmaMar. A. Oaknin: Consulting or Advisory Role: PharmaMar, Clovis Oncology, Roche, AstraZeneca. C. Tankersley, L. Maloney, S. Goble, A. Dowson, H. Giordano: Employment: Clovis

352 | Gynaecological cancers

Annals of Oncology Oncology Stock and Other Ownership Interests: Clovis Oncology. C. Unger: Employment: Clovis Oncology Stock and Other Ownership Interests: Clovis Oncology, Sillajen. A.M. Oza: Consulting or Advisory Role: Amgen, Verastem, Clovis Oncology, Immunovaccine Travel, Accommodations, Expenses: AstraZeneca Honoraria: WebRx. All other authors have declared no conflicts of interest.

989TiP

OCTOVA: A randomised phase II trial of olaparib, chemotherapy, or olaparib and cediranib in patients with BRCA-mutated platinumresistant ovarian cancer

S. Nicum1, V.Y. Strauss2, N. McGregor3, I. McNeish4, R. Roux1, M. Hall5, A. Michael6, C. Roberts2 1 Department of Oncology, Oxford University Hospitals NHS Foundation Trust, Oxford, UK, 2Centre for Statistics in Medicine, University of Oxford, Oxford, UK, 3Oncology Clinical Trials Ofﬁce, Department of Oncology, University of Oxford, Oxford, UK, 4 Institute of Cancer Sciences, University of Glasgow, Glasgow, UK, 5Mount Vernon Cancer Centre, East and North Hertfordshire NHS Trust, Middlesex, UK, 6Royal Surrey County Hospital, University of Surrey, Surrey, UK Background: Women with platinum resistant ovarian cancer (OC) have limited responses to standard therapy, and clinical trials with novel agents are therefore highly justified. Olaparib is a potent PARP inhibitor that has shown enhanced activity in women with relapsed BRCA-mutated OC in both platinum sensitive and resistant settings. Angiogenesis inhibitors, such as the oral tyrosine kinase inhibitor cediranib, are active in OC, and have shown additive effects when combined with PARP inhibitors preclinically, as hypoxia-induced downregulation of homologous recombination repair genes, BRCA1, 2 and RAD51 enhances PARP inhibitor sensitivity. Recent phase 2 trials in relapsed platinum-sensitive OC have also shown benefit from the combination of olaparib and cediranib compared to olaparib alone. The OCTOVA trial investigates the benefit of single agent olaparib compared to olaparib and cediranib or weekly paclitaxel in women with BRCA-mutated platinum-resistant OC. Trial design: Eligible patients for OCTOVA are females aged 16 years with a germline or somatic BRCA1 or 2 mutation who have progressed within 6 months of previous platinum-based therapy. Patients may have received prior PARP inhibitor and antiangiogenic therapy, with at least a 6-month interval since treatment. 132 patients will be randomised, with stratification for prior PARP inhibitor exposure and prior antiangiogenic therapy, to one of three treatment arms: paclitaxel (80mg/m2 weekly), olaparib (300mg twice daily), or cediranib (20mg once daily) and olaparib (300mg twice daily), until disease progression or unacceptable toxicity develops, and will be followed up for 18 months. Patients who progress on weekly paclitaxel will be permitted to crossover and receive single agent olaparib therapy. The primary analysis will compare the efficacy (as measured by progression-free survival) of olaparib compared to combination of olaparib and cediranib, and independently olaparib compared to paclitaxel. Secondary endpoints include safety and tolerability of the combination of olaparib and cediranib, overall survival, objective response rate, and quality of life outcomes. Clinical trial identification: OCTO_062 Legal entity responsible for the study: University of Oxford Funding: AstraZeneca (AZ) Disclosure: S. Nicum: Consultant/Advisory Boards: Roche, AstraZeneca, Abbvie, Tesaro, Clovis Speaker: AstraZeneca, Roche Clinical Trials Sponsored: AstraZeneca. All other authors have declared no conflicts of interest.

990TiP

A multicentre phase II study of AZD1775 plus chemotherapy in patients with platinum-resistant epithelial ovarian, fallopian tube, or primary peritoneal cancer

K.N. Moore1, K.A. Cadoo2, S. Chambers3, S. Ghamande4, G. Konecny5, A.M. Oza6, LM. Chen7, P.A. Konstantinopoulos8, J. Lea9, D. Spitz10, D. Uyar11, G. Mugundu12, N. Laing12, D.K. Strickland13, S. Jones14, H. Burris15, D.R. Spigel16, E. Hamilton16 1 Department of Obstetrics and Gynecology, University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA, 2Medicine, Memorial Cloan Kettering Cancer Center, New York, NY, USA, 3Cancer Center, University of Arizona, Tucson, AZ, USA, 4Georgia Cancer Center, Augusta University, Augusta, GA, USA, 5Oncology Hematology, UCLA Westwood, Los Angeles, CA, USA, 6Division of Medical Oncology and Hematology, Princess Margaret Cancer Centre, Toronto, ON, Canada, 7Gynecologic Oncology, University of California San Francisco, San Francisco, CA, USA, 8Gynecologic Oncology, Dana-Farber Cancer Institute, Boston, MA, USA, 9Obstetrics & Gynecology, University of Texas Southwestern, Dallas, TX, USA, 10Hematology & Oncology, Florida Cancer Specialists, West Palm Beach, FL, USA, 11Obstetrics & Gynecology, Medical College of Wisconsin, Milwaukee, WI, USA, 12Early Clinical Development, AstraZeneca Pharmaceuticals, Waltham, MA, USA, 13Medical and Clinical Science, Sarah Cannon Research Institute, Nashville, TN, USA, 14Strategic Development, Sarah Cannon Research Institute, Nashville, TN, USA, 15Clinical Operations, Sarah Cannon, Nashville, TN, USA, 16Medical Oncology, Sarah Cannon Research Institute/Tennessee Oncology, PLLC, Nashville, TN, USA Background: Ovarian cancers have a high rate of mutation in TP53, an alteration that produces a G1/S checkpoint deficiency and increases the level of endogenous DNA

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology

Table: 990TiP A AZD1775

175mg PO daily Days [D]1-2, 8-9, 15-16 Chemotherapy gemcitabine 1000mg/m2 IV D1, 8, 15 q28D

B

C

C2

D

225mg PO BID x 5 doses 225mg PO BID x 225mg PO BID x 5 doses D1-3, 175 or 225mg PO BID x D1-3, 8-10, 15-17 5 doses D1-3 8-10 (þ D15-17 if tolerated) 5 doses D1-3 paclitaxel 80mg/m2 carboplatin AUC carboplatin AUC 5 IV D1 q21D pegylated liposomal doxorubicin IV D1, 8, 15 q28D 5 IV D1 q21D [PLD] 40mg/m2 IV D1 q28D

Key eligibility criteria include platinum resistant OC or fallopian tube/primary peritoneal cancer (recurrence 75 years). In the DKRd arm, R 25 mg was given orally on Days 1-21. Pts received IV D as a single or split dose (Table). If C1D1 and C1D2 D infusions were not well-tolerated, C1D8 was given in 1000 mL. Pts received treatment until progression (DKd) or for 13 cycles (DKRd). To mitigate IRRs, d (20 mg) was given 3 hours before dosing on C1D1 and C1D2, and 3 hours before and the day after subsequent infusions. Paracetamol and diphenhydramine were given 3 hours before infusion. Montelukast was given prior to first D dose (optional thereafter). Results: Thirty-two pts received split-dose D. Median age (range) was 61 (34-76) years. Median number of D cycles received was 12 (1-14). Median first infusion time was 4.2

356 | Haematological malignancies

Annals of Oncology (4.0-10.3) hours. Among pts who received split-dose D, 9 (28%) pts had an IRR. Five (16%) and 4 (13%) pts had grade 1 and grade 2 IRRs, respectively. No grade 3/4 IRRs occurred. IRRs reported in 2 pts were cough, throat irritation, nausea, and headache (2 pts [6%] each). Data will be updated. Conclusions: A split first dose of D was associated with shorter infusion times and reduced incidence and lower grade of IRRs.

Table: 997PD Split dose Second dose Subsequent C1D1 and C1D2 C1D8 doses (8 mg/kg) (16 mg/kg) (16 mg/kg) Initial rate, mL/hour Rate increment increase per hour, mL/hour Maximum rate, mL/hour Total infusion, mL

50 50

50 50

100 50

200 500

200 500

200 500

Clinical trial identification: NCT01998971 Legal entity responsible for the study: Janssen Research & Development, LLC Funding: Funding provided by Janssen Research & Development Disclosure: S.Z. Usmani: Research Funding: Onyx, Janssen, Sanofi, Array Biopharma, Pharmacyclics, Takeda, Celgene, Bristo-Myers Squibb. Speakers Bureau: Celgene, Amgen, Takeda. Advisory Board: Celgene, Skyline, Onyx, Millennium, Sanofi, Janssen. A. Jakubowiak: Consultancy & Advisory Committee: Janssen. A. Chari: Consultancy & Research Funding & Advisory Committee: Amgen, Array Biopharma, Celgene, Janssen, Millenium/Takeda, Novartis. S. Lonial: Advisory Committee: Millennium, Celgene, Novartis, Bristol-Myer Squib, Onyx, Janssen. Research Funding: Janssen. M-V. Mateos: Consultancy & Honoraria: Janssen, Celgene, Takeda, Amgen. L. Benboubker: Honoraria: Takeda, Celgene, Janssen, and Amgen. K. Wu, N.Z. Khokhar: Employment: Janssen. J. Wang: Employment & Stock Ownership: Janssen. P. Doshi: Employment & Royalties & Stock Ownership: Janssen. P. Moreau: Honoraria & Consultancy: Celgen, Takeda, Janssen, Novartis, Amgen. Speakers Bureau: Janssen, Celgene.

998PD

Management of infusion-related reactions (IRRs) in patients (pts) receiving daratumumab plus standard of care for the treatment of multiple myeloma (MM) in the phase 3 studies CASTOR and POLLUX

P. Moreau1, N. Rabin2, T. Plesner3, K. Weisel4, P. Sonneveld5, M-V. Mateos6, J.M. Schecter7, H. Amin7, S. Trivedi8, M.A. Dimopoulos9 1 Hematology, University Hospital Hoˆtel-Dieu, Nantes, France, 2Department of Haematology, University College London Hospitals NHS Trust, London, UK, 3Vejle Hospital, University of Southern Denmark, Vejle, Denmark, 4Universitaetsklinikum Tuebingen der Eberhard-Karls-Universitaet, Abteilung fuer Innere Medizin II, Tuebingen, Germany, 5Department of Hematology, Erasmus MC, Rotterdam, Netherlands, 6 University of Salamanca Hospital, Salamanca, Spain, 7Janssen Research & Development, LLC, Raritan, NJ, USA, 8Janssen Research & Development, LLC, Spring House, PA, USA, 9Department of Clinical Therapeutics, National and Kapodistrian University of Athens, School of Medicine, “Alexandra” General Hospital, Athens, Greece Background: Daratumumab (D), a CD38-targeted monoclonal antibody, reduced the risk of MM progression or death by > 60% when combined with standard-of-care regimens in the phase 3 studies CASTOR (bortezomib [V] and dexamethasone [d] vs DVd; NCT02136134) and POLLUX (lenalidomide [R] and d vs DRd; NCT02076009). This analysis evaluated the management of D-related IRRs in the DVd and DRd arms of CASTOR and POLLUX. Methods: Pts had MM and had received 1 line of therapy. In CASTOR, pts were given 8 21-day cycles of Vd (V 1.3 mg/m2 subcutaneously on Days 1, 4, 8, and 11; d 20 mg per os [PO]/intravenously [IV] on Days 1-2, 4-5, 8-9, and 11-12) 6 D (16 mg/kg IV, weekly [QW] for Cycles 1-3, every 3 weeks [Q3W] for Cycles 4-8, then every 4 weeks [Q4W] thereafter). In POLLUX, pts were given 28-day cycles of Rd (R 25 mg PO on Days 1-21; d 40 mg QW) 6 D (16 mg/kg IV QW for Cycles 1-2, every 2 weeks for Cycles 3-6, then Q4W thereafter). In addition, pre-infusion medication consisted of 20 mg d (or equivalent) IV/PO, 650-1000 mg paracetamol, and 25-50 mg diphenhydramine (or equivalent). Pts with high-risk respiratory complications received diphenhydramine on Days 1 and 2, a short-acting b2 adrenergic receptor agonist and control medications for lung disease after D infusion. Results: All pts receiving D were given pre-infusion medication. In CASTOR and POLLUX, 31 (13%) and 21 pts (7%) received post-infusion medications, respectively. In both trials, the median duration of D infusion was 7.0, 4.3, and 3.4 hours for the first, second, and subsequent infusions, respectively. IRRs occurred in 45% and 48% of pts and 98% and 96% of IRRs occurred during the first infusion in CASTOR and POLLUX, respectively. Most IRRs were grade 1/2 and no grade 4 IRRs were reported.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology The median time to onset of IRRs after starting the first D infusion was 84 minutes in CASTOR and 90 minutes in POLLUX. In CASTOR, 2 pts discontinued treatment due to IRRs; in POLLUX, 1 pt discontinued D due to a grade 3 IRR, but continued Rd. Conclusions: Most D-related IRRs occurred during the first infusion and were grade 1/2. D-related IRRs were easily managed with pre- and post-infusion medications. Clinical trial identification: NCT02136134 and NCT02076009 Legal entity responsible for the study: Janssen Research & Development, LLC Funding: Funding provided by Janssen Research & Development Disclosure: P. Moreau: Honoraria & Consultancy: Celgen, Takeda, Janssen, Novartis, Amgen. Speakers Bureau: Janssen, Celgene. N. Rabin: Consultancy: Janssen, Amgen, Novartis, Takeda, Bristol-Myers Squibb. Honoraria: Janssen, Takeda. Speakers Bureau: Janssen, Celgene. T. Plesner: Consultancy: Janssen, Takeda. Research Funding: Janssen. Advisory Committee: Janssen, Genmab. K. Weisel: Consultancy & Honoraria: Amgen, Bristol-Myers Squib, Celgene, Janssen, Novartis, Takeda, Onyx. Research Funding: Janssen, Celgene, Amgen, Sanofi. P. Sonneveld: Consultancy & Research Funding & Honoraria: Amgen, Celgene, Janssen, Karyopharm, Takeda. M-V. Mateos: Consultancy & Honoraria: Janssen, Celgene, Takeda, Amgen. J.M. Schecter: Employment & Equity: Janssen. H. Amin, S. Trivedi: Employment: Janssen. M.A. Dimopoulos: Consultancy & Honoraria: Celgene, Janssen, Takeda, Amgen.

999PD

Comparison of efficacy of new therapies between younger and older patients with relapsed and refractory multiple myeloma: A metaanalysis

T. Landre1, C. Al-Nawakil2, F. Karaoud3, C. Taleb3 UCOG, Hopital Rene´ Muret APHP, Sevran, France, 2He´matologie Centre Recherche Clinique, AP-HP Hoˆpital Avicenne, Bobigny, France, 3Geriatric Oncology, AP-HP Hoˆpital Rene´ Muret, Sevran, France 1

Background: Multiple myeloma is a disease of age. With all of the new myeloma drugs being developed, there are number of treatment options for relapsed and refractory multiple myeloma (RRMM). However, in our knowledge, few data are available in patients older than 65 or 75 years. We performed a meta-analysis to compare the efficacy of new drugs to treat RRMM between younger and older patients. Methods: PubMed and the Cochrane databases were searched up to April 2016. We included phases III randomized controlled trials (RCTs) of monoclonal antibodies (mAbs) targeting CD38 or SLAMF7 (daratumumab, elotuzumab), second generation proteasome inhibitors (carfilzomib, ixazomib) and histone deacetylase (HDAC) inhibitors (vorinostat, panobinostat) reporting subgroups comparison of progressionfree survival (PFS) based of aged cut-offs. The summary hazard ratio (HR) and 95% confidential interval (CI) were calculated. Results: A total of 5241 patients from eight RCTs of RRMM new therapies were included (CASTOR, POLLUX, ELOQUENT-2, ASPIRE, ENDEAVOR, TOURMALINE-MM1, PANORAMA-1 and VANTAGE-088). When patients are dichotomized into younger and older groups with an age cut-off of 65-75 years, RRMM new therapies improved PFS in both younger (HR, 0.62; 95% CI, 0.56–0.70) and older (HR, 0.67; 95% CI, 0.60–0.74) groups. An improvement in PFS with mAbs was observed in younger (HR, 0.57; 95% CI, 0.46–0.72) and older (HR, 0.52; 95% CI, 0.42–0.64) patients. An improvement in PFS with HDAC inhibitors was also observed in both younger (HR, 0.67; 95% CI, 0.56–0.80) and older (HR, 0.78; 95% CI, 0.63– 0.97) as well as with second generation proteasome inhibitors (HR, 0.61; 95% CI, 0.52– 0.73 and HR, 0.70; 95% CI, 0.60–0.81 respectively). Conclusions: A benefit in PFS with new therapies was significant in both younger and older patients with relapsed and refractory multiple myeloma with a cut-off age of 65– 75 years. Legal entity responsible for the study: APHP Funding: None Disclosure: All authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

1000PD

Copanlisib treatment in patients with relapsed or refractory indolent B-cell lymphoma: Subgroup analyses from the CHRONOS-1 study

M. Provencio Pulla1, A. Santoro2, L. Mollica3, S. Lepp€a4, G. Follows5, G. Lenz6, W.S. Kim7, 11 € , M. Kosinova12, K. Bouabdallah13, A. Nagler8, P. Panayiotidis9, J. Demeter10, M. Ozcan F. Morschhauser14, T. Ishida15, L. Huang16, J. Garcia-Vargas17, B.H. Childs17, P.L. Zinzani18, M. Dreyling19 1 Servicio de Oncologıa Me´dica, Hospital Universitario Puerta de Hierro, Madrid, Spain, 2 Department of Oncology and Hematology, Istituto Clinico Humanitas, Rozzano, Italy, 3 Department of Hematology, Hoˆpital Maisonneuve-Rosemont-Montreal, Montreal, Canada, 4Department of Oncology, Helsinki University Central Hospital, Helsinki, Finland, 5Department of Haematology, Cambridge University Hospitals NHS Foundation Trust Addenbrooke’s Hospital j, Cambridge, UK, 6Translational Oncology, University Hospital Mu¨nster, Mu¨nster, Germany, 7Division of Hematology and Oncology, Sungkyunkwan University School of Medicine, Samsung Medical Center, Seoul, Republic of Korea, 8Hematology Division, Chaim Sheba Medical Center, Tel Aviv University, Tel-Hashomer, Israel, 9Molecular Hematology Laboratory, 1st Department of Propaedeutic Medicine, National and Kapodistrian University of Athens, School of Medicine, Laikon General Hospital, Athens, Greece, 10First Department of Internal Medicine, Division of Haematology, Semmelweis University, Budapest, Hungary, 11 Department of Hematology, Ankara University School of Medicine, Ankara, Turkey, 12 Department of Hematology, GAUZKO “Kemerovo Regional Clinical Hospital”, Kemerovo, Russian Federation, 13Service d’He´matologie et de The´rapie Cellulaire, University Hospital of Bordeaux, Pessac, France, 14Department of Hematology, CHRU ao Paulo, Hoˆpital Claude Huriez, Lille, France, 15Development Operations, Bayer SA, S~ Brazil, 16Clinical Statistics, Bayer HealthCare Pharmaceuticals USA, Whippany, NJ, USA, 17 Clinical Development, Bayer HealthCare Pharmaceuticals Inc., Whippany, NJ, USA, 18 Institute of Hematology “L. e A. Ser agnoli”, University of Bologna, Bologna, Italy, 19 Medizinische Klinik und Poliklinik III, Klinikum der Universit€ at Mu¨nchen-Grosshadern, Munich, Germany Background: Copanlisib, a pan-class I phosphatidylinositol 3-kinase (PI3K) inhibitor with predominant activity against PI3K-a and PI3K-d isoforms, has recently been shown to achieve a 59% objective tumor response rate (ORR) in a phase II study in patients with relapsed or refractory (r/r) indolent B-cell lymphoma. We report here the results of subgroup analyses conducted based on demographic and baseline disease characteristics. Methods: Patients with indolent B-cell non-Hodgkin lymphoma (4 subtypes: follicular [FL], marginal zone [MZL], small lymphocytic [SLL] and lymphoplasmacytoid/ Waldenstro¨m macroglobulinemia [LPL-WM]) and r/r to 2 prior lines of treatment were eligible. Previous treatment had to include rituximab and an alkylating agent. Copanlisib (60 mg, I.V.) was administered intermittently on days 1, 8 and 15 of a 28day cycle. The primary efficacy endpoint was ORR as assessed per independent radiologic review (Cheson et al. 2007). Results: The full analysis set comprised 142 patients, of which 141 patients had indolent lymphoma (FL/MZL/SLL/LPL-WM: 104/23/8/6). ORR per histological subgroup was 58.7%/69.6%/75.0%/16.7%, respectively. ORR based on demographics were generally consistent across categories. Likewise, there were no major differences in ORR between any of the baseline disease characteristics and prior therapy subgroups with regards to ECOG PS (0 [58.8%] vs. 1 [59.7%]), longest diameter of baseline lesion (< 7cm [59.8%] vs. 7cm [59.1%]), received prior bendamustine (yes [62.7%] vs. no [56.6%]), number of prior therapies (< 4 [59.8%] vs. 4 [57.5%]), or refractoriness to last regimen (yes [60.5%] vs. no [57.1%]). Median duration of response (DOR) by tumor histology for the subgroups with 10 responders was 370 days (range 33-687) for FL patients and had not yet been reached for MZL patients (2 of 16 responders having progressed). Conclusions: Objective response rates were consistently high in patients with r/r indolent B-cell lymphoma treated with copanlisib with the exception of LPL-WM patients. There were no major differences in the ORR between any of the baseline disease characteristics and prior therapy subgroups, confirming the robustness of the primary efficacy endpoint. Clinical trial identification: NCT01660451 Legal entity responsible for the study: Bayer AG Funding: Bayer AG Disclosure: T. Ishida: Employment: Bayer SA. L. Huang, J. Garcia-Vargas, B.H. Childs: Employment: Bayer HealthCare Pharmaceuticals Inc. M. Dreyling: Advisory boards: Bayer, Gilead Honoraria: Bayer, Gilead. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx373 | 357

abstracts 1001PD

Tumor gene expression signatures of BCR/PI3K dependence in association with copanlisib monotherapy activity in heavily pretreated patients with indolent NHL and follicular lymphoma

L. Liu1, K. Ko¨chert2, H. Seidel3, J. Garcia-Vargas4, B.H. Childs5, G. Follows6, na1 K. Bouabdallah7, M. Dreyling8, C. Pe~ 1 Biomarker Strategy, Bayer HealthCare Pharmaceuticals, Whippany, NJ, USA, 2Clinical Statistics, Bayer AG, Berlin, Germany, 3Bioinformatics, Bayer AG, Berlin, Germany, 4 Clinical Development, Bayer HealthCare Pharmaceuticals USA, Whippany, NJ, USA, 5 Clinical Development, Bayer HealthCare Pharmaceuticals, Whippany, NJ, USA, 6 Department of Haematology, Cambridge University Hospitals NHS Foundation Trust Addenbrooke’s Hospital, Cambridge, UK, 7Service d’He´matologie et de The´rapie Cellulaire, University Hospital of Bordeaux, Pessac, France, 8Medizinische Klinik und Poliklinik III, Klinikum der Universit€ at Mu¨nchen-Grosshadern, Munich, Germany Background: Copanlisib, a novel class I PI3K inhibitor with predominant activity against a and d isoforms, has shown robust single agent anti-tumor activity in a phase 2 study in heavily pretreated patients with indolent NHL (iNHL) and follicular lymphoma (FL) (NCT01660451; Part B), with response rates of 59.6% and 58.7%, respectively. Baseline tumor gene expression profiling (GEP) was performed to confirm if gene signatures identified in patients with indolent or aggressive NHL (NCT01660451; Part A) are molecular determinants for copanlisib antitumor activity in Part B. Methods: Signaling pathway gene sets (n ¼ 33) were ranked by enrichment analysis (GSEA) for association with objective response based on normalized enrichment score (NES) and false discovery rate (FDR) q values. The association of weighted geneexpression score (WGS, reflecting the overall expression level for each gene set) with response was analyzed by logistic regression. Results: Seventy-one patients with iNHL, including 54 FL, had both response data and evaluable gene expression data. All 5 gene sets reflecting upregulated PI3K/BCR signaling were top-ranked for association with higher response rates in iNHL (GSEA NES 1.93, FDR q < 0.01; WGS AUC 0.65, nominal p 0.04) and FL (GSEA NES 1.50, FDR q 0.01; WGS AUC 0.60, nominal p 0.23). Among patients with objective responses, 66% (33/47) of iNHLs and 71% (24/36) of FLs had high PI3K/BCR gene signature expression levels; for patients with CR, 86% (6/7) iNHL and 83% (5/6) FL had high levels. Further, 4 gene sets enriched with T-cell signatures were associated positively with copanlisib response (NES 1.48, FDR q 0.08). In contrast, up-regulation of macrophage/stromal gene sets was potentially associated with a lower likelihood of response to copanlisib treatment in FL (NES -1.21, q 0.21). Conclusions: Tumor gene expression profiling demonstrates that up-regulation of the BCR/PI3K pathway is frequent and dominant in iNHL and FL, and is associated with the high and durable copanlisib responses. These findings are consistent with copanlisib’s mode of action and strongly support the rationale for treatment of iNHL and FL patients with copanlisib. Clinical trial identification: NCT01660451; Part B Legal entity responsible for the study: Bayer AG Funding: Bayer AG Disclosure: L. Liu: Employee: Bayer HealthCare Pharmaceuticals. K. Ko¨chert, H. Seidel: Employee: Bayer AG. J. Garcia-Vargas, B.H. Childs, C. Pe~ na: Employment: Bayer HealthCare Pharmaceuticals. All other authors have declared no conflicts of interest. 1002PD

Standardized mortality ratios and event-free survival as new prediction tools of early increase in mortality in follicular lymphoma

F. Franco1, A. Royuela2, M. Torrente1, J. G omez-Codina3, M. Llanos Rodrıguez4, J. Guma Padr o5, C. Quero Blanco6, D. Aguiar7, A. Rueda Domınguez8, M. Provencio Pulla1 1 Medical Oncology, Hospital Universitario Puerta de Hierro Majadahonda, Madrid, Spain, 2Biostatistics, Biomedical Sciences Research Institute Puerta de HierroMajadahonda University Hospital, Madrid, Spain, 3Medical Oncology, La Fe University Hospital, Valencia, Spain, 4Medical Oncology, Hospital University of Canarias, Santa Cruz De Tenerife, Spain, 5Medical Oncology, Hospital Universitari Sant Joan de Reus, alaga, Reus, Spain, 6Medical Oncology, Hospital Universitary Virgen de la Victoria M Malaga, Spain, 7Medical Oncology, Gran Canaria Doctor Negrın University Hospital, 8 Las Palmas De Gran Canaria, Spain, Oncology, Costa del Sol Hospital, Marbella, Spain Background: There are few studies that analyze follicular lymphoma (FL) mortality compared to the general population of the same -sex and age group. Given the recent clinical relevance of the predictive event-free survival (EFS) indexes EFS12 and EFS24, we obtained them in our study cohort in order to estimate their association with overall survival (OS). Methods: Patients diagnosed with FL were prospectively enrolled from 1980 to 2013. Standardized mortality ratios (SMR) were obtained using yearly sex and age specific mortality rates in Spain, and OS was compared with age- and sex-matched general population data. EFS were defined as the time from diagnosis until relapse or progression, unplanned retreatment of lymphoma after initial management, or death due to any cause. EFS12 and 24 were defined as EFS status at 12 or 24 months from diagnosis, respectively. The crude probability of death was estimated by using the Kaplan–Meier method, and differences between patient groups were assessed by the log-rank test. In order to investigate the specific contribution of age, sex, period of diagnosis, treatment and FLIPI score, a multivariable Cox proportional hazards model was adjusted, all statistical tests were two-sided, and a p-value 60y 60y 2 0 or 1 Elevated Normal Stage IV Stage III Stage II Stage I 90% 5.53 x 109/L). The 5-year OS for LR, IR and HR pts was 86%, 65.7% and 22.5% (P < 0.001), respectively, and our PI was independent of the NCCN-IPI. Conclusions: Our data showed that SA/ANC PI could predict OS in DLBCL pts and may present a reliable, convenient and sensitive predictor to identify pts with poor prognosis in addition to NCCN-IPI. Legal entity responsible for the study: Specialized Hospital for Active Therapy of Hematological Diseases Funding: None Disclosure: All authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 1016P

Autologous stem cell transplantation (ASCT) is safe and effective for the treatment of non-hodgkin lymphoma (NHL) in an elderly population of patients over 65 years old: A single center experience

M.P. Tardy1, J. Viotti2, A. Boscagli Melaine1, A. Thyss3, F. Peyrade4, L. Gastaud1 Medical Oncology, Centre Anticancer Antoine Lacassagne, Nice, France, 2Biostatistics, Centre Anticancer Antoine Lacassagne, Nice, France, 3Medical Oncology, Centre Antoine Lacassagne, Nice, France, 4Onco-hematology, Centre Antoine Lacassagne, Nice, France

1

Background: High-dose chemotherapy followed by ASCT is a widely used treatment for agressive or recurrent NHL for young patients (pts). Limited data are available on the feasibility and the results of this strategy for older pts, who are often excluded from agressive treatment. This study aimed at comparing pts 65 years old(yo) to a younger pts’ population in term of tolerance, safety and results of ASCT. Methods: We did a retrospective study in one center in France. We included every consecutive pts treated by ASCT for NHL from May 2007 to January 2016. We collected data on the characteristics of the pts and their disease, previous treatments, and tolerance and outcome after ASCT. Results: 48 pts 65 yo (mean: 49.5) and 129 < 65 yo (mean: 68.7) at the time of the transplant were included. The most common histology was diffuse large B cell lymphoma (p ¼ 0.205). There were only 2 differences between the 2 groups. First, the number of pts with comorbidities was higher in the elderly population (p ¼ 0.016), especially cardio-vascular (p < 0.001). Secondly, the moment of the ASCT was mainly on first line for young pts vs at the time of relapse for older pts (p < 0.001). At the time of ASCT, in both groups (p ¼ 0,306), a majority of patients were in complete response. No differences were found between the 2 groups for the conditionning regimen, number of CD34þ cells reinjected, number of transfusion, weight loss, length of the hospital stay and duration of aplasia. There were no differences of grade 3 toxicities (hematologic, infections, digestive, renal, cardiac, mucositis.) for both groups (p ¼ 0,116). Treatment Related Mortality (death within 30 days following ASCT) was: 2% for pts 65yo vs 3.9% (p ¼ 1). The mean follow-up time was 56 months for young pts vs 77 months for older (p < 0.001). The specific survival was similar between the 2 groups, 65% for the young pts group vs 72% for the older (p ¼ 0.63) 3 years after ASCT. Conclusions: High-dose chemotherapy followed by ASCT is as safe and effective in a population of pts 65 yo compared to a younger population. Agressive treatment could be considered earlier in the management of elderly pts and should not be excluded only depending on the age of the patient. Legal entity responsible for the study: Centre antoine Lacassagne Funding: None Disclosure: All authors have declared no conflicts of interest.

1017P

Phenotypic and functional characterization of tumor infiltrating lymphocytes (TIL) grown from non-hodgkin lymphoma tumors: Implications for the development of novel therapies for lymphoma

L. Karyampudi1, A. Gokuldass2, M. Blaskovich2, M. Fardis3, M.T. Lotze2 Translational Research, Lion Biotechnologies, Tampa, FL, USA, 2Translational Research, Lion Biotechnologies, Tampa, FL, USA, 3Translational Research, Lion Biotechnologies, San Carlos, CA, USA

Disclosure: L. Karyampudi, A. Gokuldass, M. Blaskovich, M.T. Lotze: Employee of Lion Biotechnologies, Inc, receive compensation from Lion. Hold stock and/or stock options in Lion Biotechnologies, Inc. M. Fardis: Employee and director of Lion Biotechnologies, Inc, receive compensation from Lion. Hold stock and/or stock options in Lion Biotechnologies, Inc. Member of the Board of Directors of Lion Biotechnologies, Inc.

1018P

International prognostic scores (IPS-7, IPS-3 and IPS-3 new) for prediction of FFS and OS in cases with Hodgkin Lymphoma. Which is more practical and accurate?

S. Paydas1, M. Dogan2, V. Ercolak3, G. Seydaoglu4 Medical Oncology, Cukurova University, Adana, Turkey, 2Medical Oncology, Ankara Numune Hospital, Ankara, Turkey, 3Medical Oncology, Mersin Hospital, Mersin, Turkey, 4 Biostatistics, Cukurova University, Adana, Turkey

1

Background: Hodgkin Lymphoma (HL) is one of the curable malignant diseases. International prognostic score7 (IPS-7) was a valuable scoring system predicting FFP and OS in cases with HL. A simpler prognostic score: IPS-3 has been proposed 2 years ago and in this scoring system age and stage were found to be significant for FFP and age, stage and hemoglobin were found to be significant for OS. Here we evaluated IPS-3 new system in cases with HL. Methods: 364 patients with HL treated by ABVD have been included in this study. Two thirds of the patients had nodular sclerosing type HL, 76 had mixed cellularity type. Median follow up was 71 months. Results: Seven clinical parameters on the basis of the IPS-7 determined to be associated with adverse clinical outcome were evaluated. The prognostic ability of seven IPS factors was evaluated for both FFP and OS. A new 3-factor prognostic score (IPS-3 new) was constructed utilizing factors that were significant in multivariate Cox models: age>45 years, stage IV disease and lymphocytopenia were found to be independent factors. Thus IPS-3-new was constructed utilizing these 3 factors that were significant in multivariate Cox models. Lymphocytopenia was used instead of hemoglobin 60 years) with newly diagnosed AML. The randomized part of the study (cohort-1) was terminated due to excessive VPA-induced toxicity (Tassara et al, Blood 2014;123:4027-36.). The study was amended thereafter (cohort-2) to evaluate a cytarabine dose-intensifcation in first consolidation therapy. Here we report on the comparison of the two cohorts. Methods: Between 2004 and 2008, patients were treated in cohort-1 (n ¼ 186) and cohort-2 (n ¼ 376). 2 cycles of induction therapy (ATRA, idarubicin, cytarabine, n ¼ 93 with VPA) were followed by consolidation-1 (mitoxantrone, ATRA, cytarabine [cohort-1, 0.5g/m2; cohort-2, 1g/m2] bid, days 1-3) and consoldation-2 (idarubicin, etoposide, ATRA). Results: Median age was 68 (range, 60-84) years without difference between the cohorts (p ¼ 0.49). Complete remission (CR) rates after induction therapy were 45% and 48% (p ¼ 0.59) in cohort-1 and -2, respectively. There were no significant differences in the cumulative incidences of relapse (CIR, p ¼ 0.26) and death (p ¼ 0.51) between cohort1 and -2 with CIR of 63% (SE, 4.8%) in cohort-1 compared to 51% (SE, 6.3%) in cohort-2. A Cox regression model on overall survival revealed older age (hazard ratio (HR) for a 10 years difference, 1.97, p < 0.0001), 2010 European LeukemiaNet (ELN) unfavorable risk (HR, 1.57, p ¼ 0.0003) as well as cohort-1 (HR, 1.31, p ¼ 0.02) as unfavorable parameters and ELN favorable risk (HR 0.55, p < 0.0001) as favorable prognostic parameter. Survival was inferior (p ¼ 0.03) in cohort-1 with 21% (95%-CI, 1628%) compared to cohort-2 with 28% (95%-CI, 23-33%) at 2 years. In an age-adjusted analysis the molecular marker FLT3-ITD was associated with an unfavorable prognosis. Conclusions: Although evaluated in a cohort- rather than a randomized study, intensification of cytarabine dosage in consolidation therapy seems to improve survival. Clinical trial identification: NCT00151255 Legal entity responsible for the study: University Hospital Ulm Funding: University Ulm, Pfizer Disclosure: R. Schlenk: Research funding: Novartis, Pfizer, Amgen, AstraZeneca, PharmaMar; Speakers bureau: Novartis, Pfizer; Advisory board: Daiichi Sankyo, Novartis, Pfizer. All other authors have declared no conflicts of interest.

1031P

The EUROSKI biomarker study: Analyzing the mechanisms of treatment-free remission in chronic myeloid leukemia

R. Se´bastien1, C. Sticht2, M. Pfirrmann3, D. Nowak1, A. Fabarius1, W. Seifarth1, B. Spiess1, P. Panayiotidis4, M. Pagoni5, M. Dimou5, J. Dengler6, C.F. Waller7, T.H. Bru¨mmendorf8, A. Burchert9, G. Freunek10, W-K. Hofmann1, F-X. Mahon11, S. Saussele1 1 III. Medizinische Klinik, Medizinische Fakult€ at Mannheim, Universit€ at Heidelberg, at Mannheim, Germany, 2Zentrum fu¨r Medizinische Forschung, Medizinische Fakult€ 3 Mannheim, Universit€ at Heidelberg, Mannheim, Germany, Institut fu¨r Medizinische Informationsverarbeitung, Biometrie und Epidemiologie (IBE), Ludwig-MaximiliansUniversit€at Mu¨nchen, Munich, Germany, 4Molecular Hematology Laboratory, 1st Department of Propaedeutic Medicine, National and Kapodistrian University of Athens, School of Medicine, Laikon General Hospital, Athens, Greece, 5Of Hematology, Hellenic amatologie und Onkologie, Onkologische Society, Athens, Greece, 6Internistische H€ atsklinikum Praxis Heilbronn, Heilbronn, Germany, 7Klinik fu¨r Innere Medizin I, Universit€ 8 Freiburg, Freiburg, Germany, Med. Klinik IV, Uniklinik RWTH Aachen, Aachen, Germany, 9 Klinik fu¨r H€amatologie, Onkologie und Immunologie, Universit€ atsklinikum Gießen und Marburg, Marburg, Germany, 10Medizinisches Versorgungszentrum, Klinikum Straubing, 11 Straubing, Germany, Institut Bergonie´, University Bordeaux, Bordeaux, France Background: A substantial proportion of chronic myeloid leukemia (CML) patients in deep molecular response (DMR) reach treatment-free remission (TFR) after tyrosine kinase inhibitors (TKI) cessation. The aim of this study is to identify a gene signature predictive for TFR using whole transcriptome expression analyses. Methods: RNA from peripheral blood (PB) leukocytes of 60 CML patients who stopped TKI therapy within the EUROSKI study and 10 healthy controls were isolated. CML patients were divided into two groups of whom n ¼ 30 patients had ongoing TFR, while 30 patients encountered molecular recurrence. RNA was isolated at the last day of TKI intake. In order to investigate differentially expressed genes, whole transcriptome arrays (Clariom D, Affymetrix) were analyzed. Candidate biomarkers were tested in multivariate analyses and gene set enrichment analyses (GSEA). Results: CML patients in DMR compared to healthy controls showed 16000 differentially expressed genes (p < 0.05). The natural killer cell marker CD69 showed overexpression with highest fold change (> 8-fold) for CML patients versus healthy controls. Significant enrichment of NFjB mediated TNFa and TGFb pathways (FDR30% reduction in 3 pts. Of these 15 pts, 8 were HPVþ, 4 had PD-L1 1%, and 5 had >20% increase in target lesion at first progression. Frequencies of grade 34 treatment-related adverse events were similar in the TBP and NTBP groups. At D1 and D43, the percentage of CD8þ T cells in peripheral blood for TBP pts (n ¼ 5) was similar to that for responders (n ¼ 15), which was significantly higher vs NTBP pts (n ¼ 9). At D1, the percentage of PD-1þ CD8þ effector T cells was significantly lower in responders and TBP pts vs NTBP pts. At D1, TBP pts, similar to responders, had a significantly lower PD-1þ Treg percentage vs NTBP pts. Conclusions: Nivo treatment beyond progression in some pts with R/M SCCHN was tolerable and associated with tumor size reductions. Certain immune cell profiles in the TBP group appear similar to those of responders. Clinical trial identification: NCT02105636 Legal entity responsible for the study: Bristol-Myers Squibb Funding: Bristol-Myers Squibb Disclosure: R. Haddad: Grants to my institution from Bristol-Myers Squibb, Merck, Celgene, Pfizer, and AstraZeneca; personal fees from Bristol-Myers Squibb, Merck, AstraZeneca, Pfizer, Celgene, and Eisai. G. Blumenschein Jr: Research grants from: Merck, AstraZeneca, Celgene, AbbVie, Benentech, Xcovery, Novartis, Bayer, BristolMyers Squibb, GlaxoSmithKline; Consulting fees from: Bristol-Myers Squibb, Bayer, Celgene, Clovis, AbbVie, Ariad, AstraZeneca, Merck. J. Fayette: Personal fees from Bristol-Myers Squibb and AstraZeneca. J. Guigay: Research grants to my institution

abstracts

Annals of Oncology from: Bristol-Myers Squibb, Boehringer Ingelheim, GlaxoSmithKline, and Merck Serono. A.D. Colevas: Research grants outside of the submitted work from: BristolMyers Squibb, IRX Therapeutics, Threshold Pharmaceuticals, AstraZeneca, and Innate Pharma. L. Licitra: Consulting/grants to institution/travel fees: Merck-Serono; Consulting/grants to institution: Eisai, MSD, Merck-Serono, Boehringer Ingelheim, Novartis, AstraZeneca, Roche; Consulting/travel fees: Bayer, Debiopharm, Sobi; Consulting: Bristol-Myers Squib. S. Kasper: Personal fees for advisory board participation from Bristol-Myers Squibb. E.E. Vokes: Consulting/advisory role with: AbbVie, Amgen, AstraZeneca, Bristol-Myers Squibb, Boehringer-Ingelheim, Celgene, Eli Lilly, Genentech, Leidos, Merck, Regeneron, Serono, Takeda, VentiRx. N.F. Saba: Advisory board participation/personal fees: Bristol-Myers Squib, Lilly, Merck; Member of DSMC/personal fees: Pfizer. M. Tahara: Ad board/Research funding: Ono Pharmaceuticals, MSD, AstraZeneca, Pfizer; Ad board/Lecture honorarium: BristolMyers Squib, Bayer; Lecture honorarium/Research funding: Eisai; Lecture honorarium: Otsuka; Research funding: Boehringer-Ingelheim, Novartis, NanoCarrier. M. Monga, M. Lynch: Bristol-Myers Squibb employee. L. Li: Bristol-Myers Squibb employee. J.W. Shaw: Bristol-Myers Squibb employee and shareholder. M.L. Gillison: Honoraria, Consulting, Travel, Accommodations, and Expenses: Celgene, Lilly, Amgen, GlaxoSmithKline, Bristol-Myers Squibb, AstraZeneca, Merck; Research funding to my institution: Bristol-Myers Squibb, Merck, AstraZeneca, Kyowa. K.J. Harrington: Honoraria: AstraZeneca, Bristol-Myers Squib, Merck, MSD, Pfizer; Advisory Board membership: AstraZeneca, Bristol-Myers Squib, Merck, MSD, Pfizer; Research Grants: MSD. R.L. Ferris: Research funding: AstraZeneca/MedImmune, Bristol-Myers Squibb, VentiRx; Clinical Trial: AstraZeneca/MedImmune, Bristol-Myers Squibb, Merck; Advisory Board: Amgen, AstraZeneca/MedImmune, Bristol-Myers Squibb, EMD Serono, Lilly, Marck, Pfizer. All other authors have declared no conflicts of interest.

1044O

Long-Term Safety and Clinical Outcomes of Atezolizumab in Head and Neck Cancer: Phase Ia Trial Results

R. Bahleda1, F.S. Braiteh2, A.S. Balmanoukian3, I. Bra~ na4, F.S. Hodi5, L. Garbo6, B. Liu7, L. Molinero8, C. O’Hear9, X. Shen9, A.D. Colevas10 1 Early Drug Development Department, Gustave Roussy, Villejuif, France, 2Medical Oncoloy, Comprehensive Cancer Centers of Nevada, Las Vegas, NV, USA, 3Research, The Angeles Clinic and Research Institute, Los Angeles, CA, USA, 4Medical Oncology, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 5Medical Oncology, DanaFarber Cancer Institute, Boston, MA, USA, 6Hematology, New York Oncology, Albany, NY, USA, 7 Biostatistics, Genentech, Inc., South San Francisco, CA, USA, 8Oncology Biomarker Development, Genentech, Inc., South San Francisco, CA, USA, 9Product Development Oncology, Genentech, Inc., South San Francisco, CA, USA, 10Medicine, Stanford University, Stanford, CA, USA Background: Checkpoint inhibitors have shown efficacy in pts with recurrent and/or metastatic head and neck cancer (HNC). Atezolizumab (atezo; antiPD-L1) inhibits binding of PD-L1 to PD-1 and B7.1, thereby restoring tumor-specific T-cell immunity. Here, we examine single-agent safety and clinical activity of atezo in pts with advanced HNC. Methods: Pts with HNC received atezo IV q3w (15 or 20 mg/kg or 1200 mg) in a Ph Ia study (NCT01375842). Pts were initially enrolled non-selectively (n ¼ 10); once PD-L1 was identified as a potential biomarker, pts were selected by PD-L1 status (> 5% PD-L1 expression on IC [IC2/3]; centrally evaluated by VENTANA SP142 IHC assay, n ¼ 22). HPV status was assessed by PCR. Treatment was originally for 16 cycles or up to 1 y; pts were subsequently treated until loss of clinical benefit. Primary endpoint was safety; tumor responses were evaluated by RECIST v1.1. Results: As of 31 Dec 2016, with follow-up of 14 mo, 32 pts were safety and efficacy evaluable. 84% were male; 66% had ECOG PS of 1. Median age was 62 y (range, 32-78 y), pts were heavily pre-treated (53% had 2 prior lines of therapy) and 66% were current/previous tobacco users. Most common primary tumor sites were oropharynx (50%), oral cavity (22%) and nasopharynx (19%). Median treatment duration was 3.4 mo; 21/32 pts (66%) had a treatment-related AE. 3 pts (9%) had Gr 3 treatmentrelated AEs (tumor lysis syndrome, hyponatremia, pruritus, colitis). 1 pt (3%) had Gr 4 treatment-related cardiac tamponade. No Gr 5 treatment-related AEs were seen. In all pts (IC0/1, n ¼ 7; IC2/3, n ¼ 25), confirmed ORR was 22%, mPFS was 2.6 mo (range, 0.5-48.4 mo) and mOS was 6.0 mo (range, 0.5-51.6þ mo). Clinical activity by PD-L1 subgroups is shown in the Table. Preliminary analyses indicate that there was no association between HPV status and clinical outcomes. Conclusions: In advanced HNC, atezo was well tolerated. Encouraging response and long-term survival were seen independently of PD-L1 IHC or HPV status and warrant further investigation.

Volume 28 | Supplement 5 | September 2017

Table: 1044O Clinical Activity per RECIST v1.1 in PD-L1 Subgroups

ORR, n (%) CR PR DCR, n (%) mDOR, mo (range)b mPFS, mo (range) mOS, mo (range) 1-year OS rate 2-year OS rate 3-year OS rate

IC0/1a (n ¼ 7)

IC2/3a (n ¼ 25)

1 (14%) 0 1 (14%) 3 (43%) 7.4 5.7 (0.5-13.5) 9.0 (0.5-26.5) 43% 29% NE

6 (24%) 0 6 (24%) 7 (28%) 26.2 (2.8-45.8) 2.6 (0.5-48.4) 5.6 (1.1-51.6þ) 34% 18% 18%

þ

indicates a censored value. Data by PD-L1 expression on TC will be presented. b n ¼ 1 for IC0/1, no estimate for mDOR; n ¼ 6 for IC2/3. IC0 ¼ PD-L1 expression on < 1%; IC1 ¼ 1% to < 5%; IC2 ¼ 5% to < 10%; IC3 ¼ 10%. PD-L1 subgroups do not reﬂect the natural prevalence as not enrolled as an all-comer cohort. DCR, disease control rate: % of pts with CR, PR and SD 24 wk; IC, tumor-inﬁltrating immune cells; mDOR, median duration of response; TC, tumor cells. a

Clinical trial identification: NCT01375842 Legal entity responsible for the study: F. Hoffmann-La Roche Ltd. Funding: F. Hoffmann-La Roche Ltd. Disclosure: F.S. Braiteh: Speaking and consulting fees received from Genentech. A.S. Balmanoukian: Speaker’s bureau for Merck, Bristol-Myers Squib, Genentech, and AstraZeneca. F.S. Hodi: Consultant Merck, Bristol-Myers Squibb, Genentech, EMD Serono, Amgen; Research support to institution from Bristol-Myers Squibb. B. Liu, L. Molinero, X. Shen: Genentech employee and stock. C. O’Hear: Genentech employee and Roche stock. All other authors have declared no conflicts of interest.

1045PD

Adjuvant androgen deprivation therapy for high-risk androgen receptor-positive salivary duct carcinoma

W. Boxtel1, L.D. Locati2, S. Tooten1, E. Bos1, C. Bergamini2, A.C.H. van Engen-van Grunsven3, E. Boon1, E. Fiets4, S. Cavalieri2, G. Verhaegh5, J. Schalken5, L. Licitra2, C.M.L. Herpen1 1 Department of Medical Oncology, Radboudumc, Nijmegen, Netherlands, 2Medical Oncology Unit/Head and Neck Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy, 3Department of Pathology, Radboudumc, Nijmegen, Netherlands, 4Department of Medical Oncology, Medical Centre Leeuwarden, Leeuwarden, Netherlands, 5Laboratory of experimental urology, Radboudumc, Nijmegen, Netherlands Background: Salivary duct carcinoma (SDC) is a rare and aggressive subtype of salivary gland cancer, with a median disease-free survival (DFS) of less than 3 years. SDC is androgen receptor-positive (ARþ) in 67-96% of cases. In incurable recurrent ARþ SDC androgen deprivation therapy (ADT) has an overall response rate of 18-50%. In this study, high-risk ARþ SDC pts were treated with adjuvant ADT to study the efficacy. Methods: In this retrospective study, surgical resected pts who received adjuvant ADT for stage 4a/b ARþ SDC at the Radboudumc (Nijmegen, the Netherlands) or Istituto Nazionale dei Tumori (Milan, Italy) were collected. As control group, surgical resected pts diagnosed with stage 4a/b SDC between 1990-2014, who did not receive adjuvant ADT were collected by a search of the Dutch pathology database (PALGA). Pts were analyzed for DFS and overall survival (OS) by using Kaplan-Meier survival curves. Results: 18 ARþ SDC pts (median age 64 years [range 32-80]) were treated with adjuvant ADT (Nijmegen n ¼ 11; Milan n ¼ 7) for a median duration of 10 months [range 2-31 months]. All pts had a nuclear AR-staining pattern in > 70% of the cells. They were treated with bicalutamide monotherapy (n ¼ 10), a LHRH analog (n ¼ 1) or a combination of these (n ¼ 7). Treatment was well tolerated. 17/18 pts (94.4%) also received postoperative radiotherapy, of which 4 pts received concurrent chemoradiotherapy (22.2%). The control group consisted of 110 SDC pts (median age 70 years [range 44-100]). 103/110 pts (93.6%) received postoperative radiotherapy, of which 1 pt received concurrent chemoradiotherapy (0.9%). After a median follow-up of 22 months in the ADT-treated SDC pts and 25 months in the control SDC pts, the 3-year DFS was 53.6% and 34.2% (p ¼ 0.137), the 3-year OS was 68.2% and 52.3% (p ¼ 0.198), respectively. The median DFS and OS were not reached in the ADT-treated SDC pts and were 21 months and 46 months in the control SDC pts.

doi:10.1093/annonc/mdx374 | 373

abstracts Conclusions: Adjuvant ADT in high-risk ARþ SDC pts did not lead to a significant increase in DFS or OS, but the number of treated pts was limited. Due to the rarity of the disease we could not perform a formal phase II study. Translational research to identify pts which may benefit from ADT is warranted. Legal entity responsible for the study: Carla M.L. van Herpen Funding: None Disclosure: All authors have declared no conflicts of interest.

1046PD

Overexpression of the c-MET proto-oncogene in salivary duct carcinoma patients

C.M.L. Herpen1, W. Boxtel1, U.E. Flucke2, E. Bloemena2, E. Boon1, G. Verhaegh3, M.W. Aalders3, M.A. Jonker4, J. Schalken3, Y.M.H. Versleijen-Jonkers1 1 Department of Medical Oncology, Radboud University Medical Centre, Nijmegen, Netherlands, 2Department of Pathology, Radboud University Medical Centre, Nijmegen, Netherlands, 3Laboratory of experimental urology, Radboudumc, Nijmegen, Netherlands, 4Department for Health Evidence, Radboud University Medical Centre, Nijmegen, Netherlands Background: Salivary duct carcinoma (SDC) is a rare and aggressive subtype of salivary gland cancer (SGC). Activation of the cellular MET (c-MET) receptor tyrosine kinase has been implicated in cell proliferation, survival, migration, and invasion. The aim of this study was to evaluate the frequency of MET overexpression and its correlation to clinicopathological factors in SDC. Methods: 136 patients were collected by a retrospective search of the Nationwide Network and Registry of Histo- and Cytopathology (PALGA) in the Netherlands. Formalin-fixed paraffin-embedded tumor blocks and hematoxylin and eosin stained slides were requested for pathological review. MET expression was evaluated by immunohistochemical staining on primary tumors. These data were correlated to clinicopathological factors. Results: c-MET was positive in 54 of 136 tumors (39.7%). Of these 54 tumors, 50 had a cytoplasmic staining pattern and 23 had a membranous staining pattern, so in 19 tumors both cytoplasmic and membranous staining was observed. No correlations were found between cytoplasmic or membranous MET and high stage disease (stage 3 and 4 versus stage 1 and 2, p ¼ 0.606 and p ¼ 0.300 respectively), number of positive lymph nodes (p ¼ 0.263 and p ¼ 0.955 respectively), lymph node ratio (p ¼ 0.192 and p ¼ 0.771 respectively), androgen receptor-status (p ¼ 0.858 and p ¼ 0.258 respectively), HER2-status (p ¼ 0.257 and p ¼ 0.595 respectively), time to recurrence (p ¼ 0.559 and p ¼ 0.959 respectively), time to distant metastases (p ¼ 0.398 and p ¼ 0.666 respectively), or overall survival (p ¼ 0.754 and p ¼ 0.516 respectively). Membranous MET staining occurred more frequently in SDC ex pleomorphic adenoma (14 of 52 tumors) than in the ‘de novo’ SDC (9 of 84 tumors) (p ¼ 0.014). In SDC ex pleomorphic adenoma we also found more HER2-positive tumors (p ¼ 0.041). Conclusions: Cytoplasmic and membranous MET are overexpressed in SDC and may be a target for MET-targeted therapy. It is not a prognostic factor for overall survival, possibly because frankly invasive SGCs often show less receptor expression then minimally invasive SGCs or pleomorphic adenomas. The higher expression of c-MET and HER-2 in SDC ex pleomorphic adenoma needs further investigation. Legal entity responsible for the study: Carla M.L. van Herpen Funding: None Disclosure: All authors have declared no conflicts of interest.

1047PD

Mammary analogue secretory carcinoma (MASC): clinical characteristics in 28 ETV6-NTRK3 fusion gene confirmed patients

E. Boon1, M.H. Valstar2, W.T.A. van der Graaf1, E. Bloemena3, S.M. Willems4, C.A. Meeuwis5, L.A. Smit6, M.A.W. Merkx7, R.P. Takes8, U.E. Flucke9, C.M.L. Herpen10 1 Department of Medical Oncology, Radboud University Medical Centre, Nijmegen, Netherlands, 2Department of Head and Neck Oncology and Surgery, Antoni van Leeuwenhoek/Netherlands Cancer Institute, Amsterdam, Netherlands, 3Department of Pathology, VU University Medical Centre, Amsterdam, Netherlands, 4Department of Pathology, University Medical Centre Utrecht, Utrecht, Netherlands, 5Department of Otorhinolaryngology/Head & Neck Surgery, Erasmus University Medical Centre, Rotterdam, Netherlands, 6Department of Pathology, Antoni van Leeuwenhoek/Netherlands Cancer Institute, Amsterdam, Netherlands, 7Department of Oral and Maxillofacial Surgery, Radboud University Medical Centre, Nijmegen, Netherlands, 8Department of Otorhinolaryngology, Radboud University Medical Centre, Nijmegen, Netherlands, 9Department of Pathology, Radboud University Medical Centre, Nijmegen, Netherlands, 10Department of Medical Oncology, Radboud University Medical Centre, Nijmegen, Netherlands Background: Recently, a new subtype of salivary gland cancer (SGC), mammary analogue secretory carcinoma (MASC), has been defined, which is characterized by the presence of ETV6-NTRK3 fusion gene. Previously, MASC was mixed up with acinic cell carcinoma (AciCC), polymorphous low grade adenocarcinoma and (cyst)adenocarcinoma. At present the clinical features and outcome of MASC patients are not well known. We aimed to describe the clinical presentation and outcome of MASC. Methods: Firstly, we re-evaluated the pathological diagnosis of salivary gland cancers with a morphological resemblance to MASC, diagnosed in 4 of the 8 head and neck centres in the Netherlands, for their presence of ETV6-NTRK3 and also included genetically

374 | Head and neck cancer, excluding thyroid

Annals of Oncology confirmed prospectively diagnosed cases. The ETV6-NTRK3 fusion gene was analyzed using RT-PCR. Secondly, the clinical characteristics were retrieved from the patient files. Results: Twenty-eight patients with ETV6-NTRK3 fusion gene positive MASC were included (10 prospectively and 18 retrospectively). Of these 18 retrospective patients 13 patients were previously diagnosed as AciCC, the other 5 patients as (low-grade) adenocarcinoma. The median age at diagnosis was 49 years (range 19 83 years), 15 patients (54%) were male. The duration of symptoms varied from 6 weeks until 20 years with a median of 14 months. In 18 patients (64%) the tumour was located in the parotid gland; the other patients had tumours of the minor salivary glands (2), submandibular gland (1), oral mucosa/lip (5) or palate (2). All patients had a T1-2 tumour. One patient had lymph node metastasis at diagnosis. All patients underwent surgery of which 4 patients needed re-resection and 12 patients (43%) underwent postoperative radiotherapy. One patient had a local recurrence 50 months after primary surgery, but was cured after second resection. None of the patients had regional recurrences or distant metastases. The median follow-up was 49 months and both the 5- and 10-year overall survival rate were 94%. Conclusions: MASC is a recently acknowledged new entity of SGC characterized by the ETV6-NTRK3 fusion gene. The clinical course seems to be favourable with a very low rate of recurrences and an excellent survival. Legal entity responsible for the study: Radboudumc Funding: None Disclosure: All authors have declared no conflicts of interest.

1048PD

RetroSpective cohort stUdy of PD-L1 expression in REcurrent and/or MEtastatic squamous cell carcinoma of the head and neck (SUPREME-HN)

S. Pai1, E.E. Cohen2, D. Lin3, G. Fountzilas4, E.S. Kim5, H. Mehlhorn6, N. Baste7, D. Clayburgh8, L. Lipworth9, C. Resteghini10, N. Shara11, T. Fujii12, J. Zhang13, M. Stokes14, D. Lawrence15, A. Khaliq16, G. Melillo17, N. Shire18 1 Department of Surgery, Massachusetts General Hospital, Cancer Center, Boston, MA, USA, 2Department of Medicine, UC San Diego Health System, Moores Cancer Center, San Diego, CA, USA, 3Department of Otolaryngology, Massachusetts General Hospital, Cancer Center, Boston, MA, USA, 4Medical Oncology, Aristotle University of Thessaloniki, Thessaloniki, Greece, 5Department of Solid Tumor Oncology, Levine Cancer Institute, Carolinas Health Care System, Charlotte, USA, 6Head and Neck Surgery and Department of Head Medicine and Oral Health, Universitaetsklinikum Leipzig, Klinik und Poliklinik fur HNO-Heilkunde, Leipzig, Germany, 7Oncology, Hospital Universitari Vall d’Hebron, Barcelona, Spain, 8Department of Otolaryngology/Head and Neck Surgery, Oregon Health & Science University, Portland, OR, USA, 9Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA, 10Head and Neck Medical Oncology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 11 Biostatistics and Biomedical Informatics, MedStar Health Research Institute, Hyattsville, MD, USA, 12Otolaryngology, Osaka Medical Center for Cancer and Cardiovascular Diseases, Osaka, Japan, 13Oncology, Baylor College of Medicine, Houston, TX, USA, 14Real-World Evidence, Evidera, Lexington, MA, USA, 15Biostatistics & Information Sciences, AstraZeneca, Cambridge, UK, 16Global Medical Affairs (Oncology), AstraZeneca, Gaithersburg, MD, USA, 17Immuno-Oncology GMD, AstraZeneca, Gaithersburg, MD, USA, 18Global Medicines Development, AstraZeneca, Gaithersburg, MD, USA Background: Clinically meaningful antitumour activity and improved overall survival (OS) in recurrent and/or metastatic (R/M) head and neck squamous cell carcinoma (HNSCC) have been achieved by targeting the PD-1/PD-L1 axis. Tumoral PD-L1 expression correlates with response to blocking PD-1/PD-L1 antibodies. In a retrospective study, we investigated tumoral PD-L1 expression as a prognostic biomarker in R/M HNSCC patients (pts) treated with standard of care (SOC) therapy. Methods: Archival tumor samples from R/M HNSCC pts diagnosed between March 2011 and June 2015 at 19 institutions in 7 countries were evaluated for PD-L1 expression using the validated Ventana SP263 assay and scored as PD-L1 high (25% of tumor cells [TC]) or low/negative (10%) - 17 mo (95%CI 3-36) with strong trend of increasing OS with rise of PD-L1 expression in immune cells. Conclusions: PD-L1 expression in tumor and immune cells are favourable prognostic factors for oral cavity squamous cell carcinoma. Legal entity responsible for the study: Svetlana Kutukova Funding: None Disclosure: All authors have declared no conflicts of interest.

1062P

Immune profile analysis of head and neck squamous cell carcinoma before and after neoadjuvant treatment with the IRX-2 regimen

G.T. Wolf1, J.G. Newman2, N.L. Berinstein3, M.J. McNamara4, A. Nguyen1, J.E. Egan5, M.J. Kaplan6 1 Otolaryngology-Head and Neck Surgery, University of Michigan, Ann Arbor, MI, USA, 2 Otolaryngology, U of Penn, Philadelphia, PA, USA, 3Medical Oncology, Sunnybrook Health Sciences Center, Toronto, ON, Canada, 4Chiles Research Institute, Providence Portland Medical Center, Portland, OR, USA, 5Reearch, IRX Therapeutics, New York, NY, USA, 6Otolaryngology, Stanford University, Stanford, CA, USA Background: IRX-2 is an injectable cancer immunotherapy composed of cytokines purified from stimulated peripheral blood mononuclear cells. In a phase 2a trial

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology (n ¼ 27), neoadjuvant IRX-2 significantly increased lymphocyte infiltration (LI) into resected head and neck tumors. Increased LI was associated with changes in fibrosis and necrosis in resected tumors, 65% event-free survival (EFS) at 2 years, and 65% overall survival (OS) at 5 years, better than rates for historical matched controls. Patients with LI greater than the median had improved OS compared to those below the median. This substudy was undertaken to define the mechanisms responsible for the increase in LI with neoadjuvant IRX-2. Methods: Matched pre- and post-treatment tumor specimens from 7 phase 2a study patients were interrogated with two immune-profiling technologies, multiplex immunohistochemistry (IHC, PerkinElmer, Waltham, MA) and transcriptome analysis (NanoString Technologies, Seattle, WA). Results: Multiplex IHC provided detailed visualization and quantitation of various immune cells in the tumor microenvironment (TME), supporting previous phase 2a pathology findings. Transcriptome analysis provided a global snapshot of the TME, quantitative information on immune cell subsets, and insights into possible mechanisms for changes in LI. Consistent with IRX-2 activation of multiple immune cells in the TME, mRNA expression of B cell, CD4þ T cell, CD8þ T cell, and dendritic cell functional genes was increased on average by 87%, 106%, 6%, and 130%, respectively, following treatment with IRX-2. Increases in chemokine gene expression were observed, suggesting that IRX-2-induced production of chemokines may in part drive tumor LI. Strong evidence of functional immune activation uncovered by transcriptome analysis included an increase in interferon g pathway gene expression and induction of regulatory checkpoint pathways. Conclusions: Neoadjuvant IRX-2 promotes tumor LI and prolongs EFS and OS in patients with head and neck squamous cell carcinoma. Immune profile analyses provided insights into the pathways potentially responsible for IRX-2-induced increases in LI and overall immune activation. Legal entity responsible for the study: IRX Therapeutics, Inc. Funding: IRX Therapeutics, Inc Disclosure: G.T. Wolf: Corporate sponsored research with IRX Therapeutics, Inc. J.G. Newman, M.J. Kaplan: Research funding IRX Therapeutics. N.L. Berinstein: Consultant, IRX Therapeutics. M.J. McNamara: Research funding IRX Therapeutics. J.E. Egan: Employed by IRX Therapeutics. All other authors have declared no conflicts of interest. 1063P

Phenotyping of the immune infiltrate in oropharyngeal squamous cell carcinoma: Focus on materials and methods

A. De Meulenaere1, T. Vermassen1, S. Rottey1, L. Ferdinande2 Department of Medical Oncology, Ghent University Hospital, Ghent, Belgium, 2 Department of Medical and Forensic Pathology, Ghent University Hospital, Ghent, Belgium

1

Background: Stromal CD8þ lymphocytic infiltration constitutes an independent prognostic marker for better overall survival in patients with oropharyngeal squamous cell carcinoma (OSCC). However, scoring of (novel) biomarkers is often complicated by the lack of standardised methodology which hampers their use in daily clinical practice. We therefore performed a comparative analysis to evaluate the importance of choice of materials and methods in CD8 assessment in patients with OSCC. Other immune cell markers, that is, CD3 and FoxP3 were taken into account as well. Methods: Immunohistochemical analysis of CD3, CD8 and FoxP3 was performed on whole-tissue sections from 101 treatment-naive patients with OSCC. A comparison of different immune cell markers was made for biopsy material versus resection material when available. Also, different scoring strategies, that is, quantitative versus semiquantitative analysis were compared with each other. Results: Comparison of biopsy material versus resection material proved a good agreement for expression of the CD3þ T cells (j ¼ 0.712, q ¼ 0.853), CD8þ T cells (j ¼ 0.659, q ¼ 0.764) and FoxP3þ T cells (j ¼ 0.783, q ¼ 0.802). The comparison of

quantitative versus semi-quantitative assessment demonstrated strong correlations for the CD3þ T cells (qbiopt ¼ 0.617, qresection ¼ 0.634), CD8þ T cells (qbiopt ¼ 0.656, qresecþ tion ¼ 0.675) and FoxP3 T cells (qbiopt ¼ 0.537, qresection ¼ 0.720). Conclusions: Immunohistochemical analysis of CD3, CD8 and FoxP3 can be performed adequately on biopsy as this appears to be representative for the whole tumour. In addition, this can be done adequately in a semi-quantitative manner without the use of more expensive and time consuming machinery. Legal entity responsible for the study: Prof. Sylvie Rottey Funding: agency for Innovation by Science and Technology (IWT) Disclosure: All authors have declared no conflicts of interest.

1064P

High-dose versus low-dose cisplatin with definitive concurrent radiotherapy for squamous cell carcinoma of the head and neck (SCC): An analysis of veteran’s health registry data

J. Bauml1, R. Vinnakota2, Y-H.A. Park3, S.E. Bates2, T. Fojo2, C. Aggarwal4, S. Lunate3, N. Damjanov1, R. Mamtani4, C.J. Langer4, R.B. Cohen4, K. Sigel5 1 Hematology/Oncology, Corporal Michael J Crescenz VA Medical Center, Philadelphia, PA, USA, 2Hematology/Oncology, Columbia University Medical Center, New York, NY, USA, 3Hematology/Oncology, James J Peters VA Medical Center, New York, NY, USA, 4 Hematology/Oncology, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA, 5Infectious Diseases, James J Peters VA Medical Center, New York, NY, USA Background: Radiotherapy (RT) with concurrent high-dose cisplatin (HDC) improves outcomes for patients (pts) with SCC. Weekly low-dose cisplatin (LDC) is a widely used alternative approach. The comparative effectiveness and safety of these approaches is unknown. We compared the outcomes of pts treated with HDC and LDC within the Veteran’s Administration Corporate Data Warehouse (CDW). Methods: We identified stage III-IVb SCC patients treated non-surgically with RT and HDC or LDC from 2002 to 2014 in the CDW. Pts were grouped by the dose of their first cycle (HDC vs LDC; intent-to-treat). Variables including cancer site, stage, smoking/alcohol use, and comorbidities were used to generate propensity scores (PS) for the use of HDC. We compared overall survival (OS) by treatment group using Cox regression models, adjusting for PS. We also determined the risk of toxicities using PS-adjusted logistic regression models. Results: A total of 2,820 pts were included in the analysis: 69.7% received HDC (mean initial dose 96 mg/m2). The mean initial dose of LDC was 30 mg/m2. HDC pts were younger (p < 0.001), with lower creatinine (p ¼ 0.002), and lower incidence of baseline neuropathy (p ¼ 0.02). In an unadjusted analysis, HDC was associated with improved OS (Table). After PS adjustment, this difference was no longer statistically significant (p ¼ 0.06). On primary site sub-analysis, HDC provided a benefit only for oropharyngeal primaries (OP). Adjusting for PS, HDC was associated with more renal failure (OR 2.2, 95% CI 1.6-3), neutropenia (OR 2, 95% CI 1.1-3.5), dehydration/electrolyte disturbance (OR 1.3, 95% CI 1.04-1.6), and hearing loss (OR 1.6, 95% CI 1.3-2). Conclusions: While HDC does not improve OS over LDC for the overall cohort of patients with SCC receiving RT with definitive intent, it is associated with a survival benefit for patients with OP. HDC is associated with more adverse events. Legal entity responsible for the study: Keith Sigel Funding: None Disclosure: J. Bauml: Research Support: Merck, Bayer, Novartis, Carevive Systems Consulting: Bristol-Myers Squibb, Boehringer Ingelheim, AstraZeneca, Celgene, Genentech, Merck, Guardant Health. C. Aggarwal: Roche, Bristol-Myers Squibb, Eli Lilly: Ad boards Takeda, Macrogenics, Roche: Research support to institution. C.J. Langer: Research: Pfizer, Lilly, Genentech, OSI; GSK; Clovis; Merck; Nektar; Advantagene; Inovio; Takeda; AbbVie Advisor: Bristol-Myers Squibb, ImClone, Pfizer, Lilly; AstraZeneca; Merck; Novartis; Genentech; Bayer; Celgene; Abbott; Biodesix;

Table: 1064P OS of HDC vs LDC Group

Unadjusted HR for OS

All patients (n ¼ 2,820) Oral Cavity (n ¼ 182) Hypopharynx/Larynx (n ¼ 1,026) Oropharynx (n ¼ 1590)

0.85 0.77 1.00 0.78

Volume 28 | Supplement 5 | September 2017

95% CI 0.77-0.94 0.56-1.1 0.86-1.20 0.70-0.90

PS Adjusted HR for OS

95% CI

0.89 0.72 1.1 0.81

0.80-1.01 0.50-1.10 0.90-1.30 0.69-0.96

doi:10.1093/annonc/mdx374 | 381

abstracts Clariant; Caris; ARIAD; Boehringer Ingelheim; Synta; Clovis; Amgen; Synta; Peregrine; Incyte. R.B. Cohen: Takeda scientific ad board Zymeworks, Bristol-Myers Squibb. All other authors have declared no conflicts of interest.

1065P

3-weekly or weekly cisplatin concurrently with radiotherapy for patients with locally advanced squamous cell carcinoma of the head and neck: A multicentre, retrospective analysis

S.I. Rothschild1, S. Helfenstein1, O. Riesterer2, U.R. Meier3, A. Papachristofilou4, B. Kasenda1, M. Pless5 1 Medical Oncology, University Hospital Basel, Basel, Switzerland, 2Radiation Oncology, University Hospital Zu¨rich, Zurich, Switzerland, 3Radiation Oncology, Cantonal Hospital Winterthur, Winterthur, Switzerland, 4Clinic for Radiotherapy and Radiation Oncology, University Hospital Basel, Basel, Switzerland, 5Medical Oncology, Cantonal Hospital Winterthur, Winterthur, Switzerland Background: Concurrent chemoradiotherapy with cisplatin is standard for patients (pts) with loco-regionally advanced squamous cell carcinoma of the head and neck (LA-SCCHN). The standard regimen includes 3-weekly cisplatin, but weekly regimens are often used to lower toxicity. Reaching a cumulative dose of > 200 mg/m2 cisplatin was shown being associated with improved outcome. We herein investigated cumulative dose reached and toxicity between the both widely used 3-weekly and weekly cisplatin regimens with concurrent radiotherapy. Methods: Multicentre, retrospective analysis of all patients with LA-SCCHN treated at 3 centers in Switzerland between 06/2008 and 12/2015. We used descriptive statistics and logistic regression (uni- and multivariable) to investigate the association between the chosen cisplatin regimen (weekly versus 3-weekly) and the chance to reach the cumulative cisplatin dose of > 200 mg/m2. Landmark approach (8 weeks after start of treatment) was applied for investigating the prognostic impact of the cumulative cisplatin dose on survival using Cox regression techniques. Results: We included 314 eligible pts (3-weekly schedule, N ¼ 127; weekly schedule, N ¼ 187). Median cumulative cisplatin dose was 200 mg/m2 (IQR 150-300) for pts treated with a 3-weekly schedule and 160 mg/m2 (120-240) for the weekly schedule, consequently more pts treated with a 3-weekly schedule reached a cumulative dose >200 mg/m2 (75.6% vs. 47.1%, p < 0.001). This association was also observed in multivariable analysis adjusted for age and sex (OR 3.46, 95% confidence interval [CI], 2.1 - 5.7). The 3-weekly regimen led to a higher rate of renal toxicity (33.1% vs. 20.9%, p ¼ 0.022), but not ototoxicity (15% vs. 12.8%). In the landmark analysis, we could not confirm that a cisplatin dose >200 mg/ m2 is associated with better survival (HR 1.3, 95% CI 0.8 -1.9). Conclusions: Significantly more patients receive a cumulative of dose of > 200 mg/m2, when treated with a 3-weekly schedule compared to weekly dosing. This comes at the cost of more renal toxicity. Due to the non-randomized nature of this analysis, no conclusions on the efficacy of the respective schedules should be drawn. Legal entity responsible for the study: University Hospital Basel Funding: None Disclosure: All authors have declared no conflicts of interest.

1066P

Hyperfractionated twice daily re-irradiation (bid re-RT) and chemotherapy (CT) for locoregionally recurrent head and neck squamous cell carcinoma (LR HNSCC): A systematic review (SR)

D. Day1, G. Locke2, K. Chan3, S. Cheng3, L. Wang4, B. O’Sullivan2, A.R. Hansen1 Department Medical Oncology and Hematology, Princess Margaret Cancer Centre, Toronto, Canada, 2Department of Radiation Oncology, Princess Margaret Cancer Centre, University of Toronto, Toronto, ON, Canada, 3Medical Oncology, Sunnybrook Odette Cancer Centre, Toronto, ON, Canada, 4Biostatistics Department, Princess Margaret Cancer Centre, Toronto, Canada

1

Background: Re-RT þ/- CT is a salvage option for patients (pts) with LR HNSCC in a previously radiated field, although efficacy and toxicity, and the optimal treatment regimen remains undefined due to a lack of randomized trials. Hyperfractionated bid reRT, by reducing the dose per fraction may improve radiation (RT) therapeutic ratio and is increasingly used in LR HNSCC. The aim of this SR is to assess the treatment outcomes of bid re-RT þ CT in LR HNSCC. Methods: We conducted a SR of MEDLINE, EMBASE and the Cochrane library up to Nov 2016 for clinical trials of bid re-RT þ CT in pts with LR HNSCC. Paired reviewers selected studies for inclusion and extracted data. Individual patient level overall survival (OS) data were extracted where possible from published Kaplan-Meier (KM) curves to construct an aggregate KM curve. Results: We identified 10 clinical trials (all were phase 1 or 2) with 404 pts. Median (of reported medians) prior RT dose was 64Gy, and median time from prior RT was 30.9m. Seventy-three and 156 pts respectively had CT and surgery as part of 1st line treatment. Median re-RT dose was 60Gy administered as continuous or split courses. The re-RT fields consisted of gross tumor volume plus a minimum margin (range 12cm). All CT regimens were combinations either with cisplatin (n ¼ 6) or 5-FU (n ¼ 4), given concurrently with (n ¼ 9) or prior to (n ¼ 1) re-RT. Twenty-eight (7%) pts had debulking surgery prior to re-RT. In pts who were analyzable for toxicities, acute events (>/¼grade 3) were reported in 252 of 377 (67%) pts and late events (>90d post re-RT) in 87 of 333 (26%) pts. Treatment-related deaths occurred in 26 (6%) pts, mostly due

382 | Head and neck cancer, excluding thyroid

Annals of Oncology to infection or vascular events. Of the 5 trials with extractable KM curves, estimated median OS was 10.2m (95% CI 8.7-12.6m); 1- and 3-y OS rates were 46.8% and 11.2% respectively. No differences were observed in median OS and toxicity rates based on CT type (Wilcoxon test). Conclusions: This is the 1st aggregate analysis of bid re-RT and CT in LR HNSCC. Longterm OS was observed in a subset of pts, however treatment-related morbidity was apparent. The optimal re-RT and CT regimen is still undefined and further study is required. Legal entity responsible for the study: None Funding: None Disclosure: All authors have declared no conflicts of interest.

1067P

Raltitrexed versus 5-fluorouracil with cisplatin and concurrent radiotherapy (CCRT) for locally advanced head and neck squamous cell carcinoma (LA-HNSCC): A randomized controlled multi-centered trial

X. He1, H. Yin2, W. Guo1, X. Sun3, X. Hu2, P. Yan1, F. Li1, S. Huang1, H. Zhou4, F. Zhang5, B. Wang6, X. Bian1, Q. Wei1, J. Wu1, F. Wang1, D. Song1, J. Liu1, B. Zhou3 1 Department of Radiation Oncology, Nanjing Medical University Afﬁliated Cancer Hospital, Jiangsu Cancer Hospital, Jiangsu Institute of Cancer Research, Nanjing, China, 2 Department of Radiotherapy Oncology, Xuzhou Central Hospital, Afﬁliated to School of Medicine, Southeast University, Xuzhou, China, 3Department of Radiotherapy, Cancer Center of PLA 81 Hospital, Nanjing, China, 4Department of Radiotherapy Oncology, Nanjing BenQ Hospital, Nanjing, China, 5Department of Radiotherapy Oncology, Fourth People’s Hospital of Wuxi, Wuxi, China, 6Department of Radiotherapy Oncology, Northern Jiangsu Province People’s Hospital, Yangzhou, China Background: CCRT has been considered to be the standard of treatment for LAHNSCC. However, patients receiving CCRT experience a substantial number of treatment-related adverse events, primarily causing oropharyngeal mucositis(OM) and leading to interruption or discontinuation of treatment. Raltitrexed is a specific thymidylate synthase inhibitor with a convenient administration, acceptable toxicity and radiosensitizing property, as the published phase I/II trials have shown. This study aimed to compare the clinical efficacy and toxicity of cisplatin with raltitrexed (RP) or 5-fluorouracil (FP) for LA-HNSCC. Methods: Eligible patients with LA-HNSCC were randomly assigned in a 1:1 ratio to receive CCRT with either RP or FP. The RP group consisted of 2.5mg/m2 intravenous raltitrexed on day 1 and 25 mg/m2 intravenous cisplatin on days 1-3. The FP group consisted of continuous intravenous infusions of 600 mg/m2 5-fluorouracil on days 1-5 and 25 mg/m2 intravenous cisplatin on days 1-3. Chemotherapy was administrated concurrently with radiotherapy and was repeated every 3 weeks with completion of at least 2 cycles. Primary endpoint was PFS. Secondary endpoints were complete response rates(CRR), OS and safety. Results: A total of 108 patients with LA-HNSCC enrolled in this study, with 52 patients assigned to the RP group and 56 patients to the FP group. There was no significant difference in CRR between the two arms (42.9% vs 26.8%, respectively, p ¼ 0.074), with the RP group showing a trend of increased CRR. Data of locoregional control, patterns of failure, and survival required further follow-up. The most frequent acute toxicities were bone marrow suppression, gastrointestinal side effects and OM in both arms. The incidence rate of severe OM was significantly lower (P0.05). Conclusions: The efficacy of the RP regimen was similar to that of the FP regimen. The RP regimen had a tolerable safety profile, with a lower incidence of severe OM and, consequently, an improved quality of life. In conclusion, RP is an effective, well-tolerated regimen for LA-HNSCC. Clinical trial identification: NCT02485548 Release date: June 26, 2015 Legal entity responsible for the study: Xia He Funding: None Disclosure: All authors have declared no conflicts of interest.

1068P

The observational ENCORE study: Cetuximab 1 platinum-based therapy (PBT) for first-line (1L) treatment of patients with recurrent/ metastatic squamous cell carcinoma of the head and neck (R/M SCCHN)

C. Le Tourneau1, M. Ghiani2, M.C. Cau2, R. Depenni3, G. Ronzino4, L. Livi5, V. Montesarchio6, M. Bretagne1, M. Saint-Ghislain1, J. Schulten7, D. Messinger8, A. Sbrana9, M.G. Ghi10 1 Medical Oncology, Institut Curie, Paris, France, 2Medical Oncology, Ospedale Oncologico Armando Businco, Cagliari, Italy, 3Medical Oncology, Azienda Ospedaliero Universitaria Policlinico di Modena, Modena, Italy, 4U. O. Oncologia Medica, Ospedale Vito Fazzi, Lecce, Italy, 5Radiation Oncology, University of Florence, Florence, Italy, 6PneumoOncology, AORN dei Colli; Monaldi Hospital, Naples, Italy, 7Medical Affairs, Merck KGaA, Darmstadt, Germany, 8Biostatistics, Prometris GmbH, Mannheim, Germany, 9Medical Oncology, Azienda Ospedaliero-Universitaria Pisana; Istituto Toscano Tumori, Pisa, Italy, 10 Medical Oncology, Ospedale SS. Giovanni e Paolo, Venice, Italy Background: The randomized, phase 3 EXTREME study established cetuximab þ platinum þ 5-flurouracil (5-FU) followed by cetuximab maintenance until progressive

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology disease (PD) as the first regimen to yield survival benefits in the 1L management of patients with R/M SCCHN. In EXTREME, the addition of cetuximab increased the overall response rate from 20% to 36%, and extended progression-free survival from 3.3 to 5.6 months and overall survival from 7.4 to 10.1 months. ENCORE is a multinational, non-interventional, prospective, open-label study, seeking to determine how treatment decisions are made, planned and executed by oncologists treating patients with 1L therapy for R/M SCCHN in the real world. Methods: ENCORE prospectively enrolled 219 patients with R/M SCCHN from Algeria, France, Italy, Portugal and Russia. The recommended treatment for these patients is cetuximab þ PBT for up to 6 cycles followed by cetuximab maintenance until PD. Patient characteristics, drugs and schedule were recorded; as the study is still ongoing, safety and efficacy will not be reported here. Results: ENCORE patients and the EXTREME patients who received cetuximab þ platinum þ 5-FU had similar performance status (PS: 13.7 and 12% with PS 2, respectively), but dissimilar median age (64 and 56 years, respectively). In ENCORE, 94.1% of patients had a planned treatment of cetuximab þ PBT with cetuximab maintenance until PD. The remaining 13 (5.9%) had a fixed treatment duration of 4 to 24 weeks. 37.9% of treatment plans used cisplatin, 61.6% included carboplatin and 3.2% used a taxane. Also, only 53.4% of plans included 5-FU. When developing the treatment plan, 72.1% of all patients were discussed within the context of a multidisciplinary team (MDT). Most plans had the goal of palliative care, and 80% were formulated without a p16 or human papillomavirus status test. Updated data will be presented at congress. Conclusions: The ENCORE study shows that a real-world R/M SCCHN patient population treated with the EXTREME regimen has diverse characteristics and is treated per current recommendations (e.g. in an MDT setting, with cetuximab until PD). Clinical trial identification: EMR 62202-566 Legal entity responsible for the study: Merck KGaA, Darmstadt, Germany Funding: Merck KGaA, Darmstadt, Germany Disclosure: C. Le Tourneau: Consultancy: Novartis, MSD, Bristol-Myers Squibb, AstraZeneca; Honoraria: Merck Serono J. Schulten: Full Time Employee: Merck KGaA. D. Messinger: Employee/Consultancy: Employee of Prometris GmbH, which has a contract with Merck KGaA regarding statistical consultancy. All other authors have declared no conflicts of interest.

1069P

Primary surgery versus chemoradiotherapy for advanced oropharyngeal and hypopharyngeal cancer: A propensity-score matched study using a nationwide database

C. Cheng Medical Research, Koo Foundation Sun Yat Sen Cancer Center, Taipei, Taiwan Background: Traditionally, advanced head and neck cancer has been managed through surgery with or without postoperative radiotherapy. Studies since the 1980s have been advocating organ preservation therapies by using various combinations of chemotherapy and radiotherapy. For treatment of advanced oropharyngeal and hypopharyngeal cancer, there has been a controversy in choosing between primary surgery and chemoradiotherapy. We aimed at conducting a propensity-score matched study from a national database to investigate the survival after primary surgery with or without postoperative radiotherapy versus chemoradiotherapy in patients with advanced oropharyngeal and hypopharyngeal cancer. Methods: We identified patients with stage III & IVa oropharyngeal and hypopharyngeal cancers between 2004 and 2009 from Taiwan National Health Insurance Claims Database. The study cohort was followed until 2012. We matched patients who received primary surgery to those who received chemoradiotherapy by propensity score calculated by logistic regression. Age at diagnosis, Charlson comorbidity index score, year of cancer diagnosis, clinical stage, receiving chemoradiotherapy, and receiving radiation therapy were well matched in these two groups. Overall survival and disease-free survival were compared using the KaplanMeier method. Results: We identified 1,603 oropharyngeal and 1,512 hypopharyngeal cancer patients. After propensity score matching, 614 patients with oropharyngeal cancer and 638 patients with hypopharyngeal cancer were included in the analysis. For advanced hypopharyngeal cancer (stage III and IVa), the overall survival and disease-free survival in patients receiving primary surgery with or without radiotherapy were statistically better than the matched sample who received chemoradiotherapy. For oropharyngeal cancer, the survival benefit only existed in stage IVa patients who received primary surgery with or without radiotherapy. Conclusions: The study showed that primary surgery with or without radiotherapy might have survival benefit in patients advanced oropharyngeal or hypopharyngeal cancer as compared to chemoradiotherapy. Legal entity responsible for the study: Koo Foundation Sun-yat Sen Cancer Center Funding: Health and Welfare Surcharge of Tobacco Products grant of Taiwan Disclosure: The author has declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

1070P

Comparison of carboplatin with 5-fluorouracil (carbo-5FU) versus cisplatin as concomitant chemoradiotherapy (CRT) for locally advanced head and neck squamous cell carcinoma (LAHNSCC)

S.H. Hanemaaijer1, I.C. Kok2, R.S.N. Fehrmann2, B. van der Vegt3, J.A. Gietema2, B.E.C. Plaat1, M.A.T.M. Vugt2, M.R. Vergeer4, J. Voortman5, C.R. Leemans6, J. Buter5, J.A. Langendijk7, S.F. Oosting2 1 Otorhinolaryngology, University Medical Center Groningen, Groningen, Netherlands, 2 Medical Oncology, University Medical Center Groningen, Groningen, Netherlands, 3 Pathology, University Medical Center Groningen, Groningen, Netherlands, 4 Radiotherapy, VU University Medical Center/Cancer Center, Amsterdam, Netherlands, 5 Medical Oncology, VU University Medical Center/Cancer Center, Amsterdam, Netherlands, 6Otolaryngology/, VU University Medical Center/Cancer Center, Amsterdam, Netherlands, 7Radiation Oncology, University Medical Center Groningen, Groningen, Netherlands Background: CRT including three cycles of cisplatin is considered the standard of care for LA-HNSCC. Around one third of the patients cannot complete cisplatin due to toxicity. Carbo-5FU can be used as an alternative. The aim of this study was to compare tolerability and efficacy between CRT with carbo-5FU and cisplatin. Methods: This is a retrospective analysis of patients with LA-HNSCC treated with concomitant CRT in two Dutch cancer centers between 2007-2016. All patients received intensity modulated radiotherapy. One center routinely administered carboplatin 300350 mg/m2 at day 1, 22 and 43 followed by 5FU 600mg/m2/day for 96 hours. The other center used cisplatin 100 mg/m2 at day 1, 22 and 43. Primary endpoint was chemotherapy completion rate. Secondary endpoints included: reason for discontinuation, number of unplanned admissions, overall survival (OS) and disease free survival (DFS). Associations between clinicopathological parameters and OS were determined with multivariate Cox regression analyses. Results: In the carbo-5FU cohort (n ¼ 190), 61.6% of the patients completed chemotherapy versus 76.7% (p ¼ 0.001) of the patients in the cisplatin cohort (n ¼ 223). Discontinuation caused by chemotherapy specific toxicity occurred twice as often in the carbo-5FU cohort (odds ratio 2.2, 95%CI, 1.38-3.5). Patients in the cisplatin cohort were more likely to have an unplanned admission (OR 2.96, 95%CI, 2.21-4.27). The risk of death was higher in the carbo-5FU cohort (HR 1.50, 95%CI, 1.06-2.12, p ¼ 0.02) with a three-year OS of 64.6% compared to 76.6% in the cisplatin cohort. Similar results were observed for DFS (HR 1.39, 95%CI, 0.99-1.93, p ¼ 0.06). T-classification, Nclassification, smoking and p16 status were independently associated with OS, but chemotherapy regimen was not (HR 1.04, 95%CI, 0.72-1.51, p ¼ 0.84). Conclusions: Patients treated with carbo-5FU less frequently completed chemotherapy because of chemotherapy specific toxicity. Better OS was observed in the cisplatin cohort, but chemotherapy regimen was not independently associated with OS. Legal entity responsible for the study: S.F.Oosting Funding: None Disclosure: J.A. Gietema: Research grant to institution: Roche, Siemens. B.E.C. Plaat: Research grant to institution: Olympus NL. J.A. Langendijk: Coporate sponsored research with RaySearch, Phillips, Mirada and IBA. S.F. Oosting: Research grant to the institutuion from Pfizer. All other authors have declared no conflicts of interest.

1071P

Postoperative radiotherapy with weekly cisplatin in locally advanced head and neck cancer

A. Hervas, J. Domınguez, D. Candini, M. Martın, C. Vallejo Radiation Oncology, Hospital Universitario Ramon y Cajal, Madrid, Spain Background: The aim of this study is to analyze the results of over 10 years experience with postoperative radiochemotherapy in patients diagnosed of locally advanced squamous cell carcinoma of head and neck based in cisplatin 40mg/m2/week. Methods: From March 2004 to August 2015, 104 patients were treated in our department. All patients received chemoradiation therapy with adjuvant intent based in the same scheme: radiotherapy 50 Gy to clinical target volume (CTV) and 66-70 Gy to areas with close margin or extracapsular lymph node involvement and weekly cisplatin at 40mg/m2 concomitant to radiotherapy. Results: The median age was 59 years (range 36-76), 85 patients were male (81.7%) and 19 were female (18.3%). The pathological stage was: 2.9 stage II, 8.7% stage III and 88.4% stage IV. Locations: 38.5% larynx, 38.5% oral cavity, 17.3% oropharynx, and 5.8% hypopharynx. 76.2% of patients received at least 5 cycles of chemotherapy. G3 toxicity was observed in 33% of patients being mucositis and epitelitis the most frecuents. G4 toxicity was not detected in any patient. Median follow-up was 81 months (range 18-137). Two-year and five-year overall survival (OS) were 90% and 76% respectively and disease-free survival (DFS) were 69.07% and 52.57% respectively. In multivariate analysis two or more positive nodes and longer time between surgery and

doi:10.1093/annonc/mdx374 | 383

abstracts onset of radiotherapy were significant predictors of poorer OS and extracapsular extension, positive margin and longer time between surgery and onset of radiotherapy were significant predictors of poorer DFS. Conclusions: In our serie, postoperative radiochemotherapy based in weekly cisplatin at 40mg/m2 in patients diagnosed of locally advanced squamous cell carcinoma of head and neck offers a good toxicity profile and results comparable to those published in the literature with 3-weekly cisplatin scheme. The number of positive nodes, longer time between surgery and onset of radiotherapy, extracapsular extension and positive margin were desfavorable prognostic factor related with SLE and OS in the multivariate analysis. Legal entity responsible for the study: Hospital Ram on y Cajal Funding: None Disclosure: All authors have declared no conflicts of interest.

1072P

Effectiveness and toxicities of cetuximab in combination with concurrent chemoradiotherapy in locoregionally advanced nasopharyngeal carcinoma: A propensity score-matched analysis

L-R. Wu, J-H. Xu, W-J. Guo, X. HeDepartment of Radiation Oncology, Nanjing Medical University Afﬁliated Cancer Hospital, Jiangsu Cancer Hospital, Jiangsu Institute of Cancer Research, Nanjing, China

Background: There is increasing evidence showing that concurrent chemoradiotherapy (CCRT) may be inadequate for patients with locoregionally advanced nasopharyngeal carcinoma. Until now, no randomized controlled clinical trial has proved the effectiveness of cetuximab plus CCRT. Methods: There were 681 consecutive stage III-IVB NPC were included in this retrospective study.75 underwent CCRT with cetuximab and 606 received CCRT.The nasopharyngeal and neck tumor of all patients were treated by intensity modulation radiated therapy (IMRT). Results: After matching at a 1:2 ratio, 150 patients were treated with CCRT and 75 with CCRT plus C were selected. The 3-year PFS rates (83.7% vs 72.0%, P ¼ 0.036) and 3year LRFS rates (98.6% vs 90.2%, P ¼ 0.034) were higher for patients in the CCRT plus C arm than with CCRT alone. Furthermore, a marginal trend of increasing risk of 3year DMFS rates (83.9% vs 78.4%, P ¼ 0.301) and 3-year OS rates (91.2% vs 85.8%, P ¼ 0.123) was found. The results indicated that CCRT plus C treatment was a significant and independent protective predictor for 3-year PFS (P ¼ 0.015) and LRFS rates(P ¼ 0.047). When focusing on stage T4 and/or N3 in the subgroup, the CCRT plus C arm achieved significantly prolonged 3-year PFS (79.9% vs 62.6%, P ¼ 0.022) and a marginally increased OS (88.0% vs 77.9%, P ¼ 0.086) compared with that of CCRT alone. Additionally, the 3-year LRFS (97.0% vs 90.9%, P ¼ 0.246) and DMFS (79.9% vs 67.8%, P ¼ 0.161) were enhanced in patients with CCRT plus C compared to CCRT alone. When concentrating on stage III patients, there were no considerable statistically significant differences found in 3-year PFS, OS, LRFS, and DMFS rates between patients with and without cetuximab. No significant difference was observed in the late toxicities between the two treatments. Conclusions: This propensity-matched study reveals that patients with T4 and/or N3 stage could benefit from the combination of cetuximab with the current chemoradiotherapy in locoregionally advanced NPC, although with more acute moderate to severe toxicities. However, this strategy remains to be validated in a prospective randomized controlled study. Clinical trial identification: This retrospective study has no clinical trial identification. Legal entity responsible for the study: Department of Radiation Oncology Nanjing Medical University Affiliated Cancer Hospital, Jiangsu Cancer Hospital, Jiangsu Institute of Cancer Research Funding: The National Natural Science Foundation of China (No. 81672989); Jiangsu Clinical Medicine Science and Technology Special Fund (BL2014091); Jiangsu Provincial Commission of Health and Family Planning Youth Research Project (Q201601). Disclosure: All authors have declared no conflicts of interest.

1073P

Safety and efficacy of nimotuzumab with concurrent chemoradiotherapy in unresectable locally advanced squamous cell carcinoma of head and neck: Indian rural hospital experience

S. Rawat1, D. Pawar2, H. Tandon1, S. Patel1, S. Chaudhari2 Medical Oncology, N.S.C.B Medical College and Government Hospital, Jabalpur, India, 2 Medical Affairs, Biocon Limited, Bangalore, India

1

Background: The aim of this study was to evaluate the safety and efficacy of nimotuzumab, a humanized monoclonal antibody against epidermal growth factor receptor, in combination with chemoradiation for head and neck squamous cell cancer (HNC).

384 | Head and neck cancer, excluding thyroid

Annals of Oncology Methods: The hospital data of 42 patients with HNC who were treated with nimotuzumab from January 2012 to December 2016 were evaluated. Three patients who had undergone prior surgery were excluded and 39 patients diagnosed with locally advanced (stage III- IVb) unresectable HNC who were treated with concurrent chemoradiotherapy with weekly nimotuzumab were considered for final analysis. Tumour response was calculated as per RECIST criteria 1.1. Subgroup analysis was performed to assess association of tumour response with independent variables such as age, gender, histopathological grades and TNM stages using chi square or Fischer exact test. Overall survival (OS) and progression free survival (PFS) was calculated from date of diagnosis using Kaplan-Meier method. All patients were assessed for toxicity and adverse events (AE) were reported as per common terminology criteria for AE v 4.0. Statistical analysis was done using SPSS software (v19.0). Results: At 24 weeks after completion of treatment, objective response rate (complete response [CR] þ partial response [PR]) was 97.44% with 26 (66.67%) patients showing CR, 12 (30.77%) patients with PR and one patient (2.56%) had stable disease. Subgroup analysis did not show significant association of tumour response, although men, patients older than 65 years, laryngeal cancer, tumour grade III, TNM stage III showed more complete responses. OS at one year and two years was 100% and 72.9%, while PFS at one year and two years was 87% and 54.40%, respectively. Incidence of grade I, II, III and IV toxicity was 30%, 18.18%, 41.82%, 10%, respectively. No grade V toxicity was observed. Common AE observed were neutropenia (20.91%), mucositis (33.64%), vomiting (18.18%), diarrhea (2.73%), skin reaction (24.55%). Nimotuzumab was observed to be safe with no additional adverse events (hypersensitivity, allergic reaction and skin changes) were reported during the study period. Conclusions: Nimotuzumab is an efficacious and safe option when added to concurrent chemoradiotherapy in patients with locally advanced Head and Neck cancer. Clinical trial identification: 125/12 Legal entity responsible for the study: Dr. Shyamji Rawat Funding: None Disclosure: D. Pawar: Works in a Pharmaceutical company. S. Chaudhari: works for Pharmaceutical company. All other authors have declared no conflicts of interest.

1074P

A phase II study of combination chemotherapy with cetuximab/S-1/ low dose cisplatin as neoadjuvant manner for oral squamous cell carcinoma patients

M. Nakazawa1, T. Imai1, Y. Matsumiya1, M. Suematsu2, Y. Ohta3, T. Ohnishi4, M. Ohmae5, H. Chisoku6, I. Kato1, S-Y. Tada2, A. Takeshita1 1 Oral & Maxillofacial Surgery 2, Osaka University Graduate School of Dentistry, Suita, Japan, 2Oral & Maxillofacial Surgery, Meiwa Hospital, Nishinomiya, Japan, 3Oral & Maxillofacial Surgery, Itami City Hospital, Itami, Japan, 4Oral & Maxillofacial Surgery, Ikeda Municipal Hospital, Ikeda, Japan, 5Oral & Maxillofacial Surgery, Rinku General Medical Center, Izumisano, Japan, 6Oral & Maxillofacial Surgery, Higashiosaka City Medical Center, Higashiosaka, Japan Background: In oral cancer therapy, functional preservation, as well as survival, is a very important matter to consider. One of the methods for organ preservation is an effective preoperative (neoadjuvant) chemotherapy. We previously reported the good antitumor effect and good tolerance of low dose cisplatin/S-1 at ESMO 2004 and 2006. Cetuximab enhances the antitumor effect of cisplatin/5-Fluorouracil . We investigated the feasibility of combining cetuximab/low dose cisplatin/S-1 chemotherapy as a neoadjuvant regimen for patients with oral squamous cell carcinoma. Methods: Consecutive patients (n ¼ 14) with newly diagnosed stage II-IV oral squamous cell carcinoma were enrolled in this study from July 2014 to June 2016. Patients were administered S-1 80mg/m2/day (day 1-14), cisplatin 5 mg/m2/day (day 1-5,8-12) and cetuximab 400mg/m2/day on day 1 and 250mg m2/day on day 8. This was followed by definitive surgery. Clinical response was assessed by clinical findings and/or CT according to RECIST and histopathological effects were evaluated with surgical specimens. Results: The rate of clinical response, including complete response (CR) and partial response (PR), was 85.7%: CR 21.8%, PR 64.3%, SD (stable disease) 14.3%. The rate of histological response was 71.4%: CR 21.4%, PR 50%, no change 28.6%. Toxicities above grade 3 were neutropenia (7.1%), hypokalaemia (7.1%), leukocytopenia (7.1%), thrombocytopenia (7.1%), anorexia (21.4%), diarrhoea (7.1%) and nausea (7.1%). Most toxicities disappeared within 8 weeks after chemotherapy. No serious adverse effects were observed in the majority of patients. Conservative surgery was applied to 12 patients, except 2 patients with SD and 5 of 9 patients who needed reconstruction were able to avoid reconstructive surgery. Conclusions: Combination chemotherapy with cetuximab/low dose cisplatin/S-1 represents an effective antitumor therapy with mild to moderate toxicities. It is suggested that this regimen is superior to low dose cisplatin/S-1 and can promote function preserving surgery. Clinical trial identification: UMIN000014632 Legal entity responsible for the study: Individual person Funding: None Disclosure: All authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 1075P

Role of induction chemotherapy in locally advanced T4b oral cavity cancers: A single Institute experience

K.N. Lokesh, T. Chaudhuri, K.C. Lakshmaiah, G. Babu, D. Lokanatha, L.A. Jacob, S. Babu, A. Rudresha, L. Rajeev Medical Oncology, Kidwai Cancer Institute, Bengaluru, India Background: The standard of care for oral squamous cell carcinoma (OSCC) at present, consist of surgical resection followed by adjuvant radiotherapy and chemotherapy as indicated. However, indications of induction chemotherapy (IC) in oral cancers are not clearly defined. This retrospective analysis aimed to investigate the efficacy, toxicity and impact of induction chemotherapy in locally advanced T4b oral cavity squamous cell cancers. Methods: Patients diagnosed with locally advanced T4b OSCC from January 2013 and March 2017 at our centre, who received 2-3 cycles of IC and then assessed for resectability, were reviewed retrospectively. Patients’ profile, response and toxicity of IC, resectability status and overall survival (OS) were evaluated. Statistical analyses were done by SPSS version 17.0. Results: Total 134 patients received IC, and out of them 98 (73.1%) were males. Median age at diagnosis was 44 years (range 31-60 years). 107 (79.8%) of our patients received doublet chemotherapy (with paclitaxel þ cisplatin), and the rest of the patients received triplet regimen (with paclitaxel/docetaxel þ cisplatin þ 5-FU). Majority of the patients had buccal mucosa cancers (n ¼ 92), followed by gingivo-buccal sulcus (n ¼ 26) and oral tongue (n ¼ 16) primaries. After IC, partial response was achieved in 25 (18.7%) patients, stable disease in 83 (61.9%) patients and disease progression was noted in 26 (19.4%) patients. Post-induction chemotherapy, resectability was achieved in 28 (21%) of 134 patients, but 8 of them did not undergo surgery due to logistic and personal reasons. The median OS of patients who underwent surgery followed by adjuvant local therapy (n ¼ 20) was 18.7 months (95% CI: 16.2-21.5 months) and for those treated with non-surgical local therapy (n ¼ 114) was 7.9 months (95% CI: 6.2-9.2 months) (log-rank p ¼ 0.000). Conclusions: IC may improve the resectability in our patients with T4b OSCC with a manageable toxicity profile. Patients underwent resection had a significantly better median OS than those who received non-surgical local treatment. Legal entity responsible for the study: Kidwai cancer institute Funding: None Disclosure: All authors have declared no conflicts of interest.

1076P

Concurrent chemoradiotherapy (CCRT) versus induction docetaxel, cisplatin and 5-fluorouracil (TPF) followed by CCRT in locally advanced hypopharyngeal and base of tongue cancer: A randomized phase II study

S.H. Lim1, J-M. Sun1, J. Hong1, D. Oh2, Y.C. Ahn2, M.K. Chung3, H-S. Jeong3, Y-I. Son3, MJ. Ahn1, C-H. Baek3, K. Park1 1 Division of Hematology-Oncology, Department of Medicine, Samsung Medical Center, Seoul, Republic of Korea, 2Department of Radiation Oncology, Samsung Medical Center, Seoul, Republic of Korea, 3Department of Otorhinolaryngology-Head and Neck Surgery, Samsung Medical Center, Seoul, Republic of Korea Background: To date, clinical trials have not consistently supported the use of induction chemotherapy (IC) for locally advanced head and neck squamous cell cancer (LASCC). Hypopharyngx and base of tongue (BOT) cancer has shown relatively poor survival compared to other LASCC. We tried to investigate the role of IC for improvement over current chemoradiotherapy (CRT) in patients with locally advanced hypopharynx and BOT cancer. Methods: Treatment-naı¨ve patients with nonmetastatic stage III/IV hypopharyngeal or BOT cancer were randomly assigned to receive CRT alone (CRT arm: cisplatin 100mg/ m2 3-weekly for 2 times plus radiotherapy 68.4Gy/30fraction on weekday) versus two 21-day cycles of IC (docetaxel 75mg/m2 on day 1, cisplatin 75mg/m2 on day 1, and fluorouracil 750mg/m2 on days 1 to 4) followed by same CRT regimen (IC arm). The primary endpoint was progression-free survival (PFS) and 90 patients are required to show the superiority of IC arm with one-sided alpha 0.1 and power of 0.85. Results: This study closed early after enrollment of 36 patients (19 in CRT arm and 17 in IC arm) because of slow accrual. After a median follow up of 47.2 months, there was no significant difference in PFS: the median PFS were 26.8 months for CRT arm and not reached for IC arm (Hazard ratio: 0.55, 95% CI 0.19-1.60). However, the survival curves widely separated with a plateau after 3-years, suggesting the survival benefit from induction chemotherapy: 3-year PFS rates were 45% and 68%, and 3-year overall survival rates were 56% and 86% (HR: 0.35, 95% CI: 0.07-1.69), in CRT and IC arms, respectively. In both subgroups with BOT and hypopharyngeal cancer, survival outcomes of IC arm were also insignificantly superior to those of CRT arm. All adverse events were manageable and there was no grade 3/4 toxicity except one patient had Gr3 stomatitis in IC arm. Conclusions: This study failed to demonstrate that induction TPF chemotherapy improves survival in patients with BOT and hypopharyngeal cancer, possibly due to small number of subjects. However, it suggested favorable outcome with induction chemotherapy, and further large randomized studies are needed to this population. Clinical trial identification: NCT01312350 Legal entity responsible for the study: Samsung medical center

Volume 28 | Supplement 5 | September 2017

Funding: None Disclosure: All authors have declared no conflicts of interest.

1077P

Do patients over 70 years with advanced head and neck squamous cell carcinoma tolerate curative intent concurrent chemo-radiation? Predictors of oncological outcomes

V. Srinivasalu1, N. Subramaniam2, N. Kumar3, D. Balasubramaniam2, A. Philip1, A. Susan1, A. Remesan Nair4, C. G Prameela4, K.U. Pushpaja4, K. Thankappan2, W. Jose1, S. Iyer2, P. Keechilat1 1 Department of Medical Oncology, Amrita Institute of Medical Sciences, Kochi, India, 2 Department of Head and Neck Oncology, Amrita Institute of Medical Sciences, Kochi, India, 3GCSRT Program, Harvard Medical School, Cambridge, UK, 4Department of Radiation Oncology, Amrita Institute of Medical Sciences, Kochi, India Background: Survival benefit of adding chemotherapy to radiotherapy (RT) in patients (pts) age>70 yrs of head and neck squamous cell carcinoma (HNSCC) has not been found in literature. Our institutional policy is to offer concurrent chemoradiation (CCRT) to patients > 70 years with a good ECOG status. Methods: Retrospective analysis of stage III/IV HNSCC in pts > 70 years who received linac based radical CCRT with dose equivalent to 70Gy in conventional fractionation (n ¼ 57) between 2006 to 2014 were included. Results: Pts with stage III/IV (25.6%/75.4%) HNSCC (n ¼ 57) of oropharynx (n ¼ 15), larynx (n ¼ 18) or hypopharynx (n ¼ 24) underwent radical CCRT having mean age 75.18 yrs (range 70-86 years) and male to female ratio of 10.4:1. Pts on CCRT who got cisplatin (CIS) (n ¼ 35) and carboplatin(CARBO) (n ¼ 22) had mean weight loss of 3.53 (range 0-10) kgs. 61.4% completed chemotherapy (defined as cumulative dose of 200mg/m2 of CIS and 5 weekly dose of CARBO at AUC 2) and 98.2% completed RT without any treatment related death. Higher grades of neutropenia (33.3%) and hyponatremia (17.5%) with CIS and hypercreatinemia (10.5%) with CARBO was noted. Tube dependence (gastrostomy/tracheostomy) had 2.7-fold increase in risk of death in pts (n ¼ 25; 44%) with hypopharynx/larynx cancer, compared with stage and subsite matched pair analysis in pts < 60 years. Factors predicting good PFS were ECOG (1 vs 2) HR ¼ 0.25 (95%CI:0.09-0.70), Completion of treatment without any breaks while on CCRT, HR ¼ 2.54 (95%CI:1.02-6.32, p ¼ 0.04), and age in 70-75 years, HR ¼ 1.09 (adjusted for alcohol and smoking) (95%CI: 1.01-1.20, p ¼ 0.08). Factors suggestive of poor PFS were hyponatremia, hypercreatinemia and weight loss > 3kgs from their baseline. PFS (80%) in pts with stage III and IV disease was 22 (95%CI: 12.4-87.2) months and 15.53 (95%CI: 8.6-20.6) months respectively. Conclusions: Curative intent CCRT should be considered as standard of care in elderly patients > 70 years with good ECOG status. Both cisplatin and carboplatin showed significant benefit in PFS with fewer side effects. Aggressive swallowing rehabilitation, abstinence from smoking and alcohol are likely to improve outcomes. Legal entity responsible for the study: Amrita Institute of Medical Sciences Funding: None Disclosure: All authors have declared no conflicts of interest.

1078P

Multidisciplinary team management in head and neck cancer: The real life experience

J-B. Guy1, Y. Xia1, A. Vallard1, S. Espenel1, J-C. Trone1, J. Langrand-Escure1, A. Hamrouni1, M. Ben Mrad1, C. Rancoule1, S. Ouni1, T. Muron2, P. Fournel2, N. Magne1 1 Radiation Oncology, Institut de Cance´rologie Lucien Neuwirth, Saint-etienne, France, 2 Medical Oncology, Institut de Cance´rologie Lucien Neuwirth, Saint-etienne, France Background: Multidisciplinary team (MDT) management in oncology is integrated in a legal framework in France. This practice is essential in Head and Neck cancer management with its complex and multimodal treatments. Some trials report a positive impact of MDT on overall survival of advanced head and neck cancers. The objective of this study was to report the experience of MDT management in Head and Neck Cancer in the Lucien Neuwirth Cancer Institute over the past 6 years. Methods: Records from bi-monthly MDT meeting from 2010 to 2015 were selected for this study. Number of medical cases and type of present medical specialists were noticed. Data from MDT records were reported: clinical characteristics (performans status, weight), anatomical localisation, TNM and pathological classification, and the treatment plan decision. Impact of MDT meeting on treatment delay was also analysed. Results: As of December 2015, 1848 clinical cases were discussed with 1786 patients and 138 MDT meetings. Majority of patients were discussed only once in meeting, and 3% (52) patients were discussed twice. An average of 16 patient’s cases were discussed per-meeting. 1368 patients (74.1%) were presented at primo-diagnosis status and 481 (25.9%) in a recurrence status. 81% of patients were at stage III or IV. 969 (52.4%) patients had a treatment before MDT. Surgery (73.2%) was the main treatment operated before meeting. Radiation therapy delay after MDT was 9.8 days for dosimetric planification CT and 21 days for first radiation treatment session. Conclusions: The percentage of presented recurrent patients is reasonable regarding epidemiologic data in head and neck location. MDT seems not delay radiation treatment occurring within 21 days after MDT. Unfortunately we underlighted a majority of patients surgically treated before MDT discussion.

doi:10.1093/annonc/mdx374 | 385

abstracts Legal entity responsible for the study: Nicolas Magne´ Funding: None Disclosure: All authors have declared no conflicts of interest.

Annals of Oncology AstraZeneca. R. Galiulin, M. Tahara, K. Hoermann: COI to follow All other authors have declared no conflicts of interest. 1081P

1079P

Long-term response to second-line afatinib in patients with recurrent or metastatic head and neck squamous cell carcinoma (R/M HNSCC): Analysis of the LUX-Head & Neck 1 (LHN1) trial

J-P. Machiels1, G. de Castro Jr2, L. de Souza Viana3, R. Galiulin4, M. Tahara5, U.R. Nicolau6, C. Le Tourneau7, K. Okami8, V. Vladimirov9, A. Izmailov10, K. Hoermann11, L. Licitra12, R. Haddad13, E. Cohen14, N. Dupuis15, J. Love16, E. Zografos17, E. Ehrnrooth18, J. Fayette19 1 Institut Roi Albert II, Service d’Oncologie Me´dicale, Cliniques Universitaires Saint-Luc and Institut de Recherche Clinique et Expe´rimentale (Pole MIRO), Universite´ Catholique ancer de Louvain, Brussels, Belgium, 2Department of Clinical Oncology, Instituto do C^ ancer de do Estado, S~ao Paulo, Brazil, 3Department of Medical Oncology, Hospital de C^ Barretos, S~ao Paulo, Brazil, 4Chemotherapy Department, Omsk Regional Oncology Center, Omsk, Russian Federation, 5Division of Head and Neck Medical Oncology, National Cancer Center Hospital East, Kashiwa, Japan, 6Departamento de Oncologia, Hospital A.C. Camargo, S~ ao Paulo, Brazil, 7Department of Medical Oncology, Institut Curie, Paris, France, 8Department of Otolaryngology, Tokai University, Isehara, Japan, 9 Policlinic, Pyatigorsk Oncological Dispensary, Pyatigorsk, Russian Federation, 10 Department of Oncology, Bashkirin State Medical University, Ufa, Russian Federation, 11 Department of Otolaryngology, Head and Neck Surgery, University Hospital Mannheim, Mannheim, Germany, 12Fondazione IRCCS Istituto Nazionale Tumori, Milan, IT, and University of Milan, Milan, Italy, 13Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School and Department of Medicine, Brigham and Women’s Hospital, Boston, MA, USA, 14San Diego Moores Cancer Center, University of California, La Jolla, CA, USA, 15Development, Biodesix Inc., Boulder, CO, USA, 16Oncology Statistics, Boehringer Ingelheim Corporation, Ridgeﬁeld, CT, USA, 17 Boehringer Ingelheim, Bracknell, UK, 18TA Oncology, Boehringer Ingelheim, Danmark A/S, Copenhagen, Denmark, 19Le´on Be´rard Center and Hospices Civils de Lyon, University of Lyon, Lyon, France Background: In the Phase III LHN1 trial, second-line afatinib (A) significantly improved PFS (primary endpoint) vs methotrexate (MTX) in pts with R/M HNSCC. Tumour biomarker analyses have shown that survival benefit with A vs MTX was more pronounced in pts with p16/ErbB3-negative, EGFR-amplified, PTEN-positive disease. We present post-hoc analyses of A long-term responders (LTRs). Methods: Pts with incurable R/M HNSCC who had received first-line platinum-based therapy were randomised to A (40mg/day) or MTX (40mg/m2/week) and treated until progression/intolerable AE. LTRs were defined as pts treated with A 12 mos. Tumour biomarkers were assessed by IHC (p16, ErbB3, PTEN, cMET) and FISH (EGFR amplifiR (VS) test cation); pre-treatment (tx) serum samples were analysed with the VeriStratV and classified as VS-Good/Poor. Results: 11/322 (3%) pts treated with A were LTRs with a median (range) tx-duration of 16 (1239) mos. All pts had stopped tx at analysis. Baseline characteristics in LTRs were similar to the overall dataset, except (LTRs/overall): oral cavity primary tumour site (45%/29%); M1 disease (45%/66%); previous therapy with EGFR-antibodies (18%/59%). Median OS was 18.1 mos; median PFS (central independent review) was 14.9 mos. ORR was 45% (CR: 18%; n ¼ 2). The frequency of pts who received 1 subsequent therapy was similar to the overall dataset (LTRs, 45%; overall, 51%). In LTRs with available biomarker data, 3/3 (100%) pts were p16-negative, 4/4 (100%) pts were ErbB3-negative, 2/4 (50%) pts were PTEN-positive, 3/3 (100%) pts were cMETpositive, 2/3 (67%) pts had EGFR-amplification, and 5/5 (100%) pts were VS-Good. Tolerability-guided dose reductions were more frequent among LTRs (55% vs 32% overall). Conclusions: In the LHN1 study, some platinum-pre-treated pts with R/M HNSCC derived a long-term survival benefit from A; median OS was 1.5 yrs and >11 mos longer than in the overall dataset. Limited biomarker data available in these LTRs suggests that p16/ErbB3-negativity and EGFR-amplification might be potential predictive biomarkers for long-term benefit from A; however, results were not conclusive due to small sample size. Clinical trial identification: NCT01345682 Legal entity responsible for the study: Boehringer Ingelheim Funding: Boehringer Ingelheim Disclosure: J-P. Machiels: Reports taking part in advisory boards for Debio, Innate, Nanobiotix, MSD, AstraZeneca, and grants from Janssen, Bayer, and Novartis outside the submitted work. K. Okami: Honoraria: Merck Serono Co., Ltd C. Le Tourneau: Advisory board: MSD, Amgen, BMS. Honoraria: MSD, BMS, Merck Serono, Novartis. E. Zografos: Employment Boehringer Ingelheim L. de Souza Viana: Research funding from Boehringer Ingelheim. L. Licitra: Participated in advisory boards for Eisai, Bristol-Myers Squibb, GlaxoSmithKline, Lilly, Merck Serono, Amgen, Boehringer Ingelheim, Debiopharm, VentiRX, SOBI, Novartis, AstraZeneca and Bayer. E. Cohen: Participated in a consulting or advisory role for AstraZeneca, Bayer, Bristol-Myers Squibb, Eisai, Merck, and Novartis, and has participated in speakers’ bureaus for Bayer and Eisai. N. Dupuis: Owns stock in Biodesix. J. Love, E. Ehrnrooth: Employee of Boehringer Ingelheim. J. Fayette: Received honoraria from Bristol-Myers Squibb and

386 | Head and neck cancer, excluding thyroid

Benefit of cetuximab addition to a platinum-fluorouracil-based chemotherapy in an unselected population of metastatic head and neck cancer patients and effect of KRAS Lcs6 variation on cetuximab response

V. Bastit1, N. Bon-Mardion1, J-M. Picquenot2, V. Rainville3, C. Moldovan4, A. Franc¸ois5, F. El Ouakif6, F. Jardin3, J-P. Marie1, F. Di Fiore4, F. Clatot4 1 ORL and Head and Neck Surgery, CHU de Rouen - Hoˆpital Charle Nicolle, Rouen, France, 2Biopathology, Henri Becquerel Center, Rouen, France, 3INSERM U1245, Henri Becquerel Center, Rouen, France, 4Department of Medical Oncology, Henri Becquerel Center, Rouen, France, 5Biopathology, CHU de Rouen - Hoˆpital Charle Nicolle, Rouen, France, 6Head and Neck Surgery, Henri Becquerel Center, Rouen, France Background: While the EXTREME protocol, including platinum (P) fluorouracil (FU) and cetuximab (Cx), is the gold-standard first line chemotherapy for metastatic head neck cancer patients (MHNC), its benefit in an unselected population has never been evaluated. Furthermore, KRAS Lcs6 variation was reported as a potential marker for greater efficacy of EGFR-targeted therapy. We investigated the benefit on progressionfree survival (PFS) and overall survival (OS) of adding Cx to PFU as first line treatment for MHNC in an unselected population. We also assessed if there was a differential efficacy of Cx according to KRAS Lcs6 status. Methods: This monocentric retrospective study included all the patients treated by at least two cycles of PFUþ/-Cx between 2005 and 2014 as first line of palliative chemotherapy for MHNC. When tumor samples were available, the KRAS LCS6 variant status (rs61764370) was determined by pyrosequencing, and the p16 status by immunohisto-chemistry. Results: 134 patients were included: 59 (44%) treated with PFU and 75 (56%) with PFUCx. Baseline characteristics were comparable between the two groups. Of note 30% of the patients had a stage 2 or 3 performance status (PS). In univariate analysis, a longer median PFS was observed with PFUCx compared to PFU (6.1 vs 4.4 months respectively, HR 0.68, p ¼ 0.02). Median OS were not different (11.1 months with PFUCx versus 9.1 with PFU, p ¼ 0.2). Among the 110 tumor samples available, 29 (25%) had a KRAS-variant and 14 (12.7%) were p16 positive. No differences in OS nor PFS were observed according to the KRAS-variant status. When considering only the patients treated with PFUCx, presence of the KRAS-variant (n ¼ 17) was not associated with a better response (p ¼ 0.5). In a multivariate analysis including PS 1, addition of Cx, KRAS status, p16 status and age 55 as variables; addition of Cx to PFU was the only factor related to a better PFS (p ¼ 0.008). Conclusions: This retrospective study confirmed the effectiveness of the EXTREME protocol on PFS in an unselected population of MHNC patients. KRAS LCS6 variant was not related to a differential response to Cetuximab in MHNC population. Legal entity responsible for the study: Centre Henri Becquerel Funding: IRON - Centre Henri Becquerel Disclosure: All authors have declared no conflicts of interest.

1082P

A pilot study of apatinib in heavily pretreated metastatic adenocarcinoma of the head and neck

L. Gui, X. He Medical Oncology, Cancer Institute and Hospital, Chinese Academy of Medical Sciences (CAMS), Beijing, China Background: Although antiangiogenic therapy is effiective in advanced lung, breast, renal, hepatic, and colon cancers, limited is known about its value in the cancer of the Head and Neck. Apatinib is an oral, highly potent tyrosine kinase inhibitor targeting vascular endothelial growth factor receptor 2 (VEGFR-2). This prospective phase II study (NCT02989259) aims to investigate the efficacy and safety of apatinib in heavily pretreated patients (pts) with metastatic adenocarcinoma of the Head and Neck. Methods: This study enrolled pts with metastatic adenocarcinoma of the Head and Neck, who failed in the metastatic setting at least one prior chemotherapy regimen. The primary end point of this study was progression free survival (PFS). Secondary end points included objective response rate (ORR), disease control rate (DCR), overall survival (OS), and safety. Patients were treated with apatinib 500 mg daily. Efficacy was assessed every 6 weeks. Results: From December 2016, we recruited 10 pts, including 8 males and 2 females, with a median age of 53 years (26-71). Median number of previous chemotherapy regimens for the metastatic diseases was 2 (1-3). Median follow-up time was 4.3 months. 8 pts were eligible for efficacy analysis. ORR was 25% (2/8). DCR was 87.5% (7/8). Median PFS and median OS were not reached. The most common adverse events (AEs) of all grade were hypertension (n ¼ 5), nausea (n ¼ 4), fatigue (n ¼ 4) and hand-foot syndrome (n ¼ 3). The most common grade 3/4 AEs were hypertension (n ¼ 2), thrombocytopenia (n ¼ 1) and oral mucositis (n ¼ 1). Toxicities were tolerable and manageable. Conclusions: Our results so far indicated that apatinib exhibited objective efficacy in heavily pretreated, metastatic adenocarcinoma of the Head and Neck with acceptable safety.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Clinical trial identification: the trial protocol number: NCT02989259 release date: December 2016 Legal entity responsible for the study: The Institutional Review Board of the Cancer Institute/Hospital, CAMS & PUMC Institutional Review Board of the Cancer Institute/ Hospital, CAMS & PUMC Funding: None Disclosure: All authors have declared no conflicts of interest.

1083P

Natural history and prognostic factors of head and neck cancer patients with bone metastases: A retrospective Italian study

S. Grisanti1, C. Bergamini2, S. Bianchi1, A. Baiguini3, S. Vecchio4, L.D. Locati2, A. Bonetta5, P.F. Conte6, M. Airoldi7, M. Merlano8, P. Carlini9, T. Ibrahim10, C. Rossetto11, P. Nicolai12, R. Maroldi13, S. Tonoli5, P. Pronzato4, S.M. Magrini3, L. Licitra2, A. Berruti1 1 Medical Oncology, Azienda Ospedaliera Spedali Civili di Brescia and University of Brescia, Brescia, Italy, 2Medical Oncology Unit/Head and Neck Unit, Fondazione IRCCS Istituto Nazionale dei Tumori and University of Milan, Milan, Italy, 3Radiation Oncology, Azienda Ospedaliera Spedali Civili di Brescia and University of Brescia, Brescia, Italy, 4 Medical Oncology, IRCCS San Martino, IST National Cancer Institute and University of Genova, Genoa, Italy, 5Radiation Oncology, Cremona Hospital, Cremona, Italy, 6 Department of Surgery, Oncology and Gastroenterology, University of Padova, Istituto Oncologico Veneto IRCCS, Padua, Italy, 7Medical Oncology, AOU S. Giovanni Battista Molinette, Turin, Italy, 8Oncology Department, S. Croce & Carle Teaching Hospital, Cuneo, Italy, 9Medical Oncology, Regina Elena National Cancer Institute, Rome, Italy, 10 Osteoncology and Rare Tumors Center, IRCCS Istituto Scientiﬁco Romagnolo per lo Studio e la Cura dei Tumori (IRST), Meldola, Italy, 11Oncology, University Hospital Santa Maria della Misericordia, Udine, Italy, 12Otorhinolaryngology Unit, Azienda Ospedaliera Spedali Civili di Brescia and University of Brescia, Brescia, Italy, 13Radiology Unit, Azienda Ospedaliera Spedali Civili di Brescia and University of Brescia, Brescia, Italy Background: We performed a retrospective survey to study the natural history and the prognostic factors of patients (pts) with bone metastases (BMs) from head and neck cancers (HNCs). Methods: Clinical records of pts treated at 11 oncologic centers across Italy were retrieved. All pts were selected on the base of BMs either at diagnosis or at subsequent time points. Pts with a diagnosis of nasopharyngeal carcinoma (NPC) were analyzed separately. The time-to-first bone metastasis was calculated from initial diagnosis. The skeletal-related events (SREs) (fractures, medullary compression, hypercalcemia) were recorded from the date of BMs diagnosis. Results: From 2008 to 2016, 192 HNC pts with BMs (64 NPCs and 128 other-HNCs) were collected. Median time to first BM was 9 and 12 months for NPCs and otherHNCs, respectively. SREs occurred in 9% and 27% NPC and other-HNC pts, respectively. Pts received specific antineoplastic treatments (92%) or best supportive care (8%). Median progression free survival (PFS) and overall survival (OS) were 11 and 25 months in NPC-BM pts and only 5 and 6 months in other-HNC-BM pts, respectively. SRE did not affect the pt prognosis except for hypercalcemia that was associated with a poorer prognosis (not significant). Biphosphonates and/or denosumab were administered in 34% NPC and 33% other-HNC pts, respectively. The administration of bonedirected therapies, also including radiation therapy and surgery on BMs, was associated with a better survival at univariate analysis in both NPC and other-HNC (Hazard Ratio [HR]: 0.43, 95% confidence interval [CI] .008-.237, p 5). The primary end point was overall survival (OS). Progression free survival (PFS) was the secondary endpoint. Univariate analysis was used to identify associations and to select variables included in multivariate Cox regression analysis to determine prognostic value. Results: 146 patients (132 squamous cell carcinomas (SCC), 10 undifferencied nasopharyngeal carcinomas (UCNT) and 4 neuroendocrine carcinomas) were included in this analysis. The median follow up was 20.6 months (2.4-37.0 months). 1-year and 2year OS were 87.1% and 82.3%, respectively.1-year and 2-year PFS were 75.9% and 68.0%, respectively. On univariate analysis, OS significantly differed between groups NLR 5 vs. > 5. In the overall population (OP) (HR: 2.6; IC95%: [1,05-6,53]; p ¼ 0036) and in non-oropharyngeal subpopulation (HR: 3.67; IC95%: [1,19-11,4]; p ¼ 0,016) but not in the oropharyngeal subpopulation (p ¼ 0,51). In multivariate analysis NLR >5 was significantly associated with a poorer OS in the OP (HR: 2.89; IC95%: [1,14-7,33]; p ¼ 0.025) and in non-oropharyngeal subpopulation (HR: 4.53; IC95%: [1,34-13,5]; p ¼ 0.014). Body Mass Index (BMI) 5 was predictive of shorter overall survival. Further prospective clinical investigations are required to confirm these results and determine the clinical applicability as prognostic factor. Legal entity responsible for the study: Centre Oscar Lambret Funding: None Disclosure: All authors have declared no conflicts of interest.

1085P

Association of ERP29 genetic polymorphism in microRNA-binding site with oropharynx cancer risk and prognosis

J. Carron1, E. Costa1, L. Lopes-Aguiar1, B. Sa Carvalho2, J.A. Rinck-Junior3, A.P. Dalla Costa1, M. Ortega4, C.S.P. Lima3, G.J. Lourenc¸o1 1 Laboratory of Cancer Genetics, University of Campinas, Campinas, Brazil, 2Department of Statistics, Institute of Mathematics, Statistic and Computer Science, University of Campinas, Campinas, Brazil, 3Internal Medicine, University of Campinas, Campinas, ao Brazil, 4Laboratory of Cell and Molecular Tumor Biology and Bioactive Compounds, S~ Francisco University, Braganc¸a Paulista, Brazil Background: Our previous large-scale genotyping study identified more than 6,000 single nucleotide polymorphisms (SNPs) associated with base of tongue (BT) squamous cell carcinoma (SCC) risk in 49 patients and 49 controls. The SNP c.*293A>G of the ERP29, a tumor-suppressor chaperone, was selected for further analyses. An in silico analysis showed that microRNA (miR)-4421 shared binding site with 3’-untranslated region of variant allele of referred SNP while wild-type allele disrupt this target site. However, the role of this SNP in the risk and prognosis of oropharynx SCC (OPSCC) patients is still unknown. We aimed to verify whether the distinct genotypes or ERP29 c.*293A>G SNP influence the OPSCC risk and prognosis, and the ERP29 and miR-4421 expressions. Methods: DNA from 250 OPSCC patients and 250 controls was analyzed by RT-PCR. The patients were treated with surgery, radiotherapy and/or platinum based agents. The ERP29 and miR-4421 levels were evaluated by qPCR using RNA of 58 controls. The differences between groups were calculated by chi-square, logistic regression model and Mann-Whitney tests. Progression-free survival (PFS) and overall survival (OS) times were calculated by Kaplan-Meier and Cox regression methods. Results: ERP29 variant GG genotype was more common in OPSCC patients than in controls (6.4% vs. 3.6%, P ¼ 0.002). Individuals with GG genotype were under 8.86fold increased risk of OPSCC than others (95% CI: 2.20-35.69). Considering only the BTSCC patients, at 36 months of follow-up, shorter PFS were seen in patients with variant GG genotype (0.0% vs. 39.1%, P ¼ 0.01, Cox: HR: 2.68, 95% CI: 1.21-5.95, P ¼ 0.01). Individuals with ERP29 wild-type genotype showed higher levels of ERP29 mRNA when compared to others (1.44 vs. 1.10 arbitrary units AUs, P ¼ 0.005). In addition, lower miR-4421 expression was observed in individuals with ERP29 AA genotype when compared to those with AG or GG genotypes (0.42 vs. 0.73 AUs, P ¼ 0.05). Conclusions: Our data present, for the first time, that ERP29 c.*293A>G SNP is associated with increased risk of OPSCC and with worst survival of BTSCC patients, possibly due to variation of ERP29 mRNA levels modulated by miR-4421. Legal entity responsible for the study: University of Campinas Funding: S~ao Paulo Research Foundation (FAPESP) and Coordination for the Improvement of Higher Education Personnel (CAPES) Disclosure: All authors have declared no conflicts of interest.

Background: The neutrophil-to-lymphocyte ratio (NLR), a marker of the systemic inflammatory response, has been reported to have prognostic value in different cancer settings. In this study we aimed to assess the prognostic impact of NLR in a cohort of patients with head and neck cancers.

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx374 | 387

abstracts 1086P

Diagnostic and prognostic impact of plasma osteopontin in nasopharyngeal carcinoma

J-C. Lin1, W-Y. Wang2, Y.C. Liu1 Radiotion Oncology, Taichung Veterans General Hospital, Taichung, Taiwan, 2 Department of Nursing, Hung Kuang University, Taichung, Taiwan

1

Background: We investigated the diagnostic and prognostic impact of plasma osteopontin (pOPN) concentrations in advanced nasopharyngeal carcinoma (NPC). Methods: Pre-treatment plasma samples from 138 patients with previously untreated and biopsy-proven NPC were collected. Plasma samples from another 70 healthy volunteer were served as control. OPN concentrations were measured by the enzymelinked immunosorbent assay (ELISA). The patient characteristics were- age range 2483 and median 48 years, male/female¼97/41, WHO pathology type I/II/III¼1/105/32, stage III/IV(M0)/IVC(M1) ¼ 57/73/8. The treatment consisted of radiotherapy alone (2), concurrent chemoradiotherapy (28), and neoadjuvant chemotherapy plus radiotherapy (100) for M0 patients, and systemic chemotherapy with or without radiotherapy for M1 patients (8). Results: NPC patients (median 97.2 ng/mL; interquartile range 72.1-130.4) had significantly higher pOPN level than normal control (median 61.6 copies/ml; interquartile range 44.9-88.1), P100 ng/mL) correlated with some clinically poor prognostic factors, such as older age, male gender, advanced T-stage, and advanced overall stage. Pretreatment pOPN affected patients’ survival as well as rates of distant failure. The 5-year overall survival (56.6% vs. 81.4%, P¼0.0036) and metastasis-free survival (66.3% vs. 81.2%, P¼0.0726) were significantly lower in patients with pretreatment pOPN > 100 ng/mL than in those with pOPN < 100 ng/mL. Conclusions: Pretreatment pOPN levels can serve as a useful diagnostic and prognostic marker for advanced NPC. Legal entity responsible for the study: Taichung Veteran General Hospital Funding: Taichung Veteran General Hospital Disclosure: All authors have declared no conflicts of interest.

1087P

Head and neck cancer (HNC) and synchronic lung cancer: Impact of the lung cancer on the management and prognosis of these patients. Data from the SYNCHRON GFPC 15-04 Study

N. Paleiron1, L. Saramon2, G. Robinet3, R. Gervais4, P. Fournel5, H. Le Caer6, H. Berard7, G. Valette2, R. Marianowski2, C. Chouaid8 1 Pneumology, HIA Brest, Brest, France, 2ORL, CHU Brest Morvan, Brest, France, 3 Pneumologie, CHU Brest Morvan, Brest, France, 4Oncologie, Centre Francois Baclesse, Caen, France, 5Pneumologie, Institute de Cancerologie de la Loire, St. Priest En Jarez, France, 6Pneumologie, Hoˆpital de St Brieuc, St Brieuc, France, 7Pneumologie, Hoˆpital d’Instruction des Arme´es (HIA) Ste Anne, Toulon, France, 8Pneumologie, CHI cre´teil, Creteil, France Background: Management of synchronous head and neck and lung cancer is almost difficult. The aim of this observational study was to describe the impact of the lung cancer on the management and prognosis of HNC. Methods: Inclusion criteria were: consecutive patients diagnosed between January 2011 and December 2015 in 19 French centers with HNC and synchronous lung cancer (all stages). We describe: clinical characteristics, management and outcomes. Patient characteristics and treatment information was analyzed descriptively. Kaplan-Meier estimation was used to assess median overall survival. Results: The study included 132 patients: men: 83%; 63,7 years old, current smokers: 59,8%; performans status: 0 and 1 for 22% and 66% of the patients respectively; high rate of comorbidities: cardiovascular: 63%, COPD: 33%. Main histology for HNC was squamous: 98%, (in oral cavity: 24%, oropharyngeal: 26%, hypo-pharyngeal: 22% and laryngeal: 28%) T classification was T1, T2, T3 and T4 in 16%, 24%, 28% and 18% of cases respectively, and N classification was N0, N1, N2, N3, for 36%, 18%, 20% and 8% of cases respectively. The main treatment was surgery, 37,1%, and chemo-radiotherapy, 35,6%. The diagnosis of lung cancer impacts the HNC management in 38% of the cases. Median delay between HNC and first day treatment was 54 days. HNC progressive free survival rate was 68% at 2 years. Lung cancers were mostly localized (stages I: 46%, stages II: 10%), squamous: 39%, or adenocarcinomas: 39%. Main treatments were surgery: 29%, mainly lobectomy, radiotherapy: 13%, radio-chemotherapy: 14% and chemotherapy alone: 35%. Seven patients didn’t receive active treatment. Median delay of treatment was 82,3 days. Lung cancer progressive free survival rate was 35% at 2 years. OS was 40% at 2 years, better for stage I - II lung cancers (55%). Conclusions: Synchronous lung cancer at HNC diagnosis significantly impacts the management and outcomes of HNC. Specific recommendations and multidisciplinary approach should be elaborate to improve the management of these patients. Legal entity responsible for the study: Groupe Franc¸ais de Pneumo-cance´rologie Funding: Boringer Pierre Fabre Disclosure: All authors have declared no conflicts of interest.

388 | Head and neck cancer, excluding thyroid

Annals of Oncology 1088P

An estimation of the population survival benefit of first-course chemotherapy for head and neck cancers

V. Do, W. Ng, S. Jacob, G.P. Delaney, M. Barton Collaboration for Cancer Outcomes Research and Evaluation (CCORE), Ingham Institute for Applied Medical Research, University of New South Wales, Sydney, Liverpool, Australia Background: Randomised clinical trials describe the benefit of chemotherapy for specific head and neck patients with selected patient and tumour characteristics. This study estimates the overall survival benefit of chemotherapy above all other modalities for the whole population of head and neck cancer patients in Australia, if evidence-guidelines were followed. Methods: Decision trees with evidence-based indications for chemotherapy have been previously defined. For all defined indications, the highest level of clinical evidence available was identified. Multiple electronic citation databases were systematically queried, including Medline and Cochrane library. The benefits of first-course chemotherapy were estimated for 1-year and 5-year overall survivals. To assess the robustness of our estimates, univariate and multivariate analyses were performed. Results: The estimated 1-year and 5-year absolute population-based survival benefits of optimally utilised chemotherapy for head and neck cancer patients in Australia are 5.5% (95% Confidence Interval, CI, 4.5%-6.8%) and 4.2% (95% CI, 3.6%-5.0%), respectively. Conclusions: First-course chemotherapy improved population-based survival in head and neck cancer patients, when used in accordence with guidelines recommendations. Measurement of population survival benefits of cancer treatment is important as these can provide salient inputs for economic analyses, aid in priority setting in cancer program and guide quality improvement according to evidence-based guidelines. Legal entity responsible for the study: CCORE, Ingham Institute for Applied Medical Research, Sydney, Australia. Funding: None Disclosure: All authors have declared no conflicts of interest.

1089P

GSTP1 c.313A>G, XPD c.934G>A, XPF c.2505T>C and CASP9 c.1339A>G polymorphisms and severity of vomiting in head and neck cancer patients treated with cisplatin chemoradiation

C.S.P. Lima, J. Carron, L. Lopes-Aguiar, E.F.D. Costa, G.A.S. Nogueira, T.R.P. Lima, E.C. Pincinato, M.B. Visacri, J.C.F. Quintanilha, P. Moriel, G.J. Lourenc¸oDepartment of Internal Medicine, University of Campinas, Campinas, Brazil

Background: Cisplatin (CDDP) chemotherapy associated with radiation (RT) has been used in head and neck squamous cell carcinoma (HNSCC) patients, and vomiting is a common side effect during treatment. This prospective study aimed to identify the roles of GSTM1 and GSTT1 (presents or nulls), GSTP1 c.313A>G, XPC c.2815A>C, XPD c.934G>A and c.2251A>C, XPF c.2505T>C, ERCC1 c.354C>T, MLH1 c.93G>A, MSH2 c.211 þ 9C>G, MSH3 c.3133G>A, EXO1 c.1765G>A, TP53 c.215G>C, CASP3 c.-1191A>G and c.-1168G>T, CASP9 c.-1339A>G, CASP8 c.937_-932delAGTAAG, FAS c.-1378G>A and c.-671A>G, and FASL c.-157-687C>T single nucleotide polymorphisms, involved in CDDP metabolism, in vomiting severity in HNSCC patients treated with CDDP and RT. Methods: We evaluated 88 HNSCC patients diagnosed June 2011-February 2014 which receive CDDP chemoradiation. Ondansetron and dexamethasone were administered as antiemetic therapy and evaluated using National Cancer Institute criteria. Genotypes were analyzed in genomic DNA by polymerase chain reaction based methods. The logistic regression model was used to identify variables influencing toxicities and significant results were validated using a bootstrap (bt) resampling to investigate the stability of risk estimates (1000 replications). Results: GSTP1 c.313AG or GG genotype alone (46.7% vs 18.6%, P ¼ 0.004) and combined with XPD c.934GA or AA (50.0% vs 16.7%, P ¼ 0.02; Pbt¼ 0.008), XPF c.2505TC or CC (52.2% vs 16.7%, P ¼ 0.02; Pbt¼ 0.007) and CASP9 c.-1339AG or GG (51.9% vs 16.7%, P ¼ 0.02; Pbt¼ 0.01) genotypes were more common in patients with moderate/ severe vomiting than other genotypes. Carries with GSTP1 c.313AG or GG genotypes alone and combined with XPD c.934GA or AA, XPF c.2505TC or CC and CASP9 c.1339AG or GG genotypes had 4.28, 5.00, 5.45, and 5.38 more chances of presenting moderate/severe vomiting than others. Conclusions: Our data suggest, for the first time, that inherited abnormalities in DNA repair and apoptosis pathways are capable of modulating emesis in HNSCC patients under CDDP chemoradiation, and may be used for selecting patients who deserve to receive distinct doses of antiemetics or association of potent antiemetics in clinical practice. Legal entity responsible for the study: Fundac¸~ao de Amparo a Pesquisa do Estado de S~ao Paulo (FAPESP) Funding: None Disclosure: All authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 1090P

Analysis of the impact of the tumours committee on the multidisciplinary approach to head and neck epidermoid cancer in our institution

M.J. Martınez-Ortiz1, P. Cerezuela1, G. Marin Zafra2, J. Balsalobre1, J. Dıaz-Manzano3, J. Trujillo4 1 Medical Oncology, Hospital Universitario Santa Lucia, Cartagena, Spain, 2Hematology and Medical Oncology, Hospital Universitario Morales Meseguer, Murcia, Spain, 3ORL, HCU Virgen de la Arrixaca, El Palmar, Spain, 4Internal Medicine, Hospital Universitario Santa Lucia, Cartagena, Spain Background: The objectives are to define the characteristics of people with HNSCC treated in our area and quantify the impact of the committee on the staging, the change in the initial treatment proposed to determine whether the selection of treatment by the multidisciplinary team (MDT) influences therapeutic compliance. Methods: Observational retrospective study of two cohorts, which aims to analyse the variables in the cohort of patients handled by an MDT with respect to the patients without an MDT. We included all patients with an initial diagnosis of HNSCC at our centre between 2005 and 2012. The MDT cohort comprised those from 01/01/2009 to 31/12/ 2012. With access to the Pathological Anatomy database, the records of the MDT, the archived and computerised medical history, we collected the endpoints related to the patient (age, sex, ECOG), the tumour (date of diagnosis, location, and TNM stage), the treatment (therapy selected, change in treatment, compliance, reason for default). Definitive sample consists of 408 patients. A descriptive analysis is given of the clinical characteristics of the sample, together with the comparative bivariate analysis of these characteristics in the cohorts. Results: Our population presents age (mean) 64.2y (SD 12.4), male 82.6%, ECOGG and CASP3 c.-1191A>G polymorphisms in susceptibility and outcome of head and neck squamous cell carcinoma

E.F.D. Costa, L. Lopes-Aguiar, G.A.S. Nogueira, T.R.P. Lima, V.T. Liutti, F. Leal, V.C.A. Santos, J.A. Rinck-Junior, G.J. Lourenc¸o, C.S.P. Lima Department of Internal Medicine, University of Campinas, Campinas, Brazil Background: We analyzed herein the roles of CASP9 c.-1339A>G (rs4645978) and CASP3 c.-1191A>G (rs12108497) single nucleotide polymorphisms (SNPs) of intrinsic apoptosis pathway on risk and behavior of head and neck squamous cell carcinoma (HNSCC). Methods: DNA of 350 HNSCC patients and 350 controls was analyzed by polymerase chain reaction method and enzymatic digestion for genotyping. Patients were treated according to the Institutional protocol, including surgery, radio and chemotherapy. The statistical analyses were realized using chi-square, logistic regression model, multifactor dimensionality reduction (MDR), Kaplan-Meier, and univariate and multivariate Cox analyses. Results: CASP3 c.-1191AG or GG genotype was more common in patients with overall HNSCC (63.4% versus 53.4%, P¼ 0.013), male patients with overall HNSCC (65.5% versus 53.4%, P¼ 0.011), patients with SCC of oral cavity (OCSCC) (68.0% versus 53.4%, P¼ 0.02) and SCC of pharynx (PSCC) (62.7% versus 53.4%, P ¼ 0.010) than in controls; carriers of genotypes were under 2.15 and 2.34-fold increased risks of overall HNSCC, 2.75 and 2.67-fold increased risks of OCSCC and PSCC, respectively. Interactions of CASP9 and CASP3 SNPs and tobacco on HNSCC, OCSCC, PSCC, and laryngeal SCC risks were evident in study (P< 0.01). At 60 months of follow-up, eventfree survival was worst in patients with CASP9 c.-1339GG genotype (35.9% versus 45.1%, P ¼ 0.04) compared to others (Kaplan-Meier estimates). Patients with CASP9 c.-1339GG genotype and CASP9 c.-1339GG plus CASP3 c.-1191GG genotypes had 1.46 more chances of disease progression or relapse and 2.66 more chances of evolving to death in univariate and multivariate analyses, respectively. Conclusions: We present, for the first time, preliminary evidence that inherited abnormalities in the intrinsic apoptosis pathway, related to CASP9 c.-1339A>G and CASP3 c.1191A>G SNPs, are important determinants of HNSCC risk and outcome. Financial support: Fundac¸~ao de Amparo a Pesquisa do Estado de S~ao Paulo (FAPESP). Legal entity responsible for the study: Fundac¸~ao de Amparo a Pesquisa do Estado de S~ao Paulo (FAPESP) Funding: None Disclosure: All authors have declared no conflicts of interest.

1093P

Resource-stratification of national comprehensive cancer network R ) head and neck cancers guideline (NCCNV

R.W. Carlson1, E. Nardi2, J. Bacigalupo3, S. Darlow4, J.S. McClure4, D. Pfister5 Management Team, National Comprehensive Cancer Network (NCCN), Fort Washington, PA, USA, 2Quality of Oncology Care, National Comprehensive Cancer Network (NCCN), Fort Washington, PA, USA, 3Business Development, National Comprehensive Cancer Network (NCCN), Fort Washington, PA, USA, 4Clinical Information and Publications, National Comprehensive Cancer Network (NCCN), Fort Washington, PA, USA, 5Head and Neck Oncology Service, Memorial Sloan Kettering Cancer Center, New York, NY, USA

1

Background: Resource constraints in low- and middle- income countries (LMICs) often impede critical medical care. 65% of new cases of lip and oral cancer, and 76% of related deaths occur in LMICs, where patients lack access to standard diagnostic tests and/or treatment approaches. The development of resource-stratified clinical guidelines promotes access to critical diagnostic and treatment pathways in LMICs. Methods: To address the unmet need in LMICs, a multi-disciplinary committee of NCCN Member Institution experts developed the NCCN FrameworkTM for Head and Neck Cancers: Lip and Oral. In the evidence-based, resource-stratified Guidelines, recommendations from the NCCN Guidelines for Head and Neck Cancers were assigned to specific resource levels, based on access to various interventions and importance in

doi:10.1093/annonc/mdx374 | 389

abstracts achieving clinical outcomes. International experts reviewed the resource-stratified Guidelines to assess utility in LMICs and NCCN approved and published the finalized guidelines. Results: The NCCN Framework for Head and Neck Cancers: Lip and Oral has four resource levels: Basic, Core, Enhanced, and Parent guideline. The Framework for Basic Resources identifies essential services required for minimal standard of care for improvement in outcome; the Core Resources lead to improved outcomes but are not cost prohibitive; the Enhanced Resources recommend additional services that may improve outcomes, but may be cost prohibitive in certain settings. For initial treatment of early stage cancer of the oral cavity (T1-2, N0) as an example, the Enhanced Framework recommends surgical resection and radiation therapy (RT), but not sentinel lymph node (SLN) biopsy, which is recommended in the NCCN parent Guidelines and requires more advanced resources. The Basic Framework recommends surgical resection as the only primary treatment option, since RT may not be available at this resource level. Conclusions: The NCCN Framework for Head and Neck Cancers: Lip and Oral provide LMICs with a system to optimize care in limited resource settings, and a map to improve cancer care incrementally as resources become available. Use of this framework facilitates improved patient care in resource-constrained settings. Legal entity responsible for the study: National Comprehensive Cancer Network Funding: None Disclosure: All authors have declared no conflicts of interest.

Annals of Oncology and involved cervical nodes. Prophylacting dose up to 50 Gy was delivered to the remaining cervical areas. Survival was studied with Kaplan Meier test. Late toxicities were assessed according to SOMA-LENT and RTOG scales in patients with a minimal follow-up of 24 months. Results: Mean age was 16 years (9 20). Sex-ratio was 1,1. Seventy two percent of patients (n ¼ 55) had locally advanced tumor (T3 or T4). Cervical nodal involvement was seen in 95% of cases (n ¼ 71). There were 52 cases (68%) of N2 or N3. Sixty-six patients had neoadjuvant chemotherapy, 10 had concomitant and 5 had both. Five patients had exclusive irradiation. Radiotherapy was monofractionated in 45 cases and bifractionated in the remaining cases. Acute toxicities were tolerable. Mean follow-up was 198 months (28- 289). One patient experienced a local failure. Twenty-six presented metastatic failures. Overall survival rate at 10 years is 67,4%. Disease free survival rate at 10 years is 66,7%. Xerostomia was the most frequent late toxicity (97%). Patients experienced endocrine troubles (hypothyroidism in 19%, amenorrhea in 13%), cerebral necrosis (5cases), osteoradionecrosis (10 cases) and secondary cancer (3 cases). Conclusions: Pediatric nasopharyngeal carcinoma has good prognosis despite frequent locally advanced disease at presentation. Combining radiotherapy and chemotherapy is the standard of care. Late toxicities are often severe and affect the quality of life. Legal entity responsible for the study: University Hospital of Sfax - Tunisia Funding: None Disclosure: All authors have declared no conflicts of interest.

1098P 1094P

Long-term results of chemoradiotherapy for stage III nasopharyngeal carcinoma patients and risk grouping by pretreatment EBV viral load

P-J. Lin1, W-Y. Wang2, Y.C. Liu3, J-C. Lin3 1 Department of Radiation Oncology, Tungs’ Taichung MetroHarbor Hospital, Taichung, Taiwan, 2Department of Nursing, Hung Kuang University, Taichung, Taiwan, 3Radiotion Oncology, Taichung Veterans General Hospital, Taichung, Taiwan Background: No previous study reported the treatment outcome of stage III nasopharyngeal carcinoma (NPC) patients. The aim of this study is to investigate the long-term clinical outcome of stage III NPC patients and do risk grouping by plasma EBV DNA assay for future therapy improvement. Methods: A total of 356 previously untreated, pathologically-proven NPC patients with stage III disease and available pretreatment plasma EBV DNA data were enrolled in this retrospective study. Initial definitive treatment consisted of concurrent chemoradiotherapy or induction chemotherapy plus radiotherapy. Eighty-four of 356 (23.6%) patients also received post-RT adjuvant chemotherapy. Patients with pretreatment EBV DNA > 1000 copies/mL were defined as a high-risk subgroup (n ¼ 106) and the remaining patients as a low-risk subgroup (n ¼ 250). Results: After a median follow-up of 90 months, there were 66 recurrences (18.5%) and 57 deaths (16.0%). The 5-year overall survival (OS), progression-free survival (PFS), distant metastasis failure-free survival (DMFFS), and locoregional failure-free survival (LRFFS) for all 356 patients were 88.4%, 83.9%, 90.5%, and 90.5%, respectively. Thirty-five of 105 (33.0%) high-risk patients developed tumor relapse later, whereas only 12.4% (31/250) low-risk patients had tumor relapse (P < 0.0001) Survival analysis revealed that the high-risk subgroup had significantly worse OS (5-year rate, 79.% vs. 92.8%, P < 0.0001), PFS (73.7% vs. 88.4%, P < 0.0001), DMFFS (80.2% vs. 95.0%, P < 0.0001), and LRFFS (85.6% vs. 92.6%, P ¼ 0.0045) than those of the lowrisk subgroup. Conclusions: Long-term treatment results for Stage III NPC patients were good. Risk grouping identified a subgroup of patients with high pretreatment EBV DNA had a significantly higher relapse rates and worse survivals. Future trial should strengthen treatment intensity for these high-risk patients. Legal entity responsible for the study: Taichung Veterans General Hospital Funding: Taichung Veterans General Hospital Disclosure: All authors have declared no conflicts of interest.

1095P

Nasopharyngeal cancer in children: Long term results the experience of the university hospital of Sfax (Tunisia)

P-J. Lin1, W-Y. Wang2, Y.C. Liu3, J-C. Lin3 Department of Radiation Oncology, Tungs’ Taichung MetroHarbor Hospital, Taichung, Taiwan, 2Department of Nursing, Hung Kuang University, Taichung, Taiwan, 3Radiotion Oncology, Taichung Veterans General Hospital, Taichung, Taiwan

1

Background: Nasopharyngeal carcinoma in children is frequent in Mediterranean area. We aimed to report our experience in the treatment of this entity. Methods: We retrospectively review the records of 76 young patients (G(DPYD*5) p1543V (ATA>GTA); Exon 14 1845 G>T; (E615D) missense mutation, Exon 14 splice variant G>A and Exon 18 (DPYD*6) pV7321) - c2194 G>A (GTT>ATT)}. Descriptive statistics was performed using SPSS version 16 and RStudio. Proportions with 95% CI were described. Fisher’s exact test was performed to see the relationship between DPD mutation status and grade 3-5 adverse events. Results: Consecutive 118 patients were included in this analysis. The median age was 45 years (IQR 37.25-54.00 years). The median cycles of TPF received were 2 (range 14). DPD mutation was seen in 29 patients (24.59%, 95%CI 16.94-32.23%). The mutations were seen in exon 18 in 17 patients (14.4%), exon 13 in 9 patients (7.6%) and in both exon 13 & 18 in 3 patients (2.5%). 100 patients were eligible for assessment of adverse events (84.7%). The rate of grade 3-5 haematological and gastrointestinal adverse events was 64% and 35% respectively. The rate of grade 3-5 haematological (88.5% versus 55.4%, p-0.002) and gastrointestinal adverse events (57.7% versus 27.0%) were higher in DPYD mutated cohort. Conclusions: This study signifies the importance of ethnic difference in drug polymorphism and mutations. The impact of these adverse events in DPD mutated patients justifies doing a DPD mutation prior to subjecting patient to 5FU in head and neck cancer. Legal entity responsible for the study: Tata Memorial Hospital Centre, Mumbai Funding: None Disclosure: All authors have declared no conflicts of interest.

1099P

Radiotherapy related xerostomia in head and neck oncology: A systematic review

D. Johns Dentistry, DM WIMS Medical College, Kozhikode, India Background: Radiotherapy in the head and neck region can lead to salivary gland hypofunction and as a result dry mouth ensues. We have undertaken this systematic review and meta-analysis to estimate the effectiveness of available interventions for radiotherapy-induced xerostomia and hyposalivia. Methods: A systematic review and meta-synthesis techniques were adopted to identify, appraise and synthesize the relevant literature regarding the experience of nutritional symptoms of HNC patients conducted according to the PRISMA guidelines. Several

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology electronic databases such as PubMed, CINAHL, Scopus, PsycINFO and the Cochrane Library databases were searched. Results: 1598 patients from Eighteen studies were included in the systematic review. Cholinergic agonists like Pilocarpine, cevimeline and bethanechol were tested in most studies. Other drugs tested include malic acid, physostigmine, specific monoclonal antibodies like Rituximab, fluoroquinolones, saliva substitutes/mouthcare systems, hyperthermic humidification, acupuncture, acupuncture-like transcutaneous electrical nerve stimulation, low-level laser therapy and herbal medicine. A recent study evaluated the salivary parameters in 4 phases. Results of meta analysis suggests cholinergic agonists were the most effective to improve salivary flow, compared to placebo, although some aspects of the relevant effect size, duration of the benefit, and clinical meaningfulness remain unclear. Conclusions: Pilocarpine, bethanechol and cevimeline should represent the first line of therapy in head and neck cancer survivors with radiotherapy-induced xerostomia and hyposalivation. The use of other treatment modalities cannot be supported on the basis of current evidence. Legal entity responsible for the study: nil Funding: None Disclosure: All authors have declared no conflicts of interest.

1100P

The phase II study of HMB/Arg/Gln against oral mucositis induced by chemoradiotherapy for head and neck cancer patients

T. Yokota1, S. Hamauchi1, Y. Yoshida1, T. Yurikusa2, M. Suzuki2, A. Yamashita3, H. Ogawa4, T. Onoe4, K. Mori5, T. Onitsuka6 1 Division of Gastrointestinal Oncology, Shizuoka Cancer Center, Shizuoka, Japan, 2 Division of Dental and Oral Surgery, Shizuoka Cancer Center, Shizuoka, Japan, 3 Division of Nutrition, Shizuoka Cancer Center, Shizuoka, Japan, 4Division of Radiation Oncology, Shizuoka Cancer Center, Shizuoka, Japan, 5Clinical Trial Coordination Ofﬁce, Shizuoka Cancer Center, Shizuoka, Japan, 6Divisions of Head and Neck Surgery, Shizuoka Cancer Center, Shizuoka, Japan Background: Opioid-based pain control and systemic oral care program are effective for the chemoradiotherapy (CRT)-induced severe oral mucositis (OM) in patients with head and neck cancers (HNC). This phase II trial assessed the clinical benefit of betaHydroxy-beta-Methylbutyrate, Arginine, and Glutamine (HMB/Arg/Gln) in the prevention of CRT-induced OM in patients with HNC. Methods: Patients with HNC who were scheduled to receive definitive or postoperative cisplatin-based CRT were enrolled. HMB/Arg/Gln was administered orally or per percutaneous endoscopic gastrostomy from the first day of CRT up to completion of CRT. All patients received opioid-based pain control and oral care programs we previously published. The primary endpoint was the incidence of grade 3 OM (functional/symptomatic) according to the Common Terminology Criteria of Adverse Events version 3.0. QOL (EORTC QLQ-C30/PROMS) and intake of nutrition at baseline and 50Gy were also assessed. Results: From February 2015 to June 2016, 35 patients with HNC were enrolled. Sixteen patients (45.7%) developed grade 3 OM (i.e., functional/symptomatic). The incidence of grade 1 OM (functional/symptomatic) was 51.5% at 2 weeks and 82.9% at 4 weeks after completion of RT. Clinical examination revealed that 10 patients (28.6%) developed grade 3 OM. The incidence of grade 1 OM (clinical exam) was 80.0% at 2 weeks and 100% at 4 weeks after completion of RT. Only 5.7% of patients had unplanned breaks in radiotherapy, and all patients completed treatment. Adverse events related to HMB/Arg/Gln were increase in blood urea nitrogen and diarrhea, but were easily managed. Conclusions: Addition of HMB/Arg/Gln to opioid-based pain control and oral care programs was feasible but still insufficient in reducing the incidence of severe CRTinduced oral mucositis. However, the benefit of HMB/Arg/Gln should not be neglected in terms of findings of clinical examination and the recovery from severe oral mucositis. Clinical trial identification: UMIN000016453 Legal entity responsible for the study: None Funding: Public Interest Incorporated Foundation- Shizuoka Industrial FoundationPharma Valley Center Disclosure: All authors have declared no conflicts of interest.

1101P

Oral mucosa dose parameters predicting grade 3 acute toxicity in locally advanced nasopharyngeal carcinoma patients treated with concurrent intensity-modulated radiation therapy and chemotherapy

K. Li1, L. Yang2, Y-Y. Chen3, M. Chen3 Department of Radiation Oncology, First Hospital of Quanzhou Afﬁliated to Fujian Medical University, Quanzhou, China, 2Department of Oncology, Xinyu Peole’s Hospital, Xinyu, China, 3Radiotherapy, Zhejiang Cancer Hospital, Hangzhou, China

1

(LANPC) treated with concurrent intensity-modulated radiation therapy (IMRT) and chemotherapy. Methods: A standardized method for the oral cavity (oral cavity contours, OCC) and a novel method for the mucosal surface (mucosal surface contours, MSC) were developed for the oral mucosa and prospectively applied to the radiation treatment plans of 92 patients treated with concurrent IMRT and chemotherapy for LANPC. Dosevolume histogram (DVH) data were extracted and analyzed against patient toxicity. Receiver operating characteristic analysis and logistic regression were carried out for both contouring methods. Results: Grade 3 oral mucosa toxicity occurred in 20.7% (19/92) of patients in the study. A highly significant dosevolume relationship between oral mucosa irradiation and acute oral mucosa toxicity was supported by using both oral cavity and mucosal surface contouring techniques. In logistic regression, body weight loss was an independent factor related to grade 3 toxicity for OCC and MSC (p¼0.017 and 0.005, respectively), and the independent factor of dosimetric parameters for OCC and MSC were V30Gy (p¼0.003) and V50Gy (p¼0.003), respectively. In the receiver operating characteristics curve, the areas under V30Gy of the OCC curves was 0.753 (p¼0.001), and the areas under V50Gy of MSC curves was 0.714 (p¼0.004); the cut-off value was 73.155% (sensitivity, 0.842; specificity, 0.671) and 14.32% (sensitivity, 0.842; specificity, 0.575), respectively. Conclusions: DVH analysis of mucosal surface volumes accurately predicts grade 3 oral mucosa toxicity in patients with LANPC receiving concurrent IMRT and chemotherapy, but the MSC method is still no better than the OCC method in clinical application. Clinical trial identification: NCT02945878 Legal entity responsible for the study: Yuanyuan Chen Funding: None Disclosure: All authors have declared no conflicts of interest.

1102P

Sinonasal non-glandular cancers relapsing after multimodal treatments

S. Cavalieri1, R. Granata1, L.D. Locati1, C. Bergamini1, S. Alfieri1, C. Resteghini1, D. Galbiati1, E. Orlandi2, N.A. Iacovelli2, G. Calareso2, M. Guzzo3, P. Quattrone4, L. Licitra5, P. Bossi1 1 Head and Neck Medical Oncology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 2Radiotherapy, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 3Otorhinolaryngology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 4Pathology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 5Head and Neck Medical Oncology, Fondazione IRCCS - Istituto Nazionale dei Tumori Universit a degli studi di Milano, Milan, Italy Background: Multimodality treatment (MMT) is the current approach to advanced sinonasal cancers (SC). We lack salvage treatment standardization especially for pts already receiving MMT. No clinical factors able to predict outcome have been identified in this disease setting. Methods: We retrospectively analyzed a series of pts with recurrent/metastatic (RM) SC after multimodal curative treatment, consisting in induction chemotherapy (iCT) followed by locoregional therapy. Overall survival (OS) was measured as the interval from relapse to death. Results: Among 106 pts with SC treated with MMT at our Center from 1997 to 2016, 50 (M/F 31/19) relapsed. Median age was 53 yrs (16-73). Median follow-up was 26 months (m) (5-192). WHO 2005 histotypes were: 36% sinonasal undifferentiated carcinoma (SNUC), 34% squamous cell cancer (SCC), 30% carcinomas with neuroendocrine differentiation (CND). Median time to first relapse after curative treatment was 13.5 m. Median OS was 13 m from recurrence: 19 m in SCC, 16 m in SNUC and 6 m in CND (p ¼ .34). Relapse occurred as distant metastasis in 40%, as nodal recurrence in 6% and at primary site in 54% of cases. First line salvage treatment was surgery in 38% (14 pts received surgery on T, 2 on N and 3 on M), CT in 30%, RT in 8%, best supportive care in the remaining pts. Median OS was 31 m in surgically treated pts and 4.8 m in those receiving CT (p < .0001). In pts with disease control (PRþSD) after iCT, median OS after recurrence was longer than in pts with PD (13.4 vs 1.5 m, p ¼ .07). Median OS from relapse was 29.6 m in pts with CR after definitive treatment, 7.1 m in those with PR and 3.4 m in those with PD (p ¼ .002). Pts with an objective response to palliative CT had a longer median OS than those with PD (20 vs 4 m, p ¼ .002). Conclusions: Prognosis of SC relapsing after MMT is dismal. With the caveat of a retrospective analysis and a case series that has been collected in a long time frame, we showed that feasibility of salvage surgery, objective response to prior definitive treatment and response to palliative CT are factors associated with better outcomes. Pts with relapsed or metastatic SC not amenable to salvage surgery should be considered for enrolment in clinical trials. Legal entity responsible for the study: Fondazione IRCCS Istituto Nazionale Tumori Milano - Universita degli Studi di Milano. Italy Funding: None Disclosure: All authors have declared no conflicts of interest.

Background: To determine whether volumes based on the contours of the mucosal surface can be used instead of the contours of the oral cavity to predict for grade 3 acute oral mucosa toxicity in patients with locally advanced nasopharyngeal carcinoma

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx374 | 391

abstracts 1103P

Survival outcome and optimal treatment of intermediate-grade salivary gland carcinoma

G. Park Otolaryngology, Samsung Changwon Hospital Sung Kyun Kwan University, Changwon, Republic of Korea Background: Histological grade is the most important factor for defining treatment strategies and predicting prognosis for salivary gland carcinoma (SGC). Although several studies have addressed low- and high-grade SGCs, intermediate-grade SGC (IGSGC) has received minimal attention. Therefore, we examined factors affecting long-term recurrence and survival among IGSGC patients to define optimal treatment modalities and outcomes. Methods: We reviewed the clinical and pathological data of 108 IGSGC patients who underwent definitive surgery with or without postoperative radiotherapy at our tertiary referral center between 1994 and 2014. We performed univariate and multivariate analyses of variables predictive of locoregional control (LRC), distant metastasis-free survival (DMFS), and overall survival (OS). We compared treatment outcomes by treatment strategies such as surgical extent, primary tumor, neck dissection, or postoperative radiotherapy. Results: During a median 103 (range, 24282)-month follow-up, local, regional, and distant recurrences were detected in 14 (13.0%), 3 (2.8%), and 21 (19.4%) patients, respectively. The 10-year LRC, DMFS, and OS rates were 83.1%, 76.0%, and 80.1%, respectively. Multivariate analyses identified a non-parotid primary site as an independent prognostic factor for LRC (P ¼ 0.018), Adenoid cystic carcinoma and positive pN classification were significantly unfavorable prognostic factors for DMFS (P ¼ 0.025 and P ¼ 0.030, respectively); overall advanced stage was an independent prognostic factor for OS (P ¼ 0.020). Surgical extent, elective neck dissection, and postoperative adjuvant radiotherapy did not significantly affect treatment outcomes. Conclusions: Patients with early-stage IGSGC of parotid origin can achieve favorable treatment outcomes with conservative surgery alone. Legal entity responsible for the study: no Funding: None Disclosure: All authors have declared no conflicts of interest.

1104P

Incidence and survival of secondary malignances (SM) in oropharingeal squamous cell carcinoma (OPSCC): A homogeneous single report institution

M. Napolitano1, F. Bertolini1, E. D’Angelo2, A. Spallanzani1, S. Tassi3, B. Meduri2, S. Bettelli4, R. Depenni1, A. Ghidini3, F. Lohr2, L. Presutti3, S. Cascinu1 1 Department of Oncology, Modena University Hospital, Modena, Italy, 2Radiotherapy Service, Modena University Hospital, Modena, Italy, 3ENT Surgery Dept, Modena University Hospital, Modena, Italy, 4Molecular Biology Lab, Modena University Hospital, Modena, Italy Background: SM in HNSCC patients (pts) are common, due to the presence of risk factors (smoking habit or alcohol abuse). Aim of this report is to evaluate the incidence and characteristics of SM in a series of OPSCC. Methods: We retrospectively reviewed clinical data of 266 pts with OPSCC seen at Modena University Hospital between 2006 and 2016. We recorded data from a web platform in which every pt has a personal form filled with clinical information. In particular, we analyzed the rate of SM and described clinical and survival data. Results: SM was diagnosed in 37 pts (13,9%): 15 NSCLC (5% on all; 40,5% of SM); 7 HNSCC (18,9%), 8 GI (21,6%), 2 prostate (5,4%), 2 thyroid cancer (5,4%), 2 hematologic malignancy (5,4%) and 1 melanoma (2,7%). Clinical features at diagnosis for OPSCC: 30 (81%) male, 7 (19%) female; median age 68 years (range 37-90). Twentyfive pts (67,6%) were current/former smokers, 26 (70,3%) HPV-positive; stage at diagnosis was I-II in 5 (13,5%) and III-IV in 32 pts (86,5%). Eleven pts developed SM < 12 vs 26 12 months (mo) after the diagnosis of OPSCC. Stage at diagnosis for SM was: for lung 10 (66, 6%) I-II vs 5 (33,4%) III-IV; for HNSCC 3 (42,8%) I-II vs 4 (57,2%) III-IV; for GI 2 (25%) I-II vs 6 (75%) III-IV. Treatments for SM: 18 surgery, 2 RT, 8 CT, 5 combined treatment; 4 pts did not need or not received therapy. Death occurred in 18 pts (48,6%): SM-related in 12 (66,6%), OPSCC-related in 3 (16,7%) and not cancer-related in 3 (16,7%). mOS from diagnosis of OPSCC vs SM were 68,5 and 21,3 mo, respectively. Pts with lung or HNSCC SM (mOS 6,7 mo) had worse OS than pts with other SM (mOS 20,7 mo), but not statistically significant. Pts with SM diagnosed 12 mo vs < 12 mo after OPSCC had a significantly better OS (mOS 81 vs 25,1 mo; p < 0.001). Conclusions: In our retrospective series, we confirmed that secondary lung cancer was the most frequent SM; it was diagnosed at earlier stage, because these pts underwent a periodical follow-up for their previous OPSCC with a chest X-ray/CT. Smokers may benefit from a more intensive follow-up for a higher risk of smoking related SM (lung, HNSCC). Survival is more influenced by the occurrence of SM than by OPSCC. All these considerations should be applied to a larger series. Legal entity responsible for the study: Modena University Hospital Funding: None Disclosure: All authors have declared no conflicts of interest.

392 | Head and neck cancer, excluding thyroid

Annals of Oncology

1105TiP

Pembrolizumab plus chemoradiation (CRT) for the treatment of locally advanced head and neck squamous cell carcinoma (LAHNSCC): Phase 3 KEYNOTE-412 trial

J-P. Machiels1, L. Licitra2, D. Rischin3, J. Waldron4, B. Burtness5, V. Gre´goire1, T. Shekar6, H.M. Brown6, J. Cheng6, L.L. Siu7 1 Oncology, Cliniques Universitaires Saint-Luc, Brussels, Belgium, 2Oncology, Fondazione IRCCS Istituto Nazionale Tumori Milano, Milan, Italy, 3Oncology, Peter MacCallum Cancer Centre, East Melbourne, Australia, 4Oncology, Princess Margaret Cancer Centre, Toronto, ON, Canada, 5Medicine, Yale University School of Medicine, New Haven, CT, USA, 6Medical Oncology, Merck & Co., Inc., Kenilworth, NJ, USA, 7Division of Medical Oncology and Hematology, Princess Margaret Cancer Centre, Toronto, ON, Canada Background: Despite significant advancements in oncologic treatment, the outcome for patients with advanced or recurrent HNSCC is poor. Identification of abscopal effects by use of radiotherapy (RT) in combination with immunotherapy in a patient with metastatic melanoma has prompted interest in the use of combination regimens. The objective of the KEYNOTE-412 trial (NCT03040999) is to assess efficacy and safety of pembrolizumab in combination with CRT as maintenance therapy for subjects with LA-HNSCC. Trial design: KEYNOTE-412 is a phase 3, randomized, placebo-controlled, doubleblind trial enrolling subjects with newly diagnosed, treatment-naive, oropharyngeal p16 positive (any T4 or any N3), oropharyngeal p16 negative (any T3-T4, or N2a-N3), or larynx/hypopharynx/oral cavity (any T3-T4, any N2a-N3) SCC. Approximately 780 subjects will be randomly assigned (1:1) to receive pembrolizumab plus cisplatin-based CRT or placebo plus cisplatin-based CRT. Subjects will be stratified by RT regimen, tumor site/p16 status, and disease stage. Treatment will include a priming dose of pembrolizumab 200 mg or placebo 1 week before initiation of CRT, followed by 7 weeks’ CRT (cisplatin 100 mg/m2 every 3 weeks [Q3W] [3 doses]; accelerated RT [70 Gy, 6 fractions/week] or standard RT [70 Gy, 5 fractions/week]) plus pembrolizumab 200 mg Q3W or placebo Q3W. Treatment with pembrolizumab 200 mg Q3W or placebo Q3W will continue up to 1 year (maximum 17 doses). Treatment will continue until disease progression, unacceptable toxicity, withdrawal of consent, or investigator decision to withdraw the patient. Response will be assessed by computed topography or magnetic resonance imaging 12 weeks after completion of CRT, every 4 months for a subsequent 2 years, and every 6 months thereafter up to year 5. Safety will be monitored throughout the study. The primary end point is event-free survival by blinded independent central review per RECIST v1.1. Secondary end points include overall survival, safety, and quality of life. Exploratory biomarker analyses will be conducted. Clinical trial identification: NCT03040999, February 1, 2017 Legal entity responsible for the study: Merck & Co., Inc., Kenilworth, NJ, USA Funding: Merck & Co., Inc., Kenilworth, NJ, USA Disclosure: J-P. Machiels: Advisory board member: MSD (uncompensated), Innate, AstraZeneca, Nanobiotix, Debio; Research funding: Bayer, Janssen, Novartis. L. Licitra: Travel expenses, including accommodations: Merck-Serono, Debiopharm, Jobi, Bayer, Amger; Consulting or Advisory Role: Eisai, Bristol-Myers Squibb, MSA, Merck-Serono, Debiopharm, Jobi, Novartis, AstraZeneca, Bayer, Roche, Amgen. D. Rischin: Research funding: Genentech/Roche, Merck, Threshold Pharmaceuticals. B. Burtness: Advisory board member: Merck, Boehringer Ingelheim, Celgene, AstraZeneca, Bristol-Myers Squibb, Amgen; Research funding: Merck, Advaxis, Innate. T. Shekar, H.M. Brown: Employment and Stock ownership: Merck. J. Cheng: Employment and Stock ownership: Merck MSD. L.L. Siu: Advisory board: Merck, AstraZeneca/MedImmune, Boehringer Ingelheim, Celgene, and Pfizer; Research funding: AstraZeneca/ MedImmune, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Celgene, Genentech/Roche, GlaxoSmithKline, Merck, Novartis, and Pfizer. All other authors have declared no conflicts of interest.

1106TiP

Hope for salivary gland cancer (SGC): EORTC HNCG/UKCRN 1206 randomized phase II study to evaluate the efficacy and safety of Chemotherapy (CT) vs androgen deprivation therapy (ADT) in patients with recurrent and/or metastatic androgen receptor (AR) expressing SGC

L.D. Locati1, C. Caballero2, C. Fortpied2, F. Perrone3, S. Pilotti4, K.J. Harrington5, V. Gre´goire6, L. Licitra1 1 Head and Neck Medical Oncology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 2Medical, EORTC, Brussels, Belgium, 3Molecular Pathology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 4Anatomic Pathology, Fondazione IRCCS - Istituto Nazionale dei Tumori, Milan, Italy, 5Radiotherapy and Imaging, Royal Marsden NHS Foundation Trust/The Institute of Cancer Research, London, UK, 6Dept. Radiation Oncology, Cliniques Universitaires St. Luc, Brussels, Belgium Background: SGCs are rare and heterogenous tumors ( 350 patients (pts) in Germany. Methods: Data were extracted from pt medical records from office-based urologists and urology clinics in Germany. Adult pts (age 18 y) diagnosed with T4b, N2-3 and/ or M1 UC and received first-line (1L) or second-line (2L) palliative chemotherapy from 2009 to 2016 were included. The index date was the start date of first systemic therapy. We described tx patterns and clinical characteristics; Kaplan-Meier method assessed overall survival (OS). Cox regression adjusted for age, Eastern Cooperative Oncology Group performance status (ECOG PS) and liver metastases, stratified by hospital/office, compared tx. Results: Among 368 included pts, 356 and 107 received 1L and 2L tx, respectively. At the start of 1L therapy, mean age was 68 y, 73% of pts were male, 74% were current/exsmokers and 63% had metastatic disease. In 1L, 75% of pts received dual-combination tx, most commonly gemcitabine þ cisplatin (GemCis; 83%). In 2L, 74% received single-agent tx, most commonly vinflunine (66%). In 1L, 12-month OS was 60%, slightly higher with GemCis (65%) than with other tx (52%). No difference in OS by sex or smoking status was noted. Pts with and without renal impairment (creatinine clearance 1%) observed a 30% response rate (n ¼ 10, 1 CR and 2 PR) and a 50% DCR. Whole exon sequencing was performed on selected biopsy samples. Mutations on p53, MDM2, TAP2 et al, might contribute to the favorable response to immunotherapy. Interestingly, a divergent spectrum of mutations from mixed

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology response patients were observed on tumor cells from different metastases, which at the time of biopsy had drastically different clinical response to the treatment. Conclusions: JS001 demonstrated an acceptable safety profile in solid tumor patients. Additional phase II studies to evaluate the safety and clinical activity of JS001 in selected tumor types are ongoing. Clinical trial identification: Clinical Trial ID: NCT02857166 Legal entity responsible for the study: Sun Yat-sen University Cancer Center Funding: None Disclosure: All authors have declared no conflicts of interest.

1187P

A Phase 1/2 trial of intratumoral (i.t.) IMO-2125 (IMO) in combination with checkpoint inhibitors (CPI) in PD-(L)1-refractory melanoma

A. Diab1, C. Haymaker2, M. Uemura2, R. Murthy3, M. James4, J. Geib5, M. Cornfeld5, S. Swann5, C. Yee4, J. Wargo4, R. Amaria4, S. Patel4, H. Tawbi4, I. Glitza4, S. Woodman4, W-J. Hwu4, M.A. Davies4, W. Overwijk4, C. Bernatchez2, P. Hwu6 1 Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA, 2Melanoma Medical Oncology - Research, MD Anderson Cancer Center, Houston, TX, USA, 3 Interventional Radiology Department, The University of Texas MD Anderson Cancer Center, Houston, TX, USA, 4Department of Melanoma Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA, 5Clinical Science, Idera Pharmaceuticals, Exton, PA, USA, 6Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA Background: CPI have transformed melanoma treatment, however many patients remain refractory and subsequent treatment options are limited. IMO, a Toll-like receptor 9 agonist, may improve response to CPI by activating innate and adaptive immune responses to overcome immune escape. Initial clinical experience with IMO þ ipi is promising (Uemera, ASCO-SITC 2017). Dose-finding is now complete and is the basis for this updated report. Methods: Adults with unresectable or metastatic melanoma refractory to a PD-(L)1 inhibitor are eligible if they have tumor accessible to biopsy. IMO is administered i.t. to a single tumor at escalating doses during weeks 1,2,3,5,8, and 11 along with ipi or pem per the product label. The primary endpoint of Phase 1 is safety and for Phase 2 is overall response rate using a 2-stage design. Serial biopsies are obtained from both the injected and a non-injected lesion for immune analysis. Results: A total of 22 subjects have been treated with either IMO-ipi (N ¼ 18) or IMOpem (N ¼ 4) and dose-escalation is now complete for the IMO-ipi arm. Dose-limiting toxicities have not been reported. Immune-related AE were observed in 4 IMO-ipi subjects [hypophysitis (N ¼ 2), hepatitis (1), colitis (1)]. These responded well to standard measures. Of 9 patients treated at the RP2D of 8mg, 6 have experienced clinical benefit (1CR, 1PR, 1uPR, 3 SD). Biopsies show maturation of the mDC1 subset (CD1cþCD303-), upregulation of PD-L1 by malignant cells, and an IFNa response gene signature. Biopsies of uninjected tumors show expression of CD56þ and Ki67þ effector CD8þT cells in responding patients, indicative of an abscopal effect. Phase 2 accrual using the 8 mg IMO dose is ongoing. Conclusions: IMO þ ipi is a viable strategy to revive the immune response in CPIresistant tumors and shows preliminary clinical activity worthy of further development. Clinical trial identification: NCT02644967 Legal entity responsible for the study: Idera Pharmaceuticals Funding: Idera Pharmaceuticals Disclosure: J. Geib, S. Swann: Employment by Idera Pharmaceuticals. M. Cornfeld: Employment at Idera Pharmaceuticals. All other authors have declared no conflicts of interest.

1188P

Clinical and immune effects patients with progressive disease treated with low dose of anti-CTLA-4, bortezomib, gemcitabine, naproxen and meloxicam

J.P. Marquez-Manriquez1, J.A.M. Briseno2, K. Quayle3, L.M. Fernandez3, I.R. Barajas3, G. Torres-Montano3, F.A. Durazo-Bustamante4, A. Durazo-Acuna4, L-A. Rangel-Reyna3, D. Lopez-Hernandez3, A. Rodriguez-Jasso2, M-A. Trujillo-Acosta3, E. Ramos-Garcia4, S. Icedo-Zamora4, P.A.L. Diaz4 1 ImmunOncology Clinical Reserach, Centro de Investigacion de Cancer en Sonora campus Seattle (CICS USA), Seattle, WA, USA, 2ImmunoOncopathology, Centro de Investigacion de Cancer en Sonora campus Ciudad Obregon, Sonora, Mexico, Ciudad Obregon, Mexico, 3ImmunoOncology Clinical Reserach, Centro de Investigacion de Cancer en Sonora campus Seattle (CICS USA), Seattle, WA, USA, 4Clinical Oncology and ImmunoOncology, Centro de Investigacion de Cancer en Sonora campus Ciudad Obregon, Sonora, Mexico, Ciudad Obregon, Mexico Background: Several patients progressed with their cancer disease despite treatment and eventually they become refractory. We selected patients from several malignancies with PD despite standard of care treatment (n ¼ 30) and performed a pilot clinical study to evaluate the effect of two intravenously, two oral and one subcutaneously agent. With this in mind and with a systematic review and immunomodulatory, antiangiogenic and anti-tumoral validation of each drug was studied. We tested the

Volume 28 | Supplement 5 | September 2017

preexisting CD8 and Th1 antigen specific immune response against several clinically relevant peptides from bad prognosis proteins. Methods: 30 subjects were included after the CICS ethics committee approved the protocol. The inclusion criteria include ECOG¼0, complete CT scan from neck, thorax, abdomen and pelvis, laboratory tests such as CBC, phase acute proteins, etc. The patients were accepted after initial IFN-gamma and Elispot assays were done to make sure we have only patients with Th1 and CD8 immune response, as we know that ipilimumab unleashes every T cell. The tumors included were PDAC (n ¼ 5), HGSOC (n ¼ 12), TNBC (n ¼ 10) and MM (3). The patients received the oral and the IV treatment biweekly for 4 months. Results: We had 60% of CR and 40% of PR. The tumor with more significant response was ovarian (90%). There was an immunological correlation of CD8 immune response between in both CR (p ¼ 0.001) and PR (p ¼ 0.05). The combination was well tolerated and after 16 months of stopping the treatment some patients have persistent CD8 antigen specific immune response. Conclusions: The combination is clinically feasible, looks promising and we now understand the importance of preserving the immune response and the use of biomarkers to improve the rational and generate new combinations with this approach to improve clinical outcomes. Legal entity responsible for the study: CENTRO DE INVESTIGACION DE CANCER EN SONORA CAMPUS CIUDAD OBREGON, SONORA, MEXICO Funding: Fundacion del centro de investigacion de cancer en sonora (cics) campus ciudad obregon, sonora, Mexico. Disclosure: All authors have declared no conflicts of interest.

1189P

4SC-202 plus anti-PD1: Breaking PD1-refractoriness to increase efficacy of checkpoint inhibition in patients with advanced melanoma

F. Hermann Clinical Development, 4SC AG, Planegg-Martinsried, Germany Background: Despite successes in the treatment of melanoma patients with checkpoint inhibitors (CI), majority of patients do not respond to CI alone and a high unmet medical need remains for these patients. One promising approach is to enhance the immunogenicity and alter the tumor microenvironment from an immune-deserted to an inflamed phenotype with combination therapy. Epigenetic modulation has been reported as one key determining factor in shaping the immune microenvironment and compounds altering these processes (e.g. histone deacetylases (HDAC) inhibitors) are particularly promising. Methods: Tumor bearing animals (CT26 & C38 syngenic models) were treated with 4SC-202, an oral clinical stage combined HDAC class I/LSD1 inhibitor, or CIs PD-(L)1 alone and in combination. Tumor growth was assessed continuously and after approx. 2 weeks of treatment tumors were excised and analyzed by flow cytometry and gene expression profiling. Additionally, animals not intended for these analyses were further monitored and tumor growth/survival was monitored. Results: 4SC-202 treatment led to an increase of MHC molecules and enhanced expression of inflammatory markers like IFN-c and various chemokines in tumors. Detailed analysis of the tumors revealed that 4SC-202 strongly altered the immune cell composition; particularly the number of cytotoxic T cells (CTL) was markedly increased. Importantly, subsequent combination treatment of 4SC-202 with CIs in syngenic animal models showed a strong synergistic effect resulting in significant longer survival in both models leading to 55% of tumor free animals (C38 model). Conclusions: In an upcoming study, patients with advanced melanoma who are refractory/non-responding to anti-PD-1 antibodies will be treated with 4SC-202 plus anti-PD1. These patients do not only represent a population with a high unmet medical need but melanoma also represents a model tumor for immunotherapy in general and CI in particular. We hypothesize that addition of 4SC-202 to anti-PD-1 antibody treatment may lead to increased immunogenicity of the tumor, an inflamed tumor microenvironment and ultimately to clinical benefit in anti-PD-1 refractory/non-responding advanced-stage melanoma patients. Legal entity responsible for the study: 4SC AG Funding: 4SC AG Disclosure: F. Hermann: Employee of 4SC AG, Planegg-Martinsried, Germany.

1192P

Cytomegalovirus reactivation in patients with refractory checkpoint inhibitor-induced colitis

C. Franklin1, I. Rooms1, M. Fiedler2, H. Reis3, L. Milsch1, S. Herz1, E. Livingstone1, L. Zimmer1, K.W. Schmid3, U. Dittmer2, D. Schadendorf1, B. Schilling1 1 Department of Dermatology, Venerology and Allergology, University Hospital Essen, West German Cancer Centre, Essen, Germany, 2Institute for Virology, University Hospital Essen, West German Cancer Centre, Essen, Germany, 3Department of Pathology, University Hospital Essen, West German Cancer Centre, Essen, Germany Background: Objectives: Immune checkpoint inhibitors have become a standard treatment in patients with metastatic malignant melanoma. Showing significant anti-tumor effects by unleashing the immune-system, checkpoint inhibitors can also cause highgrade immune-related adverse events, with immune-related diarrhea and colitis

doi:10.1093/annonc/mdx376 | 421

abstracts (irColitis) being amongst the most frequent ones. While the majority of patients with irColitis respond well when treated according to standard treatment algorithms with corticosteroids þ/- other immunomodulatory drugs such as infliximab, some patients do not show resolution of diarrhea and colitis. In the present study, we analyzed the frequency of therapy-refractory irColitis, the underlying cause and useful diagnostic measures. Methods: In this retrospective, monocenter study we collected data of 370 patients with metastatic malignant melanoma. All patients had been treated with checkpoint inhibitors at the skin cancer unit of the Department of Dermatology at the University Hospital Essen from 2006-2016. Demographic and clinical data of all patients were collected. Digital patient records of all 370 patients were searched for the terms “diarrhea” and “colitis”. Results: We identified 41 patients with irColitis, the majority occurring during treatment with ipilimumab. Amongst these patients, 5 (12.2%) were refractory to standard immunomodulatory treatment with corticosteroids and infliximab. Therapyrefractory cases tended to show more severe inflammation in colonic biopsies performed during colonoscopy (p ¼ 0.04). CMV-DNA in colonic biopsies and in plasma was significantly more often detectable in therapy-refractory cases (80% vs. 6.75% in non-refractory cases in biopsies, 80% vs. 0% in plasma). Presence of serum CMV IgM as well as positive immunohistochemical stainings of colon biopsies for CMV were also strongly associated with refractory colitis (40% in refractory vs. 0% in non-refractory cases), but not reliable markers in the majority of refractory patients. Conclusions: This report on CMV reactivation during management of checkpoint inhibitor induced colitis emphasizes the need for repetitive diagnostic measures in treatment-refractory irColitis. Legal entity responsible for the study: Ethics comittee of the University Hospital Essen, University of Diusburg-Essen Funding: None Disclosure: All authors have declared no conflicts of interest.

1193P

Pooled data analysis of the safety and tolerability of intravenous pelareorep in combination with chemotherapy in 500 1 cancer patients

A.A. Gutierrez, C. Reid, M. Crawford, K. Cheetham, A. Dzugalo, M. Parsi, A. Penman, N. Noronha, D. Galindez, R. O’Flynn, M. Coffey Clinical Development, Oncolytics Biotech Inc, Calgary, Canada Background: Oncolytic viruses are promising cancer immunotherapies but questions have been raised regarding their safety. Pelareorep (REOLYSIN, R), an unmodified Reovirus Dearing strain, selectively replicates and lyses cancer cells and induces antitumor immunity. To date, 900þ patients (pts) have been treated with intravenous (IV) pelareorep. In a phase 2 trial, its combination with paclitaxel improved overall survival (17.4m) vs paclitaxel (10.4m) in metastatic breast cancer (MBC) pts (HR 0.65, 80% CI 0.46-0.91, p ¼ 0.1; Berstein et.al. AACR2017). A pooled analysis was thus conducted to better characterize pelareorep’s safety profile in combinations with paclitaxel. Methods: 1417 pts have been enrolled in 36 trials: 934 pts received IV pelareorep and 359 were in control arms. Data from 8 trials with paclitaxel (P), paclitaxel þ pelareorep (PR), carboplatin þ paclitaxel (CP) or carboplatin þ paclitaxel þ pelareorep (CPR) were pooled. Standard doses of P (weekly) and CP were administered. Pelareorep IV dose was 3x1010 TCID50 (5-6 doses q21-28 d). Various advanced solid tumors were evaluated, including the 81 pts with MBC. Results: A total of 563 pts were included in P (86), PR (95), CP (118) or CPR (264) groups. Median age (59-62 y) and ECOG 0-1 status (90-96%) were similar across the groups. All pts in P or PR had received prior chemo but only 26% in CP and 38% in CPR. Fatigue was the most common grade 3 treatment related adverse event (TRAE) in PR (9.5%) and CPR (8.3%) vs P (8.1%) and CP (2.5%). Grade 3 neutrophil count decreased and/or WBC decreased were more frequent in PR (15.8%/17.9%) than in P (5.8%/3.5%), but addition of pelareorep did not increase the frequency or severity of other grade 3 TRAEs with P or CP. Serious TRAEs (%) of interest in P vs PR and CP vs. CPR, included: fever (0 vs 3.2 & 0 vs 3.8), febrile neutropenia (0 vs 1.1 & 3.4 vs 3.4), sepsis (1.2 vs 0 & 0 vs 1.5) and flu-like syndrome (0 vs 1.1 & 0 vs 0.8). Conclusions: This is the largest database reported to date examining the safety of an IV viral agent. Pelareorep’s administration, in combination with paclitaxel or carboplatinpaclitaxel, is safe and well tolerated. Continued evaluation in a registration trial is planned. Clinical trial identification: NCI-US NCI-GOG 0186H (NCT01199263). Ongoing, but not recruiting NCI-8601 (NCT01280058). Ongoing, but not recruiting NCI- Canada (CCTG) NCIC-CTG IND.213 (NCT01656538). Ongoing, but not recruiting Oncolytics Biotech Inc: REO011 – UK (Karapanagiotou et al. CCR 2012) Completed REO015 (NCT00753038) Completed REO016 (NCT00861627) Completed REO018 (NCT01166542). Completed REO021 (NCT00998192) Completed Legal entity responsible for the study: Studies were sponsored/conducted by NCI-US, NCI-Canada (CCTG) or Oncolytics Biotech Inc. See section of Clinical Trial Identification

422 | Immunotherapy of cancer

Annals of Oncology Funding: NCI-US and NCI- Canada conducted their own studies and Oncolytics Biotech only provided the drug. Some studies were fully supported and conducted by Oncolytics. Details can be found in the Clinical Trial Identification section Disclosure: A.A. Gutierrez: Chief Medical Officer and an employee of Oncolytics Biotech Inc. (or one of its affiliated corporations). Own shares in or have options to purchase shares in Oncolytics Biotech Inc. C. Reid: Paid consultant of Oncolytics Biotech. M. Crawford: Employee of Oncolytics Biotech Inc. (or one of its affiliated corporations) and own shares in or have options to purchase shares in Oncolytics Biotech Inc. K. Cheetham, A. Penman, N. Noronha: Employee of Oncolytics Biotech Inc. (or one of its affiliated corporations) and own shares in or have options to purchase shares in Oncolytics Biotech Inc. A. Dzugalo: Employee of Oncolytics Biotech Inc. (or one of its affiliated corporations) and own shares in or have options to purchase shares in Oncolytics Biotech Inc. M. Parsi, D. Galindez, R. O’Flynn: Paid consultant of Oncolytics Biotech. M. Coffey: President and CEO of Oncolytics Biotech Inc. As an employee of Oncolytics Biotech Inc. (or one of its affiliated corporations) he owns shares in or have options to purchase shares in Oncolytics Biotech Inc.

1194P

Impact of prior immune checkpoint inhibitors on haematological toxicity in phase I patients receiving chemotherapy

A.H. Ingles Garces1, J.E. Ang1, M. Ameratunga1, M. Chenard-Poirier1, D. Dolling2, R. Sundar1, S. Kaye1, J. de Bono1, U. Banerji1, J. Lopez1 1 Drug Development Unit, The Royal Marsden NHS Foundation Trust, London, UK, 2 Oncology, Institute of Cancer Research Royal Marsden Hospital, Sutton, UK Background: Immune checkpoint inhibitors (ICI) are used increasingly and earlier to treat multiple cancers. Although rates of on-treatment myelotoxicity are low, there are no published data on the long-term effects of ICI. This is a pilot study to evaluate the impact of prior ICI exposure on chemotherapy-related myelotoxicity in patients in the Phase I setting. Methods: We conducted a retrospective chart review of patients treated between 2012 and 2016 in the Drug Development Unit, The Royal Marsden Hospital. Multivariate logistic regression (including number of previous treatment lines and type of chemotherapy) was used to assess possible relationships between G3/4 neutropenia or thrombocytopenia and previous treatment with immunotherapy in patients receiving combination chemotherapy and targeted agents. Results: We identified 99 patients (median age 62 years [range 34-79]; chemotherapy partners: cisplatin, carboplatin and paclitaxel). Fourteen patients (14%) received prior immunotherapy (PI) and 85 (86%) had no prior immunotherapy (NPI). Patient characteristics, including baseline full blood count, previous pelvic radiotherapy, sites of metastasis and serum albumin, were comparable between the 2 groups, apart from number of previous treatment lines, which was lower in the PI patients (median 1.5 vs 2, p ¼ 0.003). The odds of G4 neutropenia were higher in the PI group (OR ¼ 7.1, 95% CI ¼ 1.7-29.6, p ¼ 0.007). PI was associated with significantly increased odds of G3/4 thrombocytopenia (OR ¼ 14.4, 95% CI ¼ 2.7-77.4, p ¼ 0.002) on chemotherapy. In multivariate analysis, incorporating lines of prior chemotherapy (OR 1.3, 95% CI ¼ 1.0-1.5, p ¼ 0.037) and type of chemotherapy (carboplatin vs others: OR 2.3, 95% CI ¼ 0.9-6.2, p ¼ 0.094), the odds of developing G3/4 myelotoxicity were significantly higher in PI patients (OR 4.3, 95% CI: 1.3-14.4, p ¼ 0.02). Conclusions: In our small cohort, previous treatment with immunotherapy was associated with the development of G3/4 myelotoxicity, especially thrombocytopenia, on subsequent chemotherapy. These preliminary data require further prospective validation but may impact on decision making regarding optimal sequencing of systemic therapy. Legal entity responsible for the study: The Royal Marsden Hospital NHS Foundation Trust Funding: None Disclosure: J. de Bono: Consulting or advisory role: Astex, AstraZeneca, Genentech, Genmab, GSK, Merck, Pfizer, Sanofix Research Funding: AstraZeneca, Genentech, GSK, Sanofi, Janssen. U. Banerji: Receipt of grants/research supports: AstraZeneca, Chugai, Onyx, BTG Receipt of honoraria or consultation fees: Astex, Karus Therapeutics, Novartis, Vernalis. All other authors have declared no conflicts of interest.

1195P

Compromised efficacy of PD-L1 blockade therapy in axenic (germfree) mice with syngeneic tumors

A. Mansfield1, X. Liu2, S. Cao3, W. Bindeman4, P. Yin5, L. Till6, S. Harrington7, H. Dong8 Division of Medical Oncology, Mayo Clinic, Rochester, MN, USA, 2Urology and Immunology Research, Mayo Clinic, Rochester, MN, USA, 3Biochemistry and Molecular Biology, Mayo Clinic, Rochester, MN, USA, 4Research Services, Mayo Clinic, Rochester, MN, USA, 5Medical School, Campbell University School of Osteopathic Medicine, Lillington, USA, 6Research, Mayo Clinic, Rochester, MN, USA, 7Immunology Research, Mayo Clinic, Rochester, MN, USA, 8Immunology, Mayo Clinic, Rochester, MN, USA

1

Background: The microbiome can have profound effects on the innate immune system. Since the innate immune system regulates the adaptive immune response to antigens, we hypothesized that the microbiome may influence anti-tumor responses to immune checkpoint inhibitors. Accordingly, we sought to characterize the anti-tumor

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology effects of PD-L1 blockade therapy between mice with syngeneic tumors in conventional (specific pathogen-free, SPF) and germ-free (GF) environments. Methods: B16-OVA or Lewis Lung Cancer (LLC) cell lines were injected subcutaneously into the flanks of 10-12 week-old C57BL/6 mice in both SPF and germ-free (axenic) environments. Mice with B16-OVA tumors in SPF (n ¼ 6) and GF (n ¼ 12, 6 females and 6 males) environments, and mice with LLC tumors in GF (n ¼ 6) environments were randomized to receive the murine PD-L1 blocking antibody 10B5 or an isotype control. Tumor growth was evaluated every 2-3 days until days 35-40 when all mice were euthanized. Tumor size was compared between treatment groups in each environment at day 24 with the Mann Whitney U test. This project was approved by Mayo Clinic’s Institutional Review Board and Institutional Animal Care and Use Committee. Funding was provided by the NIH (K12 CA90628) and Mayo Clinic’s Center for Individualizing Medicine’s Microbiome Project. Results: Whereas injection of the anti-PD-L1 antibody (clone 10B5) controlled tumor growth compared to treatment with an isotype control in SPF female mice with B16OVA (p ¼ 0.05), PD-L1 blockade had no effect on tumor growth in female axenic mice with B16-OVA (p ¼ 0.20) or male axenic mice with B16-OVA (p ¼ 0.34) or axenic mice with LLC (p ¼ 0.56). Conclusions: PD-L1 blockade therapy loses its anti-tumor efficacy in axenic mice. The microbiome may influence the efficacy of PD-L1 blockade through of its effects on both innate and adaptive immune responses to tumors. Legal entity responsible for the study: Aaron Mansfield at Mayo Clinic Funding: National Institutes of Health; Mayo Clinic’s Center for Individualizing Medicine Microbiome Project Disclosure: All authors have declared no conflicts of interest.

antitumor activity. We investigated the immunomodulatory effect of regorafenib alone and in combination with a mouse-reactive anti PD1 antibody in mouse models of CRC. Methods: CT26 or MC38 syngeneic tumors were treated with regorafenib alone and in combination with anti PD1. We monitored tumor growth and analyzed the immune status of tumors ex vivo at the end of the study. Immune infiltrates were characterized by flow cytometry, intratumoral cytokines by multiplex ELISA, and expression of immunologically relevant genes by qPCR. Results: Both regorafenib and anti PD1 inhibited the growth of MC38 tumors vs control, and this effect was significantly enhanced by concomitant treatment or when regorafenib was given after anti PD1. Regorafenib treatment most consistently reduced tumor-infiltrating macrophages in both MC38 and CT26 tumors in a dose-dependent manner. Additionally, signs of M1-type macrophage conversion were detected by elevated inducible NO synthase and reduced arginase expression. This may be due to a regorafenib-mediated inhibition of the CSF1 receptor, as shown in vitro in the murine macrophage cell line RAW264.7. Anti PD1 treatment was associated with elevated interferon-g levels, indicative of enhanced T cell activation. Conclusions: These results warrant further exploration of a combination of regorafenib and PD1 for the treatment of colorectal cancer. Legal entity responsible for the study: Bayer AG Funding: Bayer AG Disclosure: S. Hoff, S. Gru¨newald, L. Ro¨se, D. Zopf: Employees of Bayer AG, and some are shareholders of Bayer AG stocks.

1199P 1197P

iRGD enhances T cells infiltration and augments response to PD-1 gene knockout immunotherapy in gastric cancer

N. Ding, S. Su, F. Meng, H. Sha, F. Chen, J. Wei, S. du, B. Liu The Comprehensive Cancer Centre of Drum Tower Hospital, Medical School of Nanjing University & Clinical Cancer Institute of Nanjing University, Nanjing, China Background: Poor infiltration of activated lymphocytes into tumors can be a fundamental factor limiting their efficacy and impeding the therapeutic effect of the checkpoint blockade immunotherapy. A tumor-penetrating peptide, iRGD, has a welldefined role in delivering drugs into extravascular tumor tissues in both the combination regimen and conjugated pattern, Here, we explored for the first time whether this cycled peptide could facilitate the infiltration of lymphocytes into tumor and furtherly overcome resistance to PD1 gene knockout immunotherapy. Methods: We used polyethylene glycol-conjugated phospholipid (PEG-lipid) derivatives, a time-efficient and versatile platform, to immobilize iRGD on T cell membrane. The ability of iRGD modified or co-applied lymphocytes infiltration was detected in both the 3D tumor spheroids in vitro and subcutaneous tumor model and peritoneal tumor model of gastric cancer in vivo. Furthermore, the synergistic effect of iRGD modification and PD-1 gene knockout in adoptive T cell transfer immunotherapy was examined in a xenograft model of EBV-associated gastric cancer. Results: In this study, we showed that T cells could be modified by the synthetic iRGDPEG-lipid without compromising their vitality, expansion, phenotype and effector function. In vitro, co-administration of iRGD could promote the infiltration of T cells while iRGD modification made T cells spread more extensively throughout the multicellular spheroids. Near infrared results showed that iRGD modification made a tenfold improvement infiltration of T cells into tumors without a parallel increase in normal tissues. Most importantly, we demonstrated that iRGD modified T cells had superior antitumor efficiency owing to sufficiently increased T cells infiltration, and exhibited robust synergistic effect with PD-1 gene knockout immunotherapy. Conclusions: Our study indicates that modification of T cell membrane with iRGD might be a potent strategy to increase T cells infiltration, thereby overcome the bottleneck of solid tumor immunotherapy. Legal entity responsible for the study: Baorui Liu Funding: None Disclosure: All authors have declared no conflicts of interest.

1198P

Immunomodulation by regorafenib alone and in combination with anti PD1 antibody on murine models of colorectal cancer

S. Hoff1, S. Gru¨newald2, L. Ro¨se1, D. Zopf2 Oncology III, Bayer AG, Berlin, Germany, 2Oncology II, Bayer AG, Berlin, Germany

1

Background: Regorafenib is a small molecule inhibitor of multiple kinases involved in normal cellular functions and in pathologic processes such as oncogenesis, angiogenesis, and tumor immunity. Regorafenib is approved for the treatment of advanced colorectal cancer (CRC) and gastrointestinal stromal tumors. In addition, an overall survival benefit has recently been shown in patients with hepatocellular carcinoma who had previously progressed on sorafenib (RESORCE trial). Immuno-oncology treatment strategies have recently expanded the arsenal of highly effective cancer therapies. In addition to their activity in monotherapy, they are being tested in combination with other therapies, including those inhibiting angiogenesis, to further improve their

Volume 28 | Supplement 5 | September 2017

Effect of MEK inhibition on PD-L1 and MCH-1 expression and on cytokines production profile in NSCLC cells and in human lymphocites

C.M. Della Corte1, G. Barra2, V. Ciaramella1, F. Papaccio1, G. Viscardi1, G. Esposito3, R. Di Liello1, M. Orditura4, T. Troiani5, R. De Palma2, F. Ciardiello4, F. Morgillo4 1 Oncology, Universit a degli Studi della Campania Luigi Vanvitelli, Naples, Italy, 2 Immunologia Clinica, AOU Seconda Universit a degli Studi di Napoli (AOU-SUN), a degli Studi di Napoli (AOU-SUN), Naples, Italy, 3Oncology, AOU Seconda Universit Naples, Italy, 4Dipartimento Medico-Chirurgico di Internistica Clinica e Sperimentale “F. Magrassi e A. Lanzara”, Universit a degli Studi della Campania “Luigi Vanvitelli”, Naples, Italy, 5Oncologia Medica, Dipartimento di Internistica Clinica e Sperimentale “F. Magrassi”, Universit a degli Studi della Campania Luigi Vanvitelli, Naples, Italy Background: Understanding of cancer-immune system interaction led to development of immunotherapy; anti-programmed cell death protein-1 (PD-1)/programmed cell death ligand-1 (PD-L1) antibodies are now used in non small cell lung cancer (NSCLC) treatment. MAPK cascade is a key intracellular network for tumor proliferation and recent data suggest that it is implicated in interplay of tumor and T-CD8þ cytotoxic lymphocytes (CTL). Methods: We evaluated PD-L1 mRNA level by Real Time qPCR (RT-qPCR) and its protein production, togheter with MAPK proteins, by western blot (WB), in NSCLC cell lines. Then, we studied the changes in PD-L1 and major histocompatibility complex class-I (MHC-I) expression and cytokines’ production, after MAPK-inhibition or -stimulation, by MEK-inhibitor, cobimetinib, or phorbol 12-myristate 13-acetate (PMA), respectively. In addition, we explored the effect of cobimetinib on cytokines’ genes by RT-qPCR on cDNA, obtained from retro-transcription of RNA extracted from T-lymphocytes, derived from Peripheral blood mononuclear cells (PBMC) of healthy volunteers, by density gradient separation, and activated with anti-CD3/antiCD28 coated beads. Results: WB and RT-qPCR for PD-L1 in NSCLC cells revealed a consistent correlation between mRNA and protein levels, togheter with activated MAPK and MEK1/2 signals, and suggested that ectopic PD-L1 mainly depends on trascriptional regulation. PDL-1 levels were significantly decreased by cobimetinib and increased by PMA, suggesting that MAPK can regulate PD-L1. Moreover, MEK-inhibition resulted on cancer cells in increased synthesis of MHC-I, IFN-gamma, IL-6, IL-1B, and TNFalpha, involved in CTL activation, and on activated human pheripheral T-lymphocytes in increment of mRNA levels of IL-12, TNFalpha and IFNgamma, that are pro-inflammatory cytokines typical of CTL subset, that seems more involved in immune response against cancer. Conclusions: These results demonstrate that MEK-inhibion induces the establishment of a pro-inflammatory microenvironment and may represent a potential mechanism to convert otherwise resistant cancers through treatment combination strategies of MEKinhibitors and anti-PD-L1/PD-1 antibodies in NSCLC. Legal entity responsible for the study: AOU Universita della Campania “Luigi Vanvitelli” Funding: AIRC Disclosure: All authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx376 | 423

abstracts 1200P

Exploring personalized immunotherapy opportunities in colorectal cancer

B. Navarro Rodrigo1, S. Bobisse1, S. Vigan o1, P. Baumgartner1, T. Nguyen-Ngoc1, P. Gannon1, R. Genolet1, B. Stevenson2, C. Sempoux3, M-O. Sauvain4, M. Hubner4, D. Hahnloser4, N. Demartines4, M. Montemurro1, L. Kandalaft1, S. Rusakiewicz1, A. Harari1, G. Coukos1 1 Department of Oncology, Ludwig Institute for Cancer Research, Lausanne, Switzerland, 2Vital-IT, SIB Swiss Institute of Bioinformatics, Lausanne, Switzerland, 3 Institute of Pathology, CHUV/UNIL, Lausanne, Switzerland, 4Department of Visceral Surgery, CHUV/UNIL, Lausanne, Switzerland Background: Adoptive T cell therapy (ACT) with tumor-infiltrating lymphocytes (TIL) can induce prolonged clinical responses in selected patients with gastrointestinal tumors while a significant fraction of patients do not respond. The association between immune profiles and antigenic specificity of TIL and clinical responses remains unclear. We addressed these issues, including the recognition of neoantigens, in order to explore the potential of personalized cell-based and vaccine therapy in colorectal cancer (CRC). Methods: Tumor specimens of primary (n ¼ 11) and metastatic (n ¼ 12) colon adenocarcinoma were dissected in fragments and cultured with IL-2 (6000U/ml) for 17-28 days. TIL were analyzed by multiparametric flow cytometric analyses and interrogated with private sets of predicted neoepitopes derived from non-synonymous mutations. T-cell responses against neoepitopes were detected by IFNc ELISpot and validated with peptide-MHC multimers. Results: TIL (i.e.>50x106 cells, mean6SEM 239652x106) were obtained from 7 and 8 patients with primary and metastatic colon adenocarcinoma, respectively. In primary tumors, the highest potential for TIL expansion was observed for microsatelliteinstable tumors as opposed to microsatellite-stable (MSS) tumors (mean6SEM 4356194x106 vs. 84634x106 cells; p ¼ 0.05, Mann-U). TIL yield was similar in primary and metastatic tumors, however in metastatic tumors the CD4/CD8 T cells ratio was higher (median 11 vs. 1; p ¼ 0.002, Mann-U) and inversely correlated with TIL expansion (rs -0.8; p ¼ 0.005). Most (>90%) T cells had a phenotype of effector-memory (CCR7-CD45RA-) activated (HLADRþPD1þTIM3þ) cells. Mutational load (ranging from 23 to 2760) and potential neoepitopes (from 25 to 2373) were determined and, of interest, initial screening experiments identified 2% of neoantigen specific-TIL (mutMALT1; V380A) in a representative MSS metastatic tumor harboring only 47 missense mutations. Conclusions: We demonstrate the spectrum of TIL expansion across CRC subtypes and stages, including the validation of neoepitopes in a non-hypermutated advanced tumor. These observations stress the potential of CRC patients for different strategies of personalized immunotherapy. Legal entity responsible for the study: Centre de The´rapies Expe´rimentales. De´partement d’oncologie. Centre Hospitalier Universitaire Vaudois Funding: Centre de The´rapies Expe´rimentales. De´partement d’oncologie. Centre Hospitalier Universitaire Vaudois. This Research Project was supported by ESMO with the aid of a grant from Amgen. Any views, opinions, findings, conclusions, or recommendations expressed in this material are those solely of the authors and do not necessarily reflect those of ESMO or Amgen Disclosure: All authors have declared no conflicts of interest.

1203P

Functional screening of B7H6-based chimeric antigen receptor (CAR) designs

B. Demoulin, L. Springuel, D. Daro, J. Bolse´e, J. Houssa, F. Huberty, C. Jacques-Hespel, C. Marchand, J. Marijsse, T.L.T. Nguyen, N. Ramelot, B. Violle, C. Lonez, D.E. Gilham, V. Steenwinckel Research & Development, Celyad SA, Mont Saint Guibert, Belgium Background: B7H6, a stress-induced ligand for the NK-activating receptor Nkp30, is widely expressed at the surface of transformed cells yet absent in healthy tissues. This makes B7H6 an attractive target for a CAR T-cell therapy with broad clinical applicability, including colon cancer and neuroblastoma. CARs are artificial receptors comprising an extracellular antigen-binding region (often a single chain variable fragment (scFv)) fused to an intracellular T-cell activation tail (usually CD3f in tandem with one or two costimulatory domain(s)). Here, we report the in vitro screening of various B7H6-based CAR designs differing by either the origin of their targeting moiety (murine versus humanized scFv), the costimulatory signaling module (either CD28 or 41BB as a 2nd generation CAR) or a combination of CD28 and 4-1BB in a 3rd generation CAR context. Methods: Primary human T-cell populations expressing the diverse B7H6-specific CAR constructs were compared for viability and fold expansion at the end of manufacturing as well as in vitro functionality (IFNg secretion and cytolytic activity when challenged with B7H6 expressing cell lines). Results: All B7H6-based CAR T-cells yielded comparable fold expansion with high viability suggesting that the CAR design has no impact on process parameters. CARs with targeting moiety of murine scFv origin were functionally superior to humanized versions in terms of killing and IFNg release potentially due to a difference in target affinity between the scFv. Second generation CARs containing CD28 endowed CAR T-cells possessed superior in vitro anti-tumor activity compared to all other constructs. Cryopreservation of these 2nd generation CAR T-cells did not significantly reduce viability and potency post-thawing.

424 | Immunotherapy of cancer

Annals of Oncology Conclusions: In these studies, a B7H6-based CAR comprised of murine scFv fused to CD28-CD3f signaling tail represented the best choice candidate after in vitro testing warranting further investigation. Subsequent studies will include in vivo xenograft models of colon cancer and neuroblastoma as well as target profiling through immunohistochemistry assessing B7H6 expression in a wide panel of tumor and normal tissues. This work focuses upon developing a package to support the clinical testing of B7H6 targeted CAR T-cell therapy. Legal entity responsible for the study: Celyad sa Funding: Celyad sa Disclosure: B. Demoulin, L. Springuel, D. Daro, J. Bolse´e, J. Houssa, F. Huberty, C. Jacques-Hespel, C. Marchand, J. Marijsse, T.L.T. Nguyen, N. Ramelot, B. Violle, C. Lonez, D.E. Gilham, V. Steenwinckel: Employee at Celyad.

1204P

Concurrent immuno-radiotherapy in lung and renal cancer- a new treatment paradigm

J. Ansari1, M. Ali1, A. Alomair2, A. Mohammed Ali1, E. Murshid1, A. Shaukat1, E. Batubara3, F. Bashir1, A. Farrag1, K. Hussein1, A. Alhamad1 1 Department of Oncology, Prince Sultan Military Medical City, Riyadh, Saudi Arabia, 2 Radiation Oncology, King Faisal Specialist Hospital and Research Centre, Riyadh, Saudi Arabia, 3Department of Pulmonology, Prince Sultan Military Medical City, Riyadh, Saudi Arabia Background: Concurrent administration of checkpoint inhibitors and radiotherapy (Immuno-RT) remains investigational and is the subject of multiple clinical trials. Nivolumab is an anti-programmed death-1 receptor monoclonal antibody that exhibits checkpoint-mediated immune response against tumor cells. Nivolumab has received regulatory approval for the second-line management of metastatic non-small cell lung cancer (NSCLC) & renal cell carcinoma (RCC). Ionising radiation could increase the diversity & quantity of tumoral antigen presentation, thereby augmenting anti-tumour immune response achieved with checkpoint inhibitors. The aim of this study was to assess the efficacy & toxicity of concurrent administration of nivolumab and radiotherapy. Methods: We identified 6 patients that received concurrent nivolumab and radiotherapy to 19 lesions; metastatic NSCLC (n ¼ 4), metastatic RCC (n ¼ 2). Treatment-related toxicities were identified by retrospective review of patient notes. Measurable lesions were assessed by RECIST 1.1 criteria. Pain score was used to assess symptomatic responses. Results: Stereotactic and conformal radiotherapy were delivered to 9 and 10 lesions, respectively. Treatment sites (number of lesions): lung (n ¼ 8), hip (n ¼ 3), brain (n ¼ 4), shoulder, scalp, ethmoid and adrenal. The gap between radiotherapy & nivolumab did not exceed 2 weeks for all patients. No grade 3-4 toxicities were observed. Two of the lung cancer patients developed grade 1 pneumonitis. Fractionation schedules included 48Gy/4 fractions (#), 40Gy/4#, 34Gy/4#, 22Gy/1#, 30Gy/10#, 25Gy/5#, 20Gy/4# and 20Gy/5#. Of the 14 measurable lesions, 86% had excellent response including complete response of 3 lesions. Symptomatic benefit was observed in 4 out of 6 treatment sites (66%). Conclusions: The of role of concurrent nivolumab & radiotherapy in patients with metastatic NSCLC and RCC has never been reported previously. In our study, concurrent administration of nivolumab and radiotherapy appears to be well tolerated with excellent radiological and symptomatic responses. Ongoing clinical trials may help determine the future role of Immuno-RT in the rapidly evolving treatment paradigm of metastatic NSCLC and RCC management. Legal entity responsible for the study: Jawaher Ansari Funding: None Disclosure: J. Ansari: Paid honoraria for lectures and/or advisory boards for Amjen, AstraZeneca, Pfizer, Novartis, Boehringer Ingleheim, Bristol-Myers Squibb, Roche and Sanofi. A. Shaukat: lecture fees and advisory board for Bristol-Myers Squibb. A. Alhamad: Advisory board for Bristol-Myers Squibb. All other authors have declared no conflicts of interest.

1205P

Optimum fractionation of radiation dose to combine anti-PD-1 mAb in MC38 mice model

F. Teng, X. Wang, J. Yu, L. Kong Radiation Oncology, Shandong Cancer Hospital, Jinan, China Background: The irradiated tumor cell death can enhance antitumor immunity by inducing antigen expression on tumor cells and activating lymphocytes. Radiotherapy (RT) combined with immunotherapy has revealed promising outcomes in various animal models. However, the optimum fractionation of radiation for priming immune response is controversial. This study aimed to explore the fractionation of radiation to maximize immunity in combinatorial treatment. Methods: Mice bearing MC38 murine colon cancer were treated with up to 24Gy radiation given in various sized fractions as 24Gy x 1f, 8Gy x 3f, 8Gy x 1f followed by 2Gy x 8f and 2Gy x 12f, and tumor growth followed. The immune response in the tumor, drainage lymph node(dLN) and spleen at 48h after radiation were assessed. 8Gy x 3f was chosen to combine anti-PD-1 immunotherapy. The abscopal effects and immune response were assessed by flow cytometry and immunohistochemistry(IHC).

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Results: Single dose of 24Gy and 8Gy x 3f brought best tumor control. No abscopal effects was observed after radiotherapy alone. Fractionation of 8Gy x 3f increased the irradiated tumor infiltrating lymphocytes(TILs). However, conventional 2Gy doses decreased CD4þTILs and CD8þTILs and increased myeloid myeloid-derived suppressor cell (MDSC) in spleen significantly. As the optimal fractionation to maximize immunity, 8Gy x 3f was chosen to combine anti-PD-1 mAb. Compared to radiotherapy or anti-PD-1 mAb alone, 8Gy x 3f combining with anti-PD-1 mAb brought obvious abscopal effect. CD8þT cells in the dLNs of the irradiated tumors were increased significantly in the combining group. Also, the combining treatment regimen increased CD4þT cells and CD8þT cells and decreased MDSC in the spleen. No serious toxicity of heart, liver, spleen, lung and kidney in each group was observed by using IHC. Conclusions: Hypofractionation of 8Gy x 3f was the fractionation of radiation dose to maximize immunity, compared to single dose of 24Gy and conventional 2Gy doses. Radiation with 8Gy x 3f combining with anti-PD-1 mAb had synergistic antitumor effect. Legal entity responsible for the study: Jinming Yu Funding: None Disclosure: All authors have declared no conflicts of interest.

1206P

Efficacy of tumor treating fields (TTFields) and anti-PD-1 in non-smallcell lung cancer (NSCLC) preclinical models

M. Giladi1, T. Voloshin1, O. Talyitzhaki2, U. Weinberg3, E.D. Kirson2 Preclinical, Novocure, Haifa, Israel, 2Research and Development, Novocure, Haifa, Israel, 3Clincal Development, Novocure, Luzern, Switzerland

1

Background: Tumor Treating Fields (TTFields) are an effective anti-neoplastic treatment modality delivered via non-invasive application of low intensity, intermediate frequency, alternating electric fields. TTFields is approved for the treatment of both newly diagnosed and recurrent glioblastoma. TTFields interrupt mitosis in cancer cells by disrupting microtubules and septin filaments, which play key roles in mitosis. The mitotic effects of TTFields include abnormal chromosome segregation that trigger different forms of cell death. We evaluated TTFields’ effect on immunogenic cell death and its efficacy when combined with an immune checkpoint inhibitor (aPD1) in NSCLC. Methods: Murine Lewis lung carcinoma (LLC) cells were treated with TTFields using the inovitroTM system. Levels of cell surface calreticulin (CRT) and intracellular ATP levels were evaluated using flow cytometry. High mobility group box 1 (HMGB1) secretion was measured using an ELISA assay. Mice inoculated with LLC cells were treated with isotype control, TTFields, aPD-1, or TTFields þ aPD-1. Tumor volume monitoring and intra-tumor immune cell profiling were performed. Results: TTFields induced elevated cell surface expression of CRT, decreased ATP levels, and promoted HMGB1 secretion. In vivo, the combined treatment of TTFields þ a-PD-1 led to a significant decrease in lung tumor volume compared to all three other groups (P < 0.001). Significant increase in CD45þ tumor infiltrating cells was observed in the TTFields þ aPD-1-treated mice. Infiltrating cells demonstrated a significant upregulation of surface PD-L1 expression. Both F4/80þCD11bþ cells and CD11cþ cells exhibited higher tumor infiltration and elevated PD-L1 expression, as compared to the control group. These findings indicate enhanced inflammatory antitumor environment conferred by the combination of TTFields þ aPD-1. Conclusions: Our results demonstrate that TTFields treatment potentiates immunogenic cell death in NSCLC cancer cells. Combining TTFields with specific immunotherapies such as anti-PD-1 may enhance antitumor immunity and result in increased tumor control. A phase III clinical study on TTFields in combination with either PD-1 inhibitors or docetaxel in NSCLC is underway. Legal entity responsible for the study: Novocure Funding: Novocure Disclosure: M. Giladi, T. Voloshin, O. Talyitzhaki, U. Weinberg, E.D. Kirson: Novocure employee.

1207TiP

An open-label, Phase IB study of NEO-PV-01 1 Adjuvant with Nivolumab in Patients with Melanoma, Non-Small Cell Lung Carcinoma, or Transitional Cell Carcinoma of the Bladder

P.A. Ott1, A. Naing2, G. Ramaswamy3, K. Margolin4, M. Moles5, R. Gaynor5, M.J. Goldstein5, S. Hu-Lieskovan6 1 Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA, 2Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA, 3Medical Oncology, Washington University School of Medicine, St. Louis, MS, USA, 4Medical Oncology, City of Hope, Duarte, CA, USA, 5R&D, Neon Therapeutics, Boston, MA, USA, 6Medical Oncology, UCLA - School of Medicine, Los Angeles, CA, USA Background: Cancer cells harbor DNA mutations that encode altered amino acid sequences known as neoantigens. Absent from normal tissues and highly specific for tumors, neoantigens bypass central tolerance and have been established as critical targets for tumor directed T cell responses. Tumor mutational burden and neoantigen load have been associated with anti-tumor activity of immune checkpoint inhibitors. Vaccines targeting neoantigens have the potential to induce de novo and expand

Volume 28 | Supplement 5 | September 2017

existing tumor directed T cell responses. NEO-PV-01 is a personalized neoantigen long peptide vaccine designed specifically for the molecular profile of an individual patient’s tumor. Trial design: NT-001 is a single-arm, phase IB study evaluating the safety of administering NEO-PV-01 þ adjuvant (Poly-ICLC) with the PD-1 directed antibody, nivolumab, in patients with advanced melanoma, smoking-associated non-small cell lung carcinoma, or transitional cell carcinoma of the bladder who have received no more than one prior systemic treatment. NEO-PV-01 is custom designed and generated for each patient by DNA and RNA sequencing of a recently biopsied tumor, HLA typing, selection of neoantigen epitopes, and synthesis of up to 20 peptides (14-35 amino acids in length). Patients receive treatment with nivolumab at a dose of 240 mg IV q2 weeks while their vaccine is produced. These peptides are formulated into four distinct pools, mixed with Poly-ICLC, and administered subcutaneously into up to 4 non-rotating anatomical sites. Beginning at Week 12, patients receive five priming immunizations over a three-week period followed by booster vaccinations at Weeks 19 and 23 while continuing nivolumab. The primary endpoint is safety. Secondary endpoints are ORR, CBR, PFS, and assessment of response conversion between Week 12 and Week 24. Exploratory endpoints include extensive immune monitoring. The study is open as of October 2016 with estimated enrollment of 90 patients. Clinical trial identification: NCT02897765 Legal entity responsible for the study: Neon Therapeutics, Inc. Funding: Neon Therapeutics, Inc. Disclosure: P.A. Ott: Reports grants and personal fees from Bristol-Myers Squibb, CytomX, Celldex, and Merck, personal fees from Neon Therapeutics, Amgen, Novartis, Pfizer, and Roche/Genentech, and grants from AstraZeneca/MedImmune, outside of the submitted work. G. Ramaswamy: Reports advisory board or consulting work for Merck, Genentech, Baxalta, Roche, Boehringer Ingelheim, Novartis, Ariad, Astellas, Bristol-Myers Squibb, Pfizer, Celgene, AstraZeneca, INC Research, AbbVie, and MSKCC. M. Moles, R. Gaynor, M.J. Goldstein: Employee of Neon Therapeutics, Inc. S. Hu-Lieskovan: Consulting: Amgen, Merck, Novartis, Vaccinex, Emergent BioSolutions. Contracted Research: Pfizer, Plexxikon, Genentech, Neon Therapeutics. Research Support: Bristol-Myers Squibb, Merck. Travel: Amgen, Merck, Novartis, Vaccinex, Emergent BioSolutions, Neon Therapeutics. All other authors have declared no conflicts of interest.

1208TiP

A phase II exploratory study of durvalumab (MEDI4736) in HIV-1 patients with advanced solid tumors

M. Gonzalez-Cao1, J. Martinez-Picado2, M. Provencio Pulla3, B. Clotet2, O. Juan4, J. Dalmau2, T. Moran5, A. Mayerhans6, J. De Castro7, J. Blanco2, R. Blanco8, R. Bernabe Caro9, N. Karachaliou10, J. Garcia-Corbacho11, M.A. Molina1, C. Brander2, R. Rosell12 1 Oncology Department, Quir on Dexeus University Hospital, Barcelona, Spain, 2AIDS Research Institute, Hospital Universitari Germans Trias i Pujol, Badalona, Spain, 3 Medical Oncology, Hospital Puerta de Hierro Majadahonda, Majadahonda, Spain, 4 Medical Oncology, Hospital Universitari i Polite`cnic La Fe, Valencia, Spain, 5Medical Oncology, Hospital Universitari Germans Trias i Pujol, Badalona, Spain, 6ICREA, Universidad Pompeu Fabra, Barcelona, Spain, 7Medical Oncology, Hospital la Paz, Madrid, Spain, 8Medical Oncology, Mutua Terrassa, Barcelona, Spain, 9Medical Oncology, Hospital Universitario Virgen del Rocio, Seville, Spain, 10Medical Oncology, Hospital Sagrado Corazon, Barcelona, Spain, 11Oncology Department, Hospital Clinic y Provincial de Barcelona, Barcelona, Spain, 12Medical Oncology, Catalan Institute of Oncology (ICO Badalona), Hospital Germans Trias i Pujol, Badalona, Spain Background: Although acquired immunodeficiency syndrome-related mortality is decreasing with the introduction of effective antiretroviral therapy, it has been reported a significant increase in the proportion of non-acquired immunodeficiency syndrome defining malignancies in HIV infected patients, often associated with premature immunosenescence and exhaustion. It has been shown in murine models and humans that programmed cell death ligand 1 (PD-L1) and its receptor, programmed cell death 1 (PD-1) play an active and reversible role mediating T-cell exhaustion both in cancer and in chronic infections. Binding PD-1 to its ligand PD-L1negatively regulates T-cell response, leading to an exhausted phenotype on CD8þ T cells. Therefore, there is a potential of immunotherapeutic intervention targeting PD-1/PD-L1 in order to enhance anti tumoral immune responses as well as to facilitate viral eradication. Durvalumab (MEDI4736) is a human monoclonal antibody (MAb) of the immunoglobulin G1 kappa (IgG1j) subclass that inhibits binding of programmed cell death ligand 1 (PDL1) to programmed cell death 1 (PD-1; CD279) and CD80 (B7-1). Durvalumab has demonstrated in cancer patients a favorable safety profile with encouraging antitumor activity, but there are no data about tolerance or anti retroviral activity in HIV patients. Trial design: This is an ongoing multicenter, open-label, phase 2 study (EUDRACT: 2016-004524-38) whose primary objective is to assess the feasibility of durvalumab at the recommended dose of 1500 mg every 4 weeks in HIV-infected patients with solid tumors for which no additional oncologic standard treatment is available. As secondary objectives the response rate (RECIST 1.1 and irRECIST), duration of response, PFS and OS will be measured. Exploratory objectives include the assessment of antiviral activity by measuring the changes in the HIV viral reservoir, the residual viral replication and the composition and function of circulating T lymphocytes and the study of molecular predictive factors of antitumoral activity on pretreatment tumor samples. Clinical trial identification: EUDRACT: 2016-004524-38 Legal entity responsible for the study: Spanish Lung Cancer Group

doi:10.1093/annonc/mdx376 | 425

abstracts Funding: AZ Spain Disclosure: All authors have declared no conflicts of interest. 1209TiP

A first-in-human, open-label, multicenter phase 1/2a study to evaluate the safety and efficacy of increased repeated doses of the first in class RORc agonist LYC-55716 in treating locally advanced or metastatic solid tumors

H.J. Wilkins1, E.P. Hamilton2, D. Mahalingam3 Clinical Development, Lycera Corp., Plymouth Meeting, PA, USA, 2Breast Cancer and Gynecologic Cancer Research Program, Sarah Cannon Research Institute/Tennessee Oncology, Nashville, TN, USA, 3Medical Oncology, University of Texas Health Science Center San Antonio, San Antonio, TX, USA

1

Background: LYC-55716, a first-in-class oral, small-molecule agonist of nuclear receptor retinoic acid–related orphan receptor c (RORc), has been shown in preclinical models to modulate gene expression to reprogram immune cell antitumor effector function and decrease immunosuppressive mechanisms, resulting in immune-mediated tumor growth control and enhanced survival. Data suggest that LYC-55716 acts via well-known antitumor mechanisms by increasing immune cell trafficking and recruitment to tumors, enhancing T cell effector function and memory development, and promoting T cell survival. LYC-55716 may enhance immune-mediated antitumor responses via its effects on T effector/Treg cell ratios, PD-1 expression, and sensitivity to PD-L1 inhibition of T cell proliferation. A first-in-human, single-arm, open-label multicenter Phase 1/2a study is ongoing to evaluate the safety and tolerability of LYC-55716 and determine the maximum tolerated dose and objective response rate. All adult subjects enrolled will have relapsed or refractory metastatic cancer and have failed to responded to standard therapies. Trial design: The Phase 1 portion of the study will follow a dose-escalation design to evaluate the occurrence of dose-limiting toxicities (DLTs) and determine the maximum tolerated dose and recommended Phase 2 dose of LYC-55716. Following a screening period, adults with locally advanced or metastatic solid tumors will receive 28-day treatment cycles of LYC-55716 BID (n ¼ 4–6/cohort). Dosing escalation considers dose and dosing regimen and is determined by PK profile and safety. Primary endpoints include safety (monitoring of adverse events, physical examination, and lab results) and incidence of DLTs (Grade 3 or 4 toxicities) during the first 28-day treatment cycle. Secondary endpoints include objective tumor response rate as assessed via response evaluation criteria in solid tumors (RECIST) v1.1 assessed at scans performed every 8 weeks, pharmacokinetics, and pharmacodynamics. Results for the first three cohorts of the Phase 1 portion of the study will be available at the time of presentation. Clinical trial identification: NCT02929862 Legal entity responsible for the study: Lycera Corp. Funding: Lycera Corp. Disclosure: H.J. Wilkins: Employee and shareholder of Lycera Corp. All other authors have declared no conflicts of interest.

1210TiP

A phase I global trial targeting multiple solid and hematologic malignancies through a NKG2D receptor-based CAR-T immunotherapy

B. Verma1, A. Awada2, P. Aftimos2, P. Lewalle3, N. Meuleman3, J-P. Machiels4, G. Catala5, E. Vandenneste6, J. Brayer7, D. Sallman7, S. Sahebjam8, T. Kerre9, S. Rottey10, K. Odunsi11, E.S. Wang12, U. Santanam13, C. Lonez14, D.E. Gilham14, F. Lehmann14 1 Immuno-Oncology, Celyad, Boston, MA, USA, 2Oncology, Institute Jules Bordet, Brussels, Belgium, 3Hematology, Institute Jules Bordet, Brussels, Belgium, 4Oncology, Cliniques Universitaires Saint-Luc, Brussels, Belgium, 5Oncology, Cliniques Universitaires Saint-Luc, Woluwe-Saint-Lambert, Belgium, 6Hematology, Cliniques Universitaires Saint-Luc, Brussels, Belgium, 7Hematology, H. Lee Mofﬁtt Cancer Center University of South Florida, Tampa, FL, USA, 8Oncology, H. Lee Mofﬁtt Cancer Center University of South Florida, Tampa, FL, USA, 9Hematology, Ghent University Hospital, Ghent, Belgium, 10Oncology, Ghent University Hospital, Ghent, Belgium, 11Oncology, Roswell Park Cancer Institute, Buffalo, NY, USA, 12Hematology, Roswell Park Cancer Institute, Buffalo, NY, USA, 13Immuno-Oncology, Celyad, Boston, MA, USA, 14Immuno-Oncology, Celyad, Mont-Saint-Guibert, Belgium Background: Because of its increasingly demonstrative successes, CAR-T therapy has been well recognized as one of the most promising therapies for cancer. We have developed a novel autologous CAR-T, NKR-2, incorporating the full-length human natural killer receptor NKG2D fused with the human CD3 zeta signaling domain. When expressed in T-cells, the naturally-expressed DAP10 provides the co-stimulatory signals to NKR-2 to be fully activated. NKR-2 selectively target tumor cells upon recognition of up to eight different NKG2D ligands expressed in many distinct cancer indications. In preclinical studies, NKR-2 demonstrated long-term anti-tumor activity towards

426 | Immunotherapy of cancer

Annals of Oncology multiple solid and hematologic tumors deploying multiple mechanisms of action targeting tumor cells and cells from the neo-vasculature and tumor suppressive immune environment, resulting in an adaptive response. In our recently completed Phase 1 study in hematologic cancers, a single administration of autologous NKR-2 was safe with initial signs of clinical benefit. Likewise, to overcome the operational challenges, our trial design incorporates strategies to harmonize multiple clinical and manufacturing processes while also enhancing patient safety and clinical outcomes. Trial design: THINK trial (THerapeutic Immunotherapy with NKR-2) is a EU/US openlabel Phase I study to assess the safety and clinical activity of NKR-2 therapy administered in three infusions, two weeks apart in five solid tumor indications (CRC, urothelial, TNBC, pancreatic, ovarian) and two hematologic indications (AML/MDS and MM). No lymphodepleting conditioning is required in this study. The study contains two consecutive segments. The dose escalation segment will enroll 18 patients in two separate hematologic and solid malignancy arms, and evaluate 3 dose levels of NKR-2 (3x108, 1x109 and 3x109 cells per injection) following a 3 þ 3 design. The expansion segment will then enroll 96 additional patients in 7 separate cohorts for each indication with 3 steps of statistical analysis (overall futility, futility within each cohort and final evaluation). At time of submission, the trial has completed enrollment in its first cohort among solid indications. Clinical trial identification: FDA: CYAD-N2T-002 Legal entity responsible for the study: CELYAD Funding: CELYAD Disclosure: B. Verma, U. Santanam, C. Lonez, D.E. Gilham, F. Lehmann: Employment with a pharmaceutical company: Employee of Celyad. A. Awada, P. Aftimos, P. Lewalle, N. Meuleman, J-P. Machiels, G. Catala, E. Vandenneste, J. Brayer, D. Sallman, S. Sahebjam, T. Kerre, S. Rottey, K. Odunsi, E.S. Wang: Corporate-sponsored research: Institute has received research finding from Celyad.

1211TiP

FAK-PD1: a phase I/IIa trial of FAK (defactinib) & PD-1 (pembrolizumab) inhibition

S. Symeonides1, T.R.J. Evans2, V. Coyle3, A. Serrels4, F. Thomson5, D. Currie6, S. Dillon6, J. Paul6, D.A. Fennell7, C. Ottensmeier8 1 Experimental Cancer Medicine Centre, University of Edinburgh, Edinburgh, UK, 2 Beatson West of Scotland Cancer Centre, University of Glasgow, Glasgow, UK, 3Medical Oncology, Northern Ireland Cancer Centre, Belfast City Hospital, Belfast, UK, 4MRC Centre for Inﬂammation Research, University of Edinburgh, Edinburgh, UK, 5 Translational Pharmacology Laboratory, Institute of Cancer Sciences, University of Glasgow, Glasgow, UK, 6Cancer Research UK Clinical Trials Unit, University of Glasgow, Glasgow, UK, 7Department of Cancer Studies, University Hospitals of Leicester NHS Trust Leicester Royal Inﬁrmary, Leicester, UK, 8Cancer Research UK Centre, Southampton General Hospital, Southampton University Hospitals NHS Trust, Southampton, UK Background: Focal Adhesion Kinase (FAK) is a pivotal intracellular mediator of extracellular contact interactions. It is over-expressed in cancer, with a long-established role in migration, invasion & survival, and is associated with poor prognosis. Recently FAK has been found to have a similar activity in recruitment of immunosuppressive cells to the tumour. We have shown that FAK inhibition can re-model the tumour immune microenvironment in vivo, shifting the balance from inhibitory Tregs, macrophages, fibroblasts and myeloid progenitors, to one which supports an active CD8þ adaptive immune response, resulting in tumour clearance and lasting immunity. FAK inhibition synergises with Programmed cell death receptor 1 (PD-1) blockade in more resistant models. Defactinib (VS-6063, Verastem) is a small molecule FAK inhibitor in Phase II development with an encouraging safety profile and biological activity. Pembrolizumab (MK-3475, MSD) is a humanized IgG4/kappa monoclonal antibody to PD-1, licensed for the treatment of an increasing number of tumour types. This recently open trial will assess the safety, tolerability and preliminary activity of defactinib plus pembrolizumab in patients with advanced solid malignancies. Trial design: FAK-PD1 is an open label, phase I/IIa clinical trial, combining 200 mg pembrolizumab as a 3-weekly IV infusion, with defactinib given orally twice daily at either 200 mg or 400 mg, before leading into three tumour-specific expansions (nonsmall cell lung cancer, mesothelioma and pancreatic cancer) at the selected dose. Up to 60 patients, PS 0-1, with adequate blood parameters, measurable disease, baseline tissue, and without contraindications to either agent, will be treated for up to 2 years until clear clinical progression, unacceptable toxicity, or withdrawal. Primary endpoint is safety (NCI-CTCAE v4.03); secondary endpoints include objective response rate (irRECIST), progression-free survival, FAK Y397 phosphorylation and immune cell infiltrate effects. Exploratory endpoints include comprehensive cellular and molecular characterisation of baseline and on-treatment tumour samples, and serial blood immune cell and cytokine profiling. Positive data will support further development of the combination. Clinical trial identification: FAK-PD1 EudraCT number: 2015-003928-31 Legal entity responsible for the study: University of Glasgow & NHS Greater Glasgow and Clyde

Volume 28 | Supplement 5 | September 2017

Annals of Oncology Funding: Cancer Research UK, Verastem Inc, and Merck Sharp and Dohme Ltd •Verastem Inc (via the Combinations Alliance program) and Cancer Research UK Disclosure: All authors have declared no conflicts of interest.

1212TiP

PIVOT-02: A phase 1/2, open-label, multicenter, dose escalation and dose expansion study of NKTR-214 and nivolumab in patients with select, locally advanced or metastatic solid tumor malignancies

A. Diab1, M.E. Hurwitz2, N. Tannir3, C. Bernatchez1, C. Haymaker1, S.E. Bentebibel1, B.D. Curti4, M.K.K. Wong1, I. Gergel5, M. Tagliaferri6, J. Zalevsky7, U. Hoch8, S. Aung9, M. Imperiale9, D. Cho10, S.S. Tykodi11, I. Puzanov12, H. Kluger13, P. Hwu1, M. Sznol14 1 Department of Melanoma Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA, 2Department of Internal Medicine, Division of Medical Oncology, Yale School of Medicine, New Haven, CT, USA, 3Genitourinary Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA, 4Department of Genitourinary Oncology Research, Providence Cancer Center and Earle A. Chiles Research Institute, Portland, OR, USA, 5 Department of Drug Development, Nektar Therapeutics, San Francisco, CA, USA, 6 Department of Clinical Development, Nektar Therapeutics, San Francisco, CA, USA, 7 Department of Biology & Preclinical Development, Nektar Therapeutics, San Francisco, CA, USA, 8Department of Clinical Pharmacology, Nektar Therapeutics, San Francisco, CA, USA, 9Department of Clinical Development, Nektar Therapeutics, San Francisco, CA, USA, 10Department of Medical Oncology, NYU Medical Oncology Associates, New York, NY, USA, 11Department of Medical Oncology, University of Washington and Fred Hutchinson Cancer Research Center, Seattle, WA, USA, 12Department of Medical Oncology, Roswell Park Cancer Institute, Buffalo, NY, USA, 13Department of Medical Oncology, Yale Breast Center, New Haven, CT, USA, 14Department of Medical Oncology, Yale University School of Medicine Medical Oncology, New Haven, CT, USA Background: Abundance and functional quality of tumor infiltrating lymphocytes are positively linked with tumor response and improved survival with checkpoint inhibitors. NKTR-214 is a CD122-biased agonist that targets the IL2 pathway and is designed to provide sustained signaling through the heterodimeric IL2 receptor pathway (IL2RbÇ) to preferentially activate and expand NK and effector CD8þ T cells over CD4þ T regulatory cells within the tumor microenvironment. NKTR-214 has been administered to 28 patients with advanced cancers. NKTR-214 as a single agent demonstrated a substantial increase in both CD8þ T and NK cells within the tumor microenvironment in patients with prior immune checkpoint therapy (Bernatchez et al 2016). Given the favorable safety profile and strong biomarker data, a trial combining NKTR-214 and nivolumab was initiated. Trial design: PIVOT-02 is a phase 1/2 open-label trial in patients (pts) with locally advanced or metastatic melanoma (mM), non-small cell lung cancer (NSCLC), renal cell carcinoma (RCC), urothelial carcinoma, or triple-negative breast cancer (TNBC). The primary objectives are to evaluate safety and tolerability, determine the recommended phase 2 dose (RP2D), and assess tumor response by RECIST 1.1. In an outpatient setting, NKTR-214 is administered at dose levels of 0.003, 0.006 and 0.009 mg/kg

Volume 28 | Supplement 5 | September 2017

abstracts in combination with nivolumab at two flat dose schedules of either 240 mg @ q2w or 360 mg @ q3w. As of May 8, 17 pts (7 mM, 8 RCC, and 2 NSCLC) have been enrolled into 4 cohorts in the dose-escalation phase. In the dose-expansion phase, approximately 250 pts will be enrolled in five tumor types and eight indications; immunotherapy naı¨ve patients and patients who are relapsed/refractory to checkpoint therapy are being studied separately. Extensive blood and tumor tissue samples are being collected to measure immune activation using immunophenotyping including flow cytometry, immunohistochemistry (IHC), T cell clonality and gene expression analyses. Enrollment is ongoing. Clinical trial identification: NCT02983045 Legal entity responsible for the study: Nektar Therapeutics Funding: Nektar Therapeutics Disclosure: A. Diab: Consulting or Advisory Role - Celgene; CureVac; Nektar Research Funding - Celgene (Inst); Idera (Inst); Nektar (Inst); Pfizer (Inst) Travel, Accommodations, Expenses – Nektar. M.E. Hurwitz: Employment - Pfizer Consulting or Advisory Role – Nektar. N. Tannir: Honoraria - Bristol-Myers Squibb; Exelixis; GSK; Nektar; Novartis; Pfizer Advisory Role - Bristol-Myers Squibb; Exelixis; GSK; Nektar; Novartis Research Funding - Bristol-Myers Squibb; Epizyme; Exelixis; Novartis Travel - Bristol-Myers Squibb; Exelixis; GSK; Nektar; Novartis; Pfizer. C. Bernatchez: Employment - Lexicon (I) Stock - Lexicon (I) Advisory Role - Lion Biotechnologies Research Funding - Idera; Nektar Patents - Patent pending on BTLA as a marker for better CD8 T cells for adoptive immunotherapy. C. Haymaker: Cara L. Haymaker Research Funding - Idera; Nektar. B.D. Curti: Honoraria - Prometheus Speakers’ Bureau - Prometheus Research Funding - Bristol-Myers Squibb; Galectin Therapeutics; MedImmune; Prometheus; Viralytics Travel, Accommodations, Expenses - Agonox; MedImmune; Nektar; Prometheus. I. Gergel: Employment - Nektar Leadership Corium International; Nektar Stock and Other Ownership Interests - Corium International; Nektar. M. Tagliaferri: Employment - Nektar Travel, Accommodations, Expenses - Nektar J. Zalevsky: Employment - Nektar. U. Hoch, S. Aung, M. Imperiale: Employment - Nektar Stock and Other Ownership Interests - Nektar D. Cho: Honoraria - Bristol-Myers Squibb; Exelixis; Roche/Genentech Consulting or Advisory Role - Pfizer; Prometheus. S.S. Tykodi: Consulting or Advisory Role - Amgen; Prometheus Research Funding - Argos Therapeutics (Inst); Bristol-Myers Squibb (Inst); Exelixis (Inst); Genentech (Inst); GlaxoSmithKline (Inst); Prometheus (Inst). I. Puzanov: Consulting or Advisory Role - Amgen; Bristol-Myers Squib; Roche/ Genentech. H. Kluger: Honoraria - Merck Consulting or Advisory Role - Alexion Pharmaceuticals; Prometheus; Regeneron Research Funding - Merck (Inst) Travel, Accommodations, Expenses - Bristol-Myers Squib P. Hwu: Stock and Other Ownership Interests - immatics; Lion Biotechnologies Consulting or Advisory Role Lion Biotechnologies Research Funding - Bristol-Myers Squibb (Inst); Genentech (Inst). M. Sznol: Stock -Adaptive Bio; Amphivena; Intensity Thera Advisor Adaptimmune; Alexion; Amgen; AstraZeneca; Biodesix; Bristol-Myers Squibb; Genentech; Immune Design; Janssen; Kyowa; Lilly; Lion Bios; Lycera; MSD; Merus; Modulate; Nektar; Novartis; Pfizer; Symphogen; Theravance. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx376 | 427

MELANOMA AND OTHER SKIN TUMOURS 1213O

Characterization of complete responses (CRs) in patients with advanced melanoma (MEL) who received the combination of nivolumab (NIVO) and ipilimumab (IPI), NIVO or IPI alone

C. Robert1, J. Larkin2, P.A. Ascierto3, G.V. Long4, J.C. Hassel5, D. Schadendorf6, F.S. Hodi7, C. Lebbe´8, J-J. Grob9, K. Grossmann10, J. Wagstaff11, J. Chesney12, D. Hogg13, O. Bechter14, I. Marquez-Rodas15, A.C. Pavlick16, D. Walker17, R. Bhore17, M.A. Postow18, J.D. Wolchok18 1 Department of Medicine, Gustave Roussy and INSERM Unite´ 981, Villejuif–Paris Sud, France, 2Medicine, Royal Marsden Hospital, London, UK, 3Melanoma, Cancer Immunotherapy and Innovative Therapy Unit, Istituto Nazionale Tumori Fondazione Pascale, Naples, Italy, 4Melanoma Medical Oncology, Melanoma Institute Australia, The University of Sydney and Royal North Shore and Mater Hospitals, Sydney, Australia, 5 Department of Dermatology, University Hospital Heidelberg, Heidelberg, Germany, 6 Department of Dermatology, University of Essen, Essen, Germany, 7Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA, 8AP-HP Dermatology CIC Departments, Saint-Louis Hospital, INSERM U976, Universite´ Paris Diderot, Paris, France, 9 Dermatology, Hospital de la Timone, Marseille, France, 10Oncology, Huntsman Cancer Institute, Salt Lake City, UT, USA, 11Medical Oncology, Singleton Hospital, South West Wales Cancer Institute & Swansea University College of Medicine, Swansea, UK, 12 Department of Medicine, University of Louisville, Louisville, KY, USA, 13Cancer Clinical Research Unit (CCRU), Princess Margaret Cancer Centre, Toronto, ON, Canada, 14 Oncology, UZ Leuven, Leuven, Belgium, 15Medical Oncology, Hospital Gregorio Maranon, Madrid, Spain, 16Department of Medicine, New York University, New York, NY, USA, 17Oncology, Bristol-Myers Squibb, Princeton, NJ, USA, 18Oncology, MemorialSloan Kettering Cancer Center and Weill Cornell Medical College, New York, NY, USA Background: In clinical studies, 10-15% of patients (pts) treated with anti-PD-1 monotherapies achieved durable CRs. Combination treatment with NIVOþIPI resulted in higher response rates, longer progression-free survival (PFS), and improved overall survival (OS) vs IPI alone, but with an increased frequency of adverse events (AEs). Here, we characterized CRs among pts who received combination therapy vs NIVO or IPI alone. Methods: In this post hoc analysis, efficacy and safety data were pooled for NIVOþIPI (N ¼ 409), NIVO (N ¼ 526), and IPI (N ¼ 362) from the phase 2 CheckMate 069, phase 3 CheckMate 066, and phase 3 CheckMate 067 studies in pts with MEL. Across studies, the minimum duration of follow-up was 24 months (median 31 months). Results: In the pooled analysis, the CR rate was 18% for NIVOþIPI, 16% for NIVO, and 4% for IPI, with partial responses (PRs) in 41%, 28%, and 14% of pts, respectively (Table). Among the 75 CR pts in the NIVOþIPI cohort, the majority (77%) are off treatment and 8% received a subsequent systemic therapy; 15% had elevated LDH levels and 32% had M1c disease. Median duration of CR has not been reached, with 63/75 pts (84%) remaining in response. After an additional follow-up of 12 months (from the 1-year initial follow-up), 24/166 pts (14%) with a PR converted to a CR. For the 75 CR pts in the NIVOþIPI cohort, 2-year PFS and OS rates were 86% and 92%, respectively. Treatment-related AEs of grade 3-4 occurred in 60% of NIVOþIPI-treated pts with a CR, 65% with a PR, and in 60% with stable disease; 31%, 36%, and 35%, respectively, led to discontinuation. There were no treatment-related deaths.

Annals of Oncology 28 (Supplement 5): v428–v448, 2017 doi:10.1093/annonc/mdx377

Legal entity responsible for the study: Bristol-Myers Squibb Funding: Bristol-Myers Squibb Disclosure: C. Robert: Served as a consultant for Amgen, Bristol-Myers Squibb, Merck, and Roche; paid honoraria from Amgen, Bristol-Myers Squibb, GSK, Merck, Novartis, and Roche. J. Larkin: Received research funding from Bristol-Myers Squibb, MSD, Novartis, and Pfizer; travel funding from Bristol-Myers Squibb, GSK, MSD, Esai, Pfizer, and Roche. P.A. Ascierto: Served as a consultant for Amgen, Array, Bristol-Myers Squibb, Merck-Serono, MSD, Novartis, Pierre-Fabre, and Roche-Genentech; institution received research funding from Array, Bristol-Myers Squibb, and Roche-Genentech. G.V. Long: Served as a consultant for Amgen, Bristol-Myers Squibb, Merck, MSD, Novartis, PierreFabre, and Roche; paid honoraria from Bristol-Myers Squibb, Merck, MSD, and Roche. J.C. Hassel: Funding for trial procedures according to study protocol from Bristol-Myers Squibb; honoraria from Amgen, Bristol-Myers Squibb, GSK, MSD, Novartis, and Roche; research grant/funding from Bristol-Myers Squibb; travel funding from Amgen, BristolMyers Squibb, GSK, MSD, Novartis, and Roche. D. Schadendorf: Served as a consultant or advisor for Roche/Genentech, Novartis, Bristol-Myers Squibb, Merck Sharp & Dohme, Merck Serono, Sysmex, Amgen, Grunenthal Group, Immunocore; participated on a speakers’ bureau for Roche, Bristol-Myers Squibb, Merck Sharp & Dohme, Novartis, Amgen, Incyte, Pierre Fabre; travel funding from Roche/Genentech, Bristol-Myers Squibb, Amgen, Merck, Merck Serono, Novartis; paid honoraria from Roche/Genentech, Novartis, Amgen, Bristol-Myers Squibb, Merck Sharp & Dohme, Sysmex, Immunocore, Grunenthal Group, Merck Serono, Agenus, Array BioPharma, LEO Pharma, Incyte, Pfizer, Pierre Fabre, Philogen, Regeneron; received institutional research funding from Bristol-Myers Squibb and Novartis. F.S. Hodi: Consultant: Amgen, EMD Serono, MSD, Novartis, Roche-GNE; travel funding: Bristol-Myers Squibb and Novartis; patent pending royalties per institutional policy; other: B Bristol-Myers Squibb MS; institutional research funding: Bristol-Myers Squibb, MSD, Novartis, Roche-GNE. C. Lebbe´: Served on an advisory board for Bristol-Myers Squibb, GSK, MSD, Novartis, and Roche. J-J. Grob: Served as a consultant to Amgen, Bristol-Myers Squibb, GSK, Merck, Novartis, and Roche; participated on speakers’ bureau for Bristol-Myers Squibb, GSK, and Roche; travel funding from Roche; recipient of research funding from Bristol-Myers Squibb and Roche. J. Wagstaff: Honoraria from Astellas, Bristol-Myers Squibb, Merck, Novartis, Pfizer, and Roche; consultant to Astellas, Bristol-Myers Squibb, Merck, Novartis, Pfizer, and Roche; served on speakers’ bureaus for Astellas, Bristol-Myers Squibb, and Novartis; travel funding from Astellas, Bristol-Myers Squibb, and Novartis. J. Chesney: Served as a consultant to Bristol-Myers Squibb; received research funding from Bristol-Myers Squibb. D. Hogg: Served as a consultant to Bristol-Myers Squibb, GSK, Novartis, and Roche. O. Bechter: Served as a consultant or advisor to Bristol-Myers Squibb; and received travel funding from Roche. I. Marquez-Rodas: Honoraria from Bristol-Myers Squibb, MSD, Novartis, and Roche; served as a consultant to Amgen, Bioncotech, Bristol-Myers Squibb, MSD, Novartis, and Roche; travel funding from Amgen, Bristol-Myers Squibb, and MSD. D. Walker: Employee of Bristol-Myers Squibb; immediate family member has stock or other ownership in Antares Pharma. R. Bhore: Employee of and owns stock in Bristol-Myers Squibb. M.A. Postow: Served on an advisory board for Bristol-Myers Squibb; recipient of research grant support from Bristol-Myers Squib. J.D. Wolchok: Consultant: BristolMyers Squibb, GSK, Jounce, MedImmune, Merck, Polaris, Polynoma, and Ziopharm; research funding: Bristol-Myers Squibb, GSK, MedImmune, and Merck; patent issued for DNA vaccine of cancer in companion animals (co-investor). All other authors have declared no conflicts of interest.

Table: 1213O NIVOþIPI (N ¼ 409) Objective response rate (%) CR, n (%) PR, n (%) Pts remaining in response (CR) CR pts continuing on treatment CR pts not continuing on treatment

NIVO (N ¼ 526)

IPI (N ¼ 362)

58.9

43.9

18.0

75 (18) 166 (41) 63/75 (84%)

83 (16) 148 (28) 75/83 (90%)

14 (4) 51 (14) 11/14 (79%)

17/75 (23%)

41/83 (49%)

4/14 (29%)

58/75 (77%)

42/83 (51%)

10/14 (71%)

Conclusions: MEL pts treated with NIVOþIPI had a high rate of durable CRs, with the majority remaining in response and often not requiring additional treatment at a median follow-up of 31 months. Some pts with a PR convert to a CR over time. Updated analyses based on 3-year data will be presented. Clinical trial identification: NCT01844505 (067) NCT01721772 (066) NCT01927419 (069) C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

1214O

Epacadostat plus pembrolizumab in patients with advanced melanoma: Phase 1 and 2 efficacy and safety results from ECHO-202/ KEYNOTE-037

O. Hamid1, T.F. Gajewski2, A.E. Frankel3, T.M. Bauer4, A.J. Olszanski5, J.J. Luke6, A.S. Balmanoukian1, E.V. Schmidt7, B. Sharkey8, J. Maleski8, M.J. Jones8, T.C. Gangadhar9 1 Research, The Angeles Clinic and Research Institute, Los Angeles, CA, USA, 2 Department of Pathology and Department of Medicine, Section of Hematology/ Oncology, University of Chicago, Chicago, IL, USA, 3Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TX, USA, 4Drug Development Unit, Sarah Cannon Research Institute/Tennessee Oncology, Nashville, TN, USA, 5Department of Medical Hematology/Oncology, Fox Chase Cancer Center, Philadelphia, PA, USA, 6 Hematology/Oncology, University of Chicago, Chicago, IL, USA, 7Oncology Clinical Research, Merck & Co., Inc., Kenilworth, NJ, USA, 8Drug Development, Incyte Corporation, Wilmington, DE, USA, 9Medicine, Hematology/Oncology Division, Abramson Cancer Center of the University of Pennsylvania, Philadelphia, PA, USA Background: Tumors can evade immunosurveillance through upregulation of indoleamine 2,3-dioxygenase 1 (IDO1). Epacadostat (E) is a potent, selective inhibitor of the IDO1 enzyme. The combination of E þ the PD-1 inhibitor pembrolizumab (P) is being evaluated in an open-label, phase 1/2 study in multiple tumor types (ECHO-202/ KEYNOTE-037). We report phase 1 and 2 efficacy and safety data for patients (pts) with advanced melanoma (27Feb2017 data cutoff). Methods: Pts previously treated with checkpoint inhibitors were excluded. Pts received E (25, 50, 100, or 300 mg PO BID) þ P (2 mg/kg or 200 mg IV Q3W) during phase 1.

abstracts

Annals of Oncology MTD was not exceeded. E (100 mg BID) þ P (200 mg Q3W) was selected for phase 2. Responses were assessed in RECIST 1.1 evaluable pts. Results: 64 pts enrolled (phase 1, n ¼ 22; phase 2, n ¼ 42). Median age, 65; male, 70%; BRAFþ, 30%; M1c disease, 52%. Median duration of follow-up was 253þ days (range, 5 to 904þ days). Among 54 efficacy evaluable pts, ORR was 56% (30/54; 8 CR, 22 PR) and DCR (CRþPRþSD) was 78% (42/54). In treatment-naı¨ve pts (n ¼ 45), ORR was 56% (25/45; 6 CR, 19 PR) and DCR was 78% (35/45). Among treatment-naı¨ve pts receiving E 100 mg BID (n ¼ 30), ORR was 60% (18/30; 2 CR, 16 PR). Responses were observed regardless of PD-L1 and BRAF mutation status. At data cutoff, 28/30 responses in the melanoma cohort were ongoing (median duration of response ¼ 287.5þ days, range 1þ to 763þ days). Median PFS was 12.4 mo; PFS rates at 6, 12, and 18 mo were 70%, 54%, and 50%, respectively. In treatment-naı¨ve pts, median PFS has not been reached; PFS rates at 6, 12, and 18 mo were 68%, 52%, and 52%. The most common (>15%) all-grade treatment-related AEs (TRAEs) were fatigue (39.1%), rash (32.8%), pruritus (26.6%), and arthralgia (15.6%). Grade 3 TRAEs were observed in 17.2% of pts (most common: lipase increased, n ¼ 4; rash, n ¼ 3; and amylase increased, n ¼ 2). 3 pts discontinued for TRAEs (lipase increased, n ¼ 1; arthralgia, n ¼ 2). No treatment-related deaths occurred. Biomarker evaluation is ongoing. Conclusions: Consistent with the phase 1 results, E þ P continues to be well tolerated and showed promising clinical activity. A phase 3 study in pts who are treatment-naive for advanced melanoma is ongoing (NCT02752074). Clinical trial identification: NCT02178722 Legal entity responsible for the study: Incyte Corporation, Wilmington, DE Funding: Incyte Corporation, Wilmington, DE; Merck & Co., Inc., Kenilworth, NJ Disclosure: O. Hamid: Advisory Board - Merck & Co., Inc, Amgen, Novartis, Roche, Bristol-Myers Squibb; Corporate-sponsored Research - Incyte Corporation (Institution), Merck & Co., Inc. (Institution); Speaker’s Bureau – Bristol-Myers Squibb, Genetech, Novartis, Amgen; Honoraria – Genetech, Bristol-Myers Squibb, Novartis. T.F. Gajewski: Advisory Board - Merck & Co., Inc; Corporate-sponsored Research - Incyte Corporation (Institution), Merck & Co., Inc. (Institution). T.M. Bauer: Corporate-sponsored Research - Incyte Corporation (Institution), Merck & Co., Inc. (Institution) A.J. Olszanski: Advisory Board - Merck & Co., Inc, Bristol-Myers Squib; Corporate-sponsored Research - Incyte Corporation (Institution), Merck & Co., Inc. (Institution), Bristol-Myers Squibb, Novartis, Teva, Takeda, Pfizer; Other Substantive Relationships - Data Safety Monitoring Board: Takeda. J.J. Luke: Consult: Amgen, Array, AstraZeneca, BeneVir, Bristol-Myers Squibb, Castle, CheckMate, EMD Serono, Gilead, Novartis, Merck; Inst res supp: AbbVie, BostonBiomedical, BristolMyers Squibb, Celldex, Corvus, Delcath, 5Prime, Genentech, Immunocore, Incyte, Intensity, MedImmune, Macrogenics, Novartis, Pharmacyclics, Merck, Tesaro. A.S. Balmanoukian: Corporate-sponsored Research – MedImmune/AstraZeneca, Merck Serono, Genentech, Incyte Corporation (Institution), Merck & Co., Inc. (Institution); Other Substantive Relationships - Speaker’s Bureau at Bristol-Myers Squibb, Merck, Genentech, AstraZeneca. E.V. Schmidt: Employment and stock ownership at Merck & Co., Inc. B. Sharkey, J. Maleski, M.J. Jones: Employment and stock ownership at Incyte Corporation T.C. Gangadhar: Corporate-sponsored Research - Incyte Corporation (Institution), Merck & Co., Inc. (Institution), Bristol-Myers Squibb, Roche, Cerulean; Honoraria - Merck & Co., Inc., Novartis; Advisory Role - Bristol-Myers Squibb All other authors have declared no conflicts of interest.

1215O

Results of COLUMBUS Part 2: A phase 3 trial of encorafenib (ENCO) plus binimetinib (BINI) versus ENCO in BRAF-mutant melanoma

R. Dummer1, P.A. Ascierto2, H. Gogas3, A. Arance4, M. Mandala5, G. Liszkay6, C. Garbe7, D. Schadendorf8, I. Krajsova9, R. Gutzmer10, V. Chiarion Sileni11, C. Dutriaux12, J.W.B. de Groot13, N. Yamazaki14, C. Loquai15, L.A. de Parseval16, M. Pickard17, V. Sandor18, C. Robert19, K.T. Flaherty20 1 Department of Dermatology, University Hospital Zu¨rich-Dermatology, Zurich, Switzerland, 2Unit of Melanoma, Cancer Immunotherapy and Innovative Therapy, Istituto Nazionale Tumori Fondazione Pascale, Naples, Italy, 3Department of Internal Medicine, National and Kapodestrian University of Athens, Laikon Hospital, Athens, Greece, 4Department of Medical Oncology, Hospital Clinic of Barcelona, Barcelona, Spain, 5Department of Oncology and Haematology, Papa Giovanni XXIII Cancer Center Hospital, Bergamo, Italy, 6Department of Dermatology, National Institute of Oncology, Budapest, Hungary, 7Department of Dermatology, Division of Dermatooncology, Eberhard Karls University, Tuebingen, Germany, 8Department of Dermatology, University Hospital Essen, Heidelberg, Germany, 9Department of Dermato-oncology, University Hospital Prague and Charles University First Medical Faculty, Prague, Czech Republic, 10Department of Dermatology and Allergy, Skin Cancer Center, Hannover Medical School, Hannover, Germany, 11Melanoma Cancer Unit, Oncology Institute of Veneto IRCCS, Padua, Italy, 12Department of Oncologic Dermatology, Centre Hospitalier Universitaire de Bordeaux, Hoˆpital Saint-Andre´, Bordeaux, France, 13Department of Medical Oncology, Isala, Zwolle, Netherlands, 14Department of Dermatologic Oncology, National Cancer Center Hospital, Tokyo, Japan, 15Department of Dermatology, University Medical Center, Mainz, Germany, 16Translational Clinical Oncology, Novartis Pharma AG, Basel, Switzerland, 17Biostatistics and Data Management, Array BioPharma Inc., Boulder, CO, USA, 18Medical Affairs, Array BioPharma, Boulder, CO, USA, 19Department of Medicine, Institut Gustave Roussy, Villejuif, France, 20Department of Medical Oncology, Massachusetts General Hospital, Boston, MA, USA Background: The addition of a MEK inhibitor (MEKi) to a BRAF inhibitor (BRAFi) in BRAF V600-mutant metastatic melanoma improves efficacy, including progression-

Volume 28 | Supplement 5 | September 2017

free survival (PFS) and objective response rate (ORR), and attenuates some BRAFiassociated toxicities. Part 1 of the COLUMBUS study met its primary endpoint. The BRAFi ENCO 450 mg once daily (QD) þ the MEKi BINI 45 mg twice daily (BID; COMBO450) improved PFS vs vemurafenib (VEM) alone and ENCO 300 mg QD (ENCO300) alone in patients (pts) with advanced BRAF V600-mutant melanoma. The tolerability of COMBO450 was favorable compared with VEM or ENCO300. In Part 2, the contribution of BINI to the combination was further evaluated by maintaining the same dose of ENCO in the combination (ENCO 300 mg QD þ BINI 45 mg BID; COMBO300) and comparator arms (ENCO300 alone; ClinicalTrials.gov, NCT01909453; EudraCT, 2013-001176-38). Methods: Pts were randomized 3:1 to COMBO300 or ENCO300. Data from ENCO300 arms in Parts 1 þ 2 were combined for the primary efficacy comparison of PFS by independent blinded central review (BCR). Other analyses included PFS for COMBO300 vs ENCO300 (Part 2 only), ORR, complete response (CR) and partial response (PR) by BCR and local review, and safety. Results: Pt characteristics are presented in the Table. Median PFS (95% CI) for COMBO300 was 12.9 mo (10.1–14.0) vs 9.2 mo (7.4–11.0) for ENCO300 (Parts 1 þ 2) and 7.4 mo (5.6–9.2) for ENCO300 (Part 2). The hazard ratio (HR) for COMBO300 was 0.77 (0.61–0.97; P ¼ 0.029, 2-sided) vs ENCO300 (Parts 1 þ 2) and 0.57 (0.41– 0.78; PT) and PD1.9 (c.644C>T) single nucleotide polymorphisms (SNPs) on PDCD1 gene influence risk, clinicopathological aspects and survival of patients with cutaneous melanoma (CM). Methods: We evaluated 250 CM patients diagnosed at the University of Campinas and 250 blood donors (controls). DNA was analyzed by real-time polymerase chain reaction (PCR) for genotyping. PDCD1 gene expression and PD1 protein expression were assessed by quantitative PCR and flow cytometry, respectively. The statistical significance of differences between groups was calculated using the Fisher’s exact or chisquare test. Bonferroni method was used in multiple comparisons. PDCD1 expression and PD1 expression on T lymphocytes were calculated, using Kruskal-Wallis and Mann-Whitney test, respectively. The prognostic impact of SNPs on recurrence-free survival (RFS) and overall survival (OS) of CM patients were examined using the Kaplan Meier and Cox analyses. Results: Individuals with PD1 CC genotype isolated and associated with PD1.5 CC genotype were under 2.20 (95% CI: 1.00-4.82, P¼ 0.04) and 2.51 (95% CI: 1.04-6.03, P¼ 0.03) times greater risks of developing CM, respectively. Individuals with phototype I or II and PD1 CC genotype or PD1 CC plus PD1.5 CC genotype had 5.89 and 6.71 more chances of presenting CM than others, respectively. PD1.5 TT genotype was associated with increased expression of PDCD1 gene when compared with CT or CC genotype (P¼ 0.03). PD1.5 CT or TT genotypes and T allele increased expression of PD1 protein in CD4þ lymphocytes (P¼ 0.01, P¼ 0.006; respectively). At 60 months of follow-up, shorter RFS was observed in patients with PD1.1 AA genotype (33.3% vs 72.5%, P¼ 0.02). Patients with PD1.1 AA genotype had 4.39 more chances of presenting tumor progression or relapse in univariate Cox analysis (P¼ 0.04) and patients with PD1.5 CC genotype had 2.38-fold increased risk of evolving to death in multivariate Cox analysis (P¼ 0.02). Conclusions: The data suggest, for the first time, preliminary evidence that inherited abnormalities in regulation of T lymphocyte activities, related to PD1.1, PD1 and PD1.5 SNPs, alter CM risk and prognosis. Legal entity responsible for the study: Faculty of Medical Sciences, University of Campinas Funding: S~ao Paulo Research Foundation (FAPESP) Disclosure: All authors have declared no conflicts of interest.

1231P

Role of an intronic polymorphism in the CREB1 gene, involved in melanogenesis, with the risk and the aggressiveness of cutaneous melanoma

J.K. Silva1, C. Oliveira1, B. Sa Carvalho2, C. Torricelli1, G.V.B. Gomez1, W.D.L. Oliveira1, J.A. Rinck-Junior3, A.M. Moraes3, M.M. Ortega4, C.S.P. Lima3, G.J. Lourenc¸o1 1 Laboratory of Cancer Genetics, Faculty of Medical Sciences, Universidade Estadual de Campinas-UNICAMP, Campinas, Brazil, 2Department of Statistics, Institute of Mathematics, Statistic and Computer Science, Universidade Estadual de CampinasUNICAMP, Campinas, Brazil, 3Department of Internal Medicine, Faculty of Medical Sciences, Universidade Estadual de Campinas-UNICAMP, Campinas, Brazil, 4Laboratory of Cell and Molecular Tumor Biology and Bioactive Compounds, S~ ao Francisco University, Braganc¸a Paulista, Brazil Background: Recently, we observed 12,882 new single nucleotide polymorphisms (SNPs) associated with cutaneous melanoma (CM) risk in 103 patients and 103 controls, using large-scale genotyping with microarrays. CREB1 c.303 þ 373G>A, involved in melanogenesis and located in regulatory sequence of mRNA processing (splicing), was selected for further analyses. An in silico analysis showed that referred SNP may alters the binding sites of splicing regulatory proteins, such as SF1 and hnRNP A1. However, the role of this SNP in the risk, aggressiveness and prognosis of CM is unknown. Verify whether the distinct genotypes of CREB1 c.303 þ 373G>A influence the CM risk and prognosis, clinicopathological aspects, and CREB1, SF1 and HNRNPA1 mRNA levels. Methods: Genomic DNA of 262 patients and 280 controls was analyzed by RT-PCR. Patients were treated with conventional procedures. Gene expressions were determined

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology by qPCR using total RNA of 56 controls. Chi-square, logistic regression model, MannWhitney and Student’s t tests analyzed the differences between groups. Progressionfree survival (PFS) and overall survival (OS) times were calculated using Kaplan-Meier and Cox regression analyses. Results: CREB1 GA or AA genotypes were more frequent in CM patients than in controls (72.0% vs. 61.1%, P ¼ 0.02). Carriers of the genotypes were under 1.61-fold increased risk of CM (95% CI: 1.07-2.41) than others. An excess of CREB1 AA variant genotype was seen in patients with Breslow’s thickness higher than 1.5mm (28.2% vs. 18.5%, P ¼ 0.04) and high Clark’s level (26.2% vs. 13.3%, P ¼ 0.02). The median of follow-up of CM patients was 76 months; no association of referred SNP and patients’ PFS and OS was observed in this study. Individuals with CREB1 GA or AA genotypes presented higher mRNA expression of CREB1 (0.94 vs. 0.60 arbitrary units (UAs), P ¼ 0.007), SF1 (1.33 vs. 1.05 UAs, P ¼ 0.03) and HNRNPA1 (0.77 vs. 0.57 UAs, P ¼ 0.02) than those with GG wild-type genotype. Conclusions: Our data suggest, for the first time, that CREB1 c.303 þ 373G>A SNP is an important hereditary factor for the risk and aggressiveness of CM, possibly due to variation of the splicing factors. Legal entity responsible for the study: University of Campinas Funding: Foundation for protection of research in the state of S~ao Paulo (FAPESP) Disclosure: All authors have declared no conflicts of interest.

1232P

Investigation of AMBRA1 as a melanoma susceptibility gene

V. Hoeiom1, M. Yang1, K. Nosrati1, A. Azimi2, S. Egyhazi Brage1, R. Tuominen1 Oncology and Pathology, Karolinska Institutet, Stockholm, Sweden, 2OncologyPathology, Karolinska Institutet, Stockholm, Sweden

1

Background: Melanoma is the most lethal form of skin cancer, which shows a rapid increase in incidence in many countries including Sweden. To date, the annual increase is over 5% and there is an urgent need to improve possibilities for prevention and early diagnoses when the prognosis is far favorable compared to disseminated disease. Melanoma is caused by an interplay of environmental and genetic factors and is one of the cancer forms showing highest heritability. Still a substantial extent of the genes underlying melanoma susceptibility is unknown. Methods: We have executed whole-exome sequencing of melanoma-prone families to identify novel melanoma predisposing genes. Further genetic and functional studies of strong candidate genes using patient samples and melanoma cell lines has been performed. Various in vitro assays have been used to determine the role of these genes in for example autophagy and cell proliferation. Results: One gene discovered was the autophagy/beclin-1 regulator 1 (AMBRA1), where a putative splice variant was co-segregating with the melanoma phenotype in a 4-case family. This mutation was not found among over 6000 Swedish populationbased controls nor in any additional melanoma patients. AMBRA1 is essential in the regulation of autophagy and apoptosis and has been suggested to function as a tumor suppressor. By gene expression analysis we identified several transcripts of AMBRA1, with differential expression in melanoma tumors and in various melanoma cell lines. In tumor material from the splice variant carrier AMBRA1 showed low levels of expression. In melanoma cell lines, AMBRA1 was up-regulated when adding an autophagy activating reagent while down-regulated when treating the cells with Chloroquine, a drug inhibiting autophagy. AMBRA1 was also significantly up regulated when treating the cells with Crizotinib, a drug that targets the tyrosine kinase receptor c-MET and may induce autophagy, whereas no effect was seen when using the BRAF-inhibitor Vemurafenib. Thus, AMBRA1 may be involved in the Crizotinib-induced autophagy pathway. Conclusions: Preliminary data suggest AMBRA1 as a candidate melanoma susceptibility gene with a role during autophagy in melanoma cells. Further studies are needed to elucidate the specific role of this gene in melanoma development. Legal entity responsible for the study: Karolinska Institutet Funding: The Swedish Cancer Society, The Swedish Research Council, Regne´rs foundation Disclosure: All authors have declared no conflicts of interest.

1233P

Influence of an intronic polymorphism in the MITF gene, of melanogenic pathway, in the risk and the prognosis of cutaneous melanoma

C. Torricelli1, C. Oliveira1, B. Sa Carvalho2, J.K. Silva1, G.V.B. Gomez1, W.D.L. Oliveira1, J.A. Rinck-Junior3, A.M. Moraes3, M.M. Ortega4, C.S.P. Lima3, G.J. Lourenc¸o1 1 Laboratory of Cancer Genetics, Faculty of Medical Sciences, University of Campinas, Campinas, Brazil, 2Department of Statistics - Institute of Mathematics, Statistics and Computer Science, University of Campinas, Campinas, Brazil, 3Department of Internal Medicine, University of Campinas, Campinas, Brazil, 4Laboratory of Cell and Molecular Tumor Biology and Bioactive Compounds, S~ ao Francisco University, Braganc¸a Paulista, Brazil Background: We identified more than 12,000 new single nucleotide polymorphisms (SNPs) associated with cutaneous melanoma (CM) risk in 103 patients and 103 controls, using large-scale genotyping with DNA microarrays. A bioinformatics analysis

Volume 28 | Supplement 5 | September 2017

showed that MITF c.938-325G>A SNP, involved in melanogenesis and located in regulatory sequence of mRNA processing (splicing), may alter the binding sites of splicing proteins, such as SF1 and hnRNP A1. However, the role of this SNP in the risk and prognosis of CM patients is still unknown. We aim to evaluate the influence of this SNP on the risk and prognosis of CM, clinical and tumor characteristics, and MITF, SF1 and HNRNPA1 levels. Methods: MITF genotypes of 262 CM patients and 280 controls were identified in DNA by RT-PCR. Patients were treated with conventional protocols. Gene expressions were evaluated by qPCR using RNA of 73 controls. The differences between groups were assessed by chi-square, logistic regression, t test and ANOVA. Progression-free (PFS) and overall survival (OS) times were estimated by Kaplan-Meier and Cox methods. Results: The frequency of the AA variant genotype was higher in patients than in controls (26.8% vs. 21.1%, P ¼ 0.03). Individuals with referred genotype were under 1.60fold increased risk of CM (95% CI: 1.02-2.52) than others. The frequency of GA or AA genotypes was more common in patients with lower phototype (I-III) (90.8% vs. 80.9%, P ¼ 0.04) and with vertical tumors (83.7% vs. 67.5%, P ¼ 0.04). The median of follow-up was 76 months. At 60 months, PFS (53.4% vs. 71.6%, P ¼ 0.005, Cox: HR: 1.84, P ¼ 0.006) and OS (76.2% vs. 82.4%, P ¼ 0.02, Cox: HR: 1.79, P ¼ 0.03) were shorter in patients with AA genotype than others. We observed similar frequencies of MITF (1.2 vs. 1.1 vs. 1.0 arbitrary units (AUs), P ¼ 0.30), SF1 (1.1 vs. 1.2 vs. 1.0 AUs, P ¼ 0.94) and HNRNPA1 (1.1 vs. 1.3 vs. 1.3 AUs, P ¼ 0.61) mRNA levels in individuals with distinct genotypes. Conclusions: Our results suggest, for the first time, that MITF c.938-325G> SNP is an important inherited factor for the risk and prognosis of CM. Our findings, once validated in additional studies, will contribute to personalize the therapy of CM patients. Legal entity responsible for the study: University of Campinas (UNICAMP) Funding: S~ao Paulo Research Foundation (FAPESP) Disclosure: All authors have declared no conflicts of interest.

1234P

Hybrid-capture based genomic profiling identifies BRAF V600 and non-V600 alterations in melanoma samples negative by prior testing

L. Boussemart1, A. Wang2, M.K.K. Wong3, J.S. Ross4, P.J. Stephens5, S.M. Ali6, J. Sosman7, J.M. Mehnert8, G. Daniels9, K. Kendra10, A.B. Schrock6, V.A. Miller11 1 Department of Dermatology, Pontchaillou Hospital, CHU de Rennes, Rennes, France, 2 Clinical Development, Foundation Medicine, Inc., Cambridge, MA, USA, 3Department of Melanoma Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA, 4 Pathology, Albany Medical Center, Albany, NY, USA, 5R & D, Foundation Medicine, Cambridge, MA, USA, 6Clinical Development, Foundation Medicine, Cambridge, MA, USA, 7Oncology, Northwestern University, Chicago, IL, USA, 8Division of Medical Oncology, The Cancer Institute of New Jersey, New Brunswick, NJ, USA, 9Medicine, University of California San Diego, La Jolla, CA, USA, 10Oncology, The Ohio State University, Columbus, OH, USA, 11Medical Affairs, Foundation Medicine, Cambridge, MA, USA Background: BRAF and MEK inhibitors are approved for V600-mutated melanoma, and response rates of up to 70% are seen for patients with V600 mutations. Responses to targeted therapies have also been observed for a variety of non-V600 BRAF alterations. Thus, sensitive, accurate, and broad detection of BRAF alterations is critical to match patients with available targeted therapies. Methods: Pathology reports were reviewed for 385 consecutive melanoma cases (Mar 2016 - Mar 2017) with BRAF mutations or rearrangements identified using a hybridcapture based next generation sequencing (NGS) assay during the course of clinical care. Results: Records of prior BRAF molecular testing were available for 79 (21%) cases, utilizing PCR (n ¼ 30), Sanger sequencing (n ¼ 13), IHC (n ¼ 10), non-hybrid capture based NGS (n ¼ 9), or other or unspecified methodology (n ¼ 17). Of cases with BRAF V600 mutations 11/57 (19%) with available data were negative by prior BRAF testing, including 2/11 (18%) with confirmation that the same biopsy was tested. In cases with BRAF V600 mutations, there was no significant difference in mutant allele frequencies (median 35% vs. 40%, p ¼ 0.25) or percentage of tumor nuclei (median 50% for both, p ¼ 0.97) between samples with prior negative and prior positive results. Prior negative results were also identified in 16/20 (80%) cases with non-V600 mutations, two of which harbored multiple BRAF alterations [K601E (4), D594A/G/N (4), S467L (2), L584F (2), G464V, G466V, G469V, E586K, N581I, L597Q, A589_T599insT]. Two of 2 (100%) cases with activating BRAF fusions also had prior negative BRAF results. Clinical outcomes for a subset of patients will be presented. Conclusions: Despite approved companion diagnostics, significant variability exists in methods for BRAF testing in the clinical setting. Hybrid-capture based NGS identifies diverse activating mutations and fusions, including BRAF V600E, in a significant fraction of cases for which prior BRAF testing returned negative results. Given the proven clinical benefit in patients with BRAF alterations treated with match targeted therapies, hybrid-capture based NGS should be considered for patients with metastatic melanoma, particularly if other testing is negative. Legal entity responsible for the study: Foundation Medicine, Inc. Funding: Foundation Medicine, Inc. Disclosure: A. Wang, J.S. Ross, P.J. Stephens, S.M. Ali, A.B. Schrock, V.A. Miller: Employee with stock ownership in Foundation Medicine, Inc. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx377 | 437

abstracts 1235P

Post-transcriptional regulation of immune checkpoint genes by mir16 in melanoma

R. Leibowitz-Amit1, A. Layani2, J. Roszik3, Y. Sidi2, D. Avni2, E. Grimm4 Oncology, Chaim Sheba Medical Center, Ramat Gan, Israel, 2Cancer research center, Chaim Sheba Medical Center, Ramat Gan, Israel, 3Genomic medicine, MD Anderson Cancer center, Houston, TX, USA, 4Melanoma medical oncology, MD Anderson Cancer center, Houston, TX, USA

1

Background: The complex interface between T lymphocytes and cancer (’the immunological synapse’) comprises of both co-stimulatory and co-inhibitory proteins that modulate lymphocytes towards activation or anergy. ’Checkpoint inhibitors’ have impressive activity in melanoma, but not all patients respond and drug resistance often develops. MiRNAs are master regulators of gene expression. Our aim is to study the regulation of the immunological synapse by miRNAs in melanoma. Methods: Bioinformatic analyses of mRNAs and miRNA expression in 451 samples from the melanoma TCGA database was performed. Spearman rho correlation coefficients were calculated and survival analysis was performed using the Kaplan-Meier method. Direct mRNA targets of miRNAs were found using luciferase reporter assays, and mRNA/miRNA expression was assessed by qRT-PCR following either ectopic expression or depletion of specific miRNAs. Results: Of 15 checkpoint mRNAs and 8 miRNAs examined, nine checkpoint mRNAs showed a highly statistically significant positive correlation to each other and, to a lesser extent, to mir-16. These results were fully corroborated in vitro. Mir-16 may potentially target the 3’UTR of 3 of these mRNAs. CD80 (B7.1) was found to a direct target of mir16 in vitro. Overexpression of mir-16 in melanoma cell lines led to downregulation of CD80, CD274 (PD-L1) and CD40, while downregulation of mir-16 increased the expression of these genes. Survival data from 163 stage III melanoma patients show that high levels of mir-16 and low levels of any of six checkpoint mRNAs (among them CD80) is significantly associated with poor prognosis. Conclusions: Our results suggest that mir-16 and many checkpoint mRNAs are generally under a strict joint transcriptional regulation. The ability of mir-16 to decrease CD80 expression suggests that it serves as a key regulator of the immunological sample. We hypothesize that in vivo, an aberrantly high expression of mir-16 decreases the expression of the co-stimulatory checkpoint CD80 in melanoma and other checkpoint mRNAs, leading to immune evasion and compromised outcome. Further elucidation of both the transcriptional and post-transcriptional regulation of the immunological synapse may help point to novel targets and means for immune modulation. Legal entity responsible for the study: Raya Leibowitz-Amit Funding: Israeli Scientific Foundation (ISF) Disclosure: All authors have declared no conflicts of interest.

1236P

Does melanoma or other skin cancers belong to the BRCA2 phenotype?

R. Vitorino1, F. Vaz2, A.L. Carvalho3, S. Bento2, A. Luıs2, A. Opini~ao3, A. Clara2, J. Dupont4, S. Santos5, P. Machado6, S. Fragoso5, P. Rodrigues4, J. Parreira4, C. Moura4 1 Medical Oncology, Instituto Portugue`s de Oncologia de Lisboa Francisco Gentil, E.P.E. (IPOLFG EPE), Lisbon, Portugal, 2Familiar Cancer Risk Clinic, Instituto Portugue`s de Oncologia de Lisboa Francisco Gentil, E.P.E. (IPOLFG EPE), Lisbon, Portugal, 3Medical Oncology, Instituto Portugu^es de Oncologia de Lisboa Francisco Gentil, Lisbon, Portugal, 4Familiar Cancer Risk Clinic, Instituto Portugu^es de Oncologia de Lisboa Francisco Gentil, E.P.E. (IPOLFG EPE), Lisbon, Portugal, 5Molecular Pathobiology Research Department, Instituto Portugu^es de Oncologia de Lisboa Francisco Gentil, E.P.E. (IPOLFG EPE), Lisbon, Portugal, 6Molecular Pathobiology Investigation Unit, Instituto Portugu^es de Oncologia de Lisboa Francisco Gentil, E.P.E. (IPOLFG EPE), Lisbon, Portugal Background: The BRCA2 phenotype includes breast (BC) and ovarian cancer (OC) as well as, less frequently, prostate (PC), gastric (GC) and pancreatic cancer. The association with melanoma remains unclear, because previous studies were retrospective and included non- confirmed carriers. With this study we intend to determine the rate of melanoma and nonmelanoma skin cancers diagnoses in a consecutive prospective cohort of confirmed BRCA2 carriers. Methods: Review of all skin cancer diagnoses in BRCA2 carriers under prospective surveillance. Results: Four hundred and eighty six BRCA2 carriers (376 female, 110 male) belonging to 216 families were identified. The median age for genetic diagnosis was 48,3yrs with the BRCA2 c.156_157inAlu being the mutation most frequently observed (43,6%). Most carriers (359/486) had yearly full skin examinations. Although a majority of women (226/376) were cancer survivors (209 BC, 29 OC and 12 with breast/ovarian cancer), only 30/110 men had a previous cancer diagnosis (20 BC, 8 PC and 2 GCs). For a median follow up of 4 yrs, melanoma diagnoses were 3 in 2 female with bilateral BC. The patient with 2 melanomas had the first melanoma at 28yrs before BC diagnoses. Other skin cancers: 14 squamous cell carcinoma (SCC) (8 invasive, 6 in situ SCC) and 14 pts with basal cell carcinoma (BCC), 6 of which with multiple BCC. Four pts were diagnosed with actinic keratosis (AK). The rate for melanoma is 0,4% and for nonmelanoma skin cancers 5%. The rate for skin cancer in BRCA2 carriers with a previous cancer diagnosis is 2%. No statistical significance was found either for the association of skin cancer (p ¼ 0,221) or melanoma (p ¼ 0,9) with specific BRCA2 mutations.

438 | Melanoma and other skin tumours

Annals of Oncology Conclusions: The low rates of melanoma diagnosis in our prospective confirmed BRCA2 cohort, raises questions about the previously described association of melanoma and BRCA2 mutations. Also, no association was found between the Portuguese founder mutation and melanoma or other skin cancers. Although more follow up may be needed, there is insufficient evidence to warrant increased skin surveillance of BRCA2 carriers in the absence of standard skin cancer risk factors. Legal entity responsible for the study: IPOL FG, E.P.E. Funding: None Disclosure: All authors have declared no conflicts of interest.

1237P

Resected malignant melanoma at high risk of recurrence in SEERMedicare

N. Sadetsky1, J. Yi2, A. Hernandez1, D. Colburn1, G. Goodman1 Product Development, Genentech, Inc., South San Francisco, CA, USA, 2Epidemiology/ Analysis, Genesis Research Group, Hoboken, NJ, USA

1

Background: While surgery remains a mainstay in the management of high-risk resectable malignant melanoma (MM), there is a high chance of recurrence. Utilization of approved adjuvant therapies (e.g. interferon a and ipilimumab) are limited by the common occurrence of debilitating side effects. The objective of our study was to describe characteristics of patients (pts) with resected MM at high risk of recurrence in the older US population. Methods: A retrospective cohort study was undertaken using the Surveillance, Epidemiology, and End Results (SEER)-Medicare population-based linked database. The study population included pts with Stage IIC-IIIC surgically resected MM diagnosed between 2004 and 2011. Demographic and clinical characteristics, adjuvant therapies, including radiation (XRT) and/or systemic therapy (eg, interferon a, interleukin, pegylated interferon), and overall survival (OS) were evaluated. Results: We identified 1016 pts; the mean age was 75.2 years (interquartile range [IQR], 72–82) and 66.2% were males. The majority of pts had Stage IIC-IIIB disease at diagnosis (Cohort 1; n ¼ 877 [86.3%]); the remainder had stage IIIC disease (Cohort 2; n ¼ 139 [13.7%]). Adjuvant therapy was utilized in 27.3% (n ¼ 239) and 43.2% (n ¼ 60) of pts in Cohorts 1 and 2, respectively, and consisted of XRT in 74% and 78% of pts, systemic therapy in 16% and 10% of pts (with interferon a representing 98.6% of systemic therapies), and a combination of XRT and systemic therapy in 10% and 12% of pts. OS differed between cohorts, with a median of 32.3 months (IQR, 17.9– 53.3) for Cohort 1 and 19.8 months (IQR, 11.5–36.2) for Cohort 2. Landmark OS at 5 years was 20.8% for Cohort 1 and 12.2% for Cohort 2. Conclusions: Among pts with resected MM at high risk of recurrence in the older US population, utilization of adjuvant therapy and OS varied based on disease stage at diagnosis. Pts with Stage IIIC disease were exposed to more medical interventions; however, use of highly toxic systemic therapy available during the study period was limited in both cohorts. As more therapies for the adjuvant setting are being developed, the evaluation of clinical and demographic characteristics may help tailor treatment regimens. Legal entity responsible for the study: F. Hoffmann-La Roche Ltd. Funding: F. Hoffmann-La Roche Ltd. Disclosure: N. Sadetsky, A. Hernandez, D. Colburn: Employee, Genentech, Inc. G. Goodman: Employee, Genentech, Inc.; owns stock in Roche. All other authors have declared no conflicts of interest.

1238P

Independent prognostic impact of lympho-vascular invasion in cutaneous melanoma patients with sentinel lymph node biopsy

R. Luca, M. Rizzo, P. Mando, C. Perez de La Puente, A. Blanco, S. Rivero, G. Lutter, F. Cappuccio, M. Amat, J. Kaplan, R. Chacon, M. Chacon Medical Oncology, Instituto Alexander Fleming, Caba, Argentina Background: Incidence of cutaneous melanoma (CM) is increasing worldwide. The primary treatment of CM is surgery. Prognosis is determined by characteristics of the lesion such as depth of invasion, ulceration and sentinel lymph node (SLN) status. The aim of this study was to analyze the prognostic impact of lympho-vascular invasion (LVI) in CM patients (pts) undergoing SLN biopsy since LVI has not been established as a clear prognostic factor in the current AJCC 8th ed. cancer staging system. Methods: Retrospective, descriptive and observational analytical study. We used the institutional database of pts with diagnosis of CM, submitted to SLN biopsy between November 1994 and August 2016. The association between pathological characteristics and SLN were analyzed using Chi2 and logistic regression model. Kaplan Meier and Log rank were used for disease free survival (DFS) analysis. Results: 385 pts with a diagnosis of CM were analyzed. Median follow-up 45.2 months (IQR: 15.66-91.77). Median age: 52 years (IQR 42-65). SLNþ: 47/384 (12.2%). Evaluated prognostic factors: Breslow (Br) 1.5 mm md (IQR 1-2.67), ulceration þ 94/ 385 (24.4%), LVI þ 32/144 (22.2%). Relapse 86/367 (23.4%). In the univariate analysis we found association between relapse and the following factors: LVI þ (OR: 2.97, p ¼ 0.0125), SLN þ (OR: 3.97, p < 0.01), Br 1mm (OR: 4.13, p ¼ 0.01) and ulceration þ (OR: 2.08, p < 0.01). There was no association with age and sex. In the multivariate analysis LVI þ (OR: 2.47, p ¼ 0.049) and SLN þ (OR: 3.91, p ¼ 0.048)

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology were associated with relapse, whereas neither Br 1 mm, sex nor ulceration were associated with relapses. 5-year-DFS was higher in SLN - (79.3% vs 56.1%, p < 0.01), LVI (80.7% vs 57.9%, p 0.019), Br < 1 mm (89.5% vs 71%, p < 0.01) and ulceration (80.4% vs 59.7%, p < 0.01). Conclusions: In our retrospective series, after a long period of follow-up, the presence of LVI as an independent factor was associated with relapse and DFS. Within CM pts the best candidate for adjuvant therapy is yet to be defined, LVI þ as a prognostic factor should be validated in prospective trials in this scenario. Legal entity responsible for the study: Instituto Alexander Fleming Funding: Instituto Alexander Fleming Disclosure: All authors have declared no conflicts of interest.

1239P

Validating prognostic models in metastatic uveal melanoma (MUM), an international rare cancers initiative

L. Khoja1, E. Atenafu2, A.M. Joshua3, I.R.C.I. Ocular Melanoma Group4 Clinical Development Unit, AstraZeneca plc, Melbourn, UK, 2Biostatisticis, University of Toronto, Toronto, ON, Canada, 3Medical Oncology, Kinghorn Cancer Centre, Sydney, Australia, 4Cancer Research UK, International Rare Cancers Initiative, London, UK

1

Background: We validated 2 models (the 7thAmerican joint committee on cancer (AJCC) and the Helsinki university central hospital (HUCH) staging) and 1 nomogram; the Padova-Mayo (PMN), for progression free (PFS) and overall survival (OS) using patient (pt) level data from the PUMMA meta-analysis. Methods: 29 prospective trials’ (1988-2015) pt data was analysed. Models were validated with cox regression analysis for survival in months (m). Concordance index (CCI) was used to test predictive value. Results: Comparable data was available for 463 pt; see table for variables used in each system. Models were prognostic differentiating into M1a, M1b and M1c groups. Median PFS for AJCC was 4m for M1a, 3 for M1b and 2 for M1c. Median PFS for HUCH was 3.5m for M1a, 2.5 for M1b and 1 for M1c. CCI for PFS using AJCC was 0.69 (SE 0.02, 95%CI 0.65-0.73), for HUCH it was 0.79 (SE 0.02, 95%CI 0.74-0.83). Median OS for AJCC was 15m for M1a, 9 for M1b and 5 for M1c. Median OS for HUCH was 13m for M1a, 6 for M1b and 2 for M1c. CCI for OS for AJCC was 0.69 (SE 0.02, 95%CI 0.65-0.73). For HUCH it was 0.79 (SE 0.02, 95%CI 0.74-0.83). Using ECOG and LDH (available variables used in PMN) median PFS was 4m (95% CI 4-5) for normal LDH and ECOG 0, 7 (3-9) for normal LDH and ECOG > 0, 2.6 (2-3) for elevated LDH and ECOG 0 and 2.5 (2-3) for elevated LDH and ECOG > 0. Corresponding median OS was 17m (95%CI 15-18), 12.7 (95%CI 10-19), 7.4 (95%CI 6.3-8.9) and 5.3 (95%CI 3.86.1). CCI were PFS 0.72 (SE 0.02, 95% CI 0.69-0.75), OS 0.73 (SE 0.02, 95% CI 0.70.76).

Conclusions: Prognostic models in MUM remain imprecise in an externally validated dataset. Further validation is needed to find clinical utility Legal entity responsible for the study: Princess Margaret Cancer Centre Funding: None Disclosure: L. Khoja: Employed by Astrazeneca plc. All other authors have declared no conflicts of interest.

1240P

Impact of an active surveillance programme on outcome of patients (pts) with uveal melanoma (UM) after primary curative therapy (PTx): results of a single-institution experience

C.L. Lee1, G. Gullo1, N. Horgan2, C. Baily2, R. Abdul Rahman1, D. Smith1, C. Buckley1, J. Crown1 1 Medical Oncology, St Vincents University Hospital, Dublin, Ireland, 2Opthalmology, Royal Victoria Eye and Ear Hospital, Dublin, Ireland Background: About 30% of pts with UM develop metastatic disease (MUM) despite PTx. Liver is by far the commonest site of metastases. MUM has poor prognosis and no systemic treatment (STx) has been proven to improve overall survival (OS). However, the role of active surveillance for metastatic disease is still controversial. Methods: We performed an outcome analysis of all UM pts prospectively registered onto our active surveillance programme after PTx. All pts had systemic staging at initial diagnosis of UM and then 6-monthly liver imaging (CT triple-phase or ultrasound) and clinical review for the first 5 years and 12-monthly afterwards. Progression-free survival (PFS) was calculated from time of first systemic relapse to first disease progression, OS from time of first systemic relapse to death or latest FU. Results: Out of 166 pts registered between April 2009 and April 2017, 36 (22%) developed MUM: 14 pts relapsed 5 yrs from PTx. MUM pts characteristics: males 19 (53%); median age 58 (range 34-85); median tumour thickness at diagnosis 9mm (2-22); sites of metastases: liver only 13 (36%), liver þ other sites 21 (58%), extra-hepatic only 2 (6%). Relapses were asymptomatic and detected on surveillance imaging in 29 (80%) pts. Nine pts (7 detected from surveillance) underwent primary hepatic metastasectomy (HM), 27 (75%) pts were non-resectable (NR) and underwent STx (n ¼ 18), locoregional Tx (n ¼ 4), best supportive care (n ¼ 5). Overall, 29/36 MUM pts received immunotherapy with either ipilimumab or nivolumab/pembrolizumab. At a median FU of 36.5 mos (1-103), 27 pts have died and the median OS is 16.6 mos (95%CI: 7.8-25.3). Both PFS and OS were statistically significantly longer for HM pts compared to NR pts (PFS: 10.8 vs 4.4mos, p ¼ 0.01/OS: 24.9 vs 13.4mos, p ¼ 0.04). Eight out of 9 pts developed further disease relapse after HM. Conclusions: Our data indicate that active surveillance after PTx of UM can allow detection of asymptomatic potentially resectable liver metastases, especially in pts with

Table: 1239P 7thAJCC

Variable (n 5 463) (n (%), median, range) ECOG

Diameter in cm of largest metastasis

Diameter of largest liver lesion % liver involvement LDH ALP Disease free interval OS (%, 95% CI) M1a

M1b

M1c

HUCH

0 1 >2 < 3 cm 3-8 cm > 8 cm

PMN 296 (64) 156 (34) 11 (2)

1.9 (0-2.9) 4.4 (3-8) 10.8 (8.1-22.5) 3.8 (0-22.5)

< 20 20-50 > 50 Missing

3 (0-65) 5 (1) 3 (1) 452 (98) 344 (39-8198) 89 (24-1178) Missing

6 12 24 6 12 24 6 12 24

Volume 28 | Supplement 5 | September 2017

89 (83-93) 60 (52-68) 25 (18-33) 68 (62-74) 38 (31-44) 14 (94-19) 45 (33-57) 19 (10-29) 78 (27-17)

84 (80-88) 55 (49-60) 22 (18-27) 48 (39-57) 17 (11-24) 5 (2-1) 8 (0-29) 0 0

doi:10.1093/annonc/mdx377 | 439

abstracts

Annals of Oncology

high risk UM (i.e. tumour thickness >5mm). Although durable remission after HM is rare PFS and OS may be significantly prolonged. Legal entity responsible for the study: St Vincent’s Healthcare Group Funding: None Disclosure: All authors have declared no conflicts of interest.

1241P

Impact of duration of response (DOR) on overall survival (OS) in patients with metastatic melanoma treated with dacarbazine (DTIC), vemurafenib (V), or cobimetinib plus vemurafenib (C1V): a pooled analysis

K. Lewis1, J. Larkin2, A. Ribas3, K.T. Flaherty4, G.A. McArthur5, P.A. Ascierto6, B. Dre´no7, E. McKenna8, Q. Zhu9, Y. Mun8, A. Hauschild10 1 Medicine, University of Colorado Comprehensive Cancer Center, Aurora, CO, USA, 2 Oncology, Royal Marsden Hospital NHS Foundation Trust, London, UK, 3Jonsson Comprehensive Cancer Center, University of California, Los Angeles, Los Angeles, CA, USA, 4Medicine, Massachusetts General Hospital, Boston, MA, USA, 5Department of Medical Oncology, Peter MacCallum Cancer Centre, East Melbourne, Australia, 6 Melanoma, Cancer Immunotherapy and Innovative Therapy Unit, Istituto Nazionale Tumori Fondazione Pascale, Naples, Italy, 7Dermatology, Nantes University, Nantes, France, 8Product Development, Genentech, Inc., South San Francisco, CA, USA, 9 Oncology, Genentech, Inc., South San Francisco, CA, USA, 10Dermatology, University Hospital Schleswig-Holstein, Kiel, Germany Background: Evaluation of treatment efficacy in oncology using OS is confounded by survival benefit from post-progression treatment. We pooled data from the BRIM-2, 3, -7, and coBRIM studies (BRAF inhibitor–naive patients with BRAFV600-mutated metastatic melanoma) to evaluate whether DOR could be a surrogate for OS. Methods: Time-dependent Cox proportional hazards regression was used to model the association of DOR (interval from date of first RECIST response to progressive disease [PD] or death) with OS. The risk of death for DORs of 1–10 months (in 1-month increments) was evaluated. Patients with best response of stable disease or PD [nonresponders (NR)] were assigned a DOR of zero. Models were adjusted for time-fixed baseline covariates (ECOG status, demographics, disease covariates, and first-line treatment), and time-dependent covariates (DOR and post-progression treatment [immunotherapy, targeted therapy, or other]). Results: This analysis included 1365 patients (DTIC ¼ 338; V ¼ 717; CþV ¼ 310). Objective response was 47.5% for the overall population and 11.5%, 53.6%, and 72.9% for the DTIC, V, and CþV cohorts, respectively. Median DOR was 9.3 months in the overall population and 6.4, 7.6, and 14.6 months in the DTIC, V, and C þ V cohorts, respectively. Cox proportional hazards adjusted for time-dependent covariates showed a significant and progressive reduction in the risk of death with increasing DOR vs NR. The absolute risk of death decreased by a mean of 6.3–7.7% per month increase in DOR in the overall population and across treatment cohorts (Table). Sensitivity analyses in responders only showed similar results. Conclusions: These exploratory analyses suggest that DOR is independently associated with OS outcomes regardless of treatment and merits further exploration as a surrogate endpoint to assess long-term treatment benefit. Legal entity responsible for the study: F. Hoffmann-La Roche Ltd. Funding: F. Hoffmann-LaRoche Ltd. Disclosure: K. Lewis: Grants from Roche/Genentech, Amgen, EMD Serono, and Incyte, and personal fees from Roche/Genentech, Incyte, and SunPharma. J. Larkin: Institutional research support from MSD, Bristol-Myers Squibb, Pfizer, and Novartis and nonremunerated consultant for GSK, Novartis, MSD, Bristol-Myers Squibb,

Pfizer, and Roche/Genentech. A. Ribas: Owns stock in Kite Pharma and has received honoraria from Roche, Amgen, Pfizer, and Merck. All monies paid to Dr. Ribas are deposited into the Division Account at the David Geffen School of Medicine and do not constitute personal income. K.T. Flaherty: Consultant for Roche. G.A. McArthur: Research grant support from Pfizer, Celgene, Ventana; consultant for Provectus; uncompensated consultancy for Pfizer, Millennium, GSK, Roche-Genentech, Novartis, Bristol-Myers Squibb, and Amgen P.A. Ascierto: Consulting or advisory role for Amgen, Array, Bristol-Myers Squibb, Genentech/Roche, Merck Serono, Merck Sharp & Dohme, Novartis, and Pierre-Fabre, and research funding from Bristol-Myers Squibb, Genentech/Roche, and Array. B. Dre´no: Personal fees from Roche, BristolMyers Squibb, Novartis, GlaxoSmithKline, and Amgen. E. McKenna: Employee, Genentech, Inc. Q. Zhu, Y. Mun: Employment, Genentech, Inc. A. Hauschild: Personal fees from Roche, Amgen, Bristol-Myers Squibb, Celgene, Eisai, GlaxoSmithKline, MedImmune, Merck Serono, Merck Sharp & Dohme/Merck, Novartis, Oncosec, and MELA Sciences.

1242P

Neutrophil to Lymphocyte Ratio (NLR) as an independent prognostic measure in patients receiving targeted therapy or immunotherapy for stage IV melanoma

E. Blackley1, L.E. Lim1, M. Moore1, M. Voskoboynik1, C. McLean2, A. Haydon1 Medical Oncology, Alfred Health, Melbourne, Australia, 2Department of Pathology, Alfred Health, Melbourne, Australia

1

Background: Treatment of metastatic melanoma has rapidly evolved with the introduction of targeted and immunotherapies in recent years. An elevated NLR (neutrophil-lymphocyte ratio) has been shown to be an independent marker of poor prognosis in malignancies including melanoma. Here we present an updated survival analysis demonstrating the utility of NLR as a marker of prognosis in patients with metastatic melanoma receiving targeted and immunotherapy. Methods: We identified patients with stage 4 melanoma who received systemic therapy with targeted therapy (BRAF þ/- MEK inhibitor) or immunotherapy (Anti-CTLA-4 or Anti-PD-1) at our institution. Patients not receiving any systemic therapy were excluded. We retrospectively reviewed all medical records collecting data on baseline demographics, prognostic factors (stage, LDH, CNS and Liver metastases), treatments received, pre-treatment NLR and outcomes. Overall survival (OS) and Progressionfree survival (PFS) were measured from date of first dose received. Results: 174 patients were treated between August 2010 to November 2016, 74 received targeting therapy and 100 receiving immunotherapy. Median follow up was 10 months. At time of interim analysis median OS for patients with NLR < 5 was 11.7 months compared to 4.8 months in NLR >5 (HR 0.45, 95% C.I. 0.31-0.67, p ¼ 0.00007), this was seen in patients treated with both targeted therapies (HR 0.48, p ¼ 0.012) and immunotherapies (HR 0.40, p ¼ 0.0009). Median PFS was also longer in patients with NLR 5 is a strong independent predictor of poor outcome in patients with metastatic melanoma regardless of targeted or immunotherapy. We hypothesis that at final data lock in July 2017 this association will remain strong given it was a clear predictor of outcome at the time of interim analysis. NLR may assist selection of initial therapy, for example, a favourable ratio may indicate suitability for single agent rather than doublet immunotherapy with its greater toxicity profile. Legal entity responsible for the study: Alfred Health Funding: None Disclosure: All authors have declared no conflicts of interest.

Table: 1241P Hazard of death by duration in response (excluding time-dependent PD variables) Patient cohort

DOR of 1 month HR (95% CI)

DOR of 10 months HR (95% CI)

Mean per month HR decrease

Range of HR decrease

P-valuea

All patients V CþV DTIC All respondersb V CþV DTIC

0.85 (0.82–0.87) 0.80 (0.77–0.84) 0.89 (0.86–0.93) 0.81 (0.71–0.92) 0.88 (0.85–0.90) 0.84 (0.80–0.88) 0.91 (0.88–0.94) 0.87 (0.78–0.97)

0.19 (0.14–0.25) 0.11 (0.07–0.18) 0.33 (0.23–0.46) 0.12 (0.03–0.45) 0.27 (0.21–0.35) 0.17 (0.11–0.27) 0.37 (0.27–0.51) 0.26 (0.09–0.77)

0.073 0.077 0.063 0.076 0.068 0.074 0.059 0.069

0.034–0.130 0.028–0.157 0.039–0.095 0.029–0.154 0.038–0.108 0.033–0.135 0.039–0.086 0.037–0.111

60%) who received 1 ipi dose may discontinue ipi. Pts with SD or better who subsequently have PD may be eligible for a second course of pembro þ ipi or pembro monotherapy (maximum 17 doses pembro and 4 doses ipi). Eligible pts with PD may remain on treatment until a confirmatory scan 4 wk later. Primary end points are safety and ORR; secondary end points include PFS, OS, and DOR. Enrollment is ongoing in the US, Australia, and New Zealand. Clinical trial identification: NCT02089685 Legal entity responsible for the study: Merck & Co., Inc., Kenilworth, New Jersey, USA Funding: Merck & Co., Inc., Kenilworth, New Jersey, USA Disclosure: M.B. Atkins: Advisory board member for Bristol-Myers Squibb, Merck, Roche, Novartis, Pfizer, Celldex, Allexion. M.S. Carlino: Advisory board member for Bristol-Myers Squibb, Merck, Amgen, Novartis. Honoraria from Merck, Bristol-Myers Squibb. C.M. McNeil: Advisory board member for Merck, Sharp & Dohme. Speakers bureau for Merck, Sharp & Dohme and Bristol-Myers Squibb. Research funding from Merck, Sharp & Dohme. Travel expenses, including accommodations from Merck, Sharp & Dohme. A. Ribas: Stock ownership in Kite Pharma. Honoraria from Amgen, Pfizer, Merck, Roche. V. Atkinson: Advisory board member: MSD, Bristol-Myers Squib, Novartis, Pierre Fabre. Speakers bureau: MSD, Bristol-Myers Squib, Novartis. Honararia: MSD, Bristol-Myers Squib, Novartis. Travel expenses, including accommodations: MSD, Bristol-Myers Squib, Novartis. M.B. Jameson: Travel expenses, including accommodations for Merck Sharp & Dohme W-J. Hwu: Advisory board member for Merck. Speakers bureau for Merck, Bristol-Myers Squibb, GlaxoSmithKline, MedImmune. J.A. Thompson: Advisory board: CellDex. Research funding to institution: Merck, Agensys, Seattle Genetics, Pfizer. Honoraria: CellDex. J. Anderson: Employment with Merck. B. Homet Moreno: Employee of Merck & Co, Inc. Stock ownership in Merck & Co., Inc. N. Ibrahim: Employment: Merck. Stock: Merck, GSK. G.V. Long: Advisory board member Amgen; Bristol-Myers Squibb; Array; Merck; Merck, Sharp & Dohme; Novartis, Roche, Pierre Fabre. Honararia from Bristol-Myers Squibb; Roche; Merch, Sharp & Dohme. All other authors have declared no conflicts of interest.

1258TiP

Phase 2 Study Comparing Pembrolizumab with Intermittent/shortterm dual MAPK pathway inhibition plus Pembrolizumab(PEM) in patients harboring the BRAFV600 mutation (IMPemBra Trial)

E.A. Rozeman1, M.A. Deken2, J. Gadiot3, M.H. Geukes Foppen1, L. Pronk4, J.V. Thienen1, J.B. Haanen1, C.U. Blank1 1 Medical Oncology Department, Het Nederlands Kanker Instituut Antoni van Leeuwenhoek (NKI-AVL), Amsterdam, Netherlands, 2Molecular Oncology and Immunology, Het Nederlands Kanker Instituut Antoni van Leeuwenhoek (NKI-AVL), Amsterdam, Netherlands, 3Immunology, Het Nederlands Kanker Instituut Antoni van Leeuwenhoek (NKI-AVL), Amsterdam, Netherlands, 4Department of Biometrics, Het Nederlands Kanker Instituut Antoni van Leeuwenhoek (NKI-AVL), Amsterdam, Netherlands Background: Continuous combinations of targeted therapy (TT), e.g. BRAFþMEK inhibitors(BRAFiþMEKi), with immunotherapy (IT), e.g. CTLA-4 or PD-1 blockade are currently tested in several phase1/2 trials with the aim to improve response rate and response duration in melanoma patients with a BRAFV600 mutation. However, high toxicity rates have been observed, revealing PD-1 blockade currently being the only possible combination partner for TT. Recently we have published preclinical data, showing that short-time TT induces strong T cell infiltration and is synergistic with PD-1 blockade. Analysis of biopsies of patients during TT indicate that long-term TT might be counterproductive, as T cell infiltration decreases in some patients already beyond 2 weeks. This raises the question which time period of MAPK pathway inhibition is optimal for combination with anti-PD-1. The IMPemBra trial will address this question, comparing PEM monotherapy with combination schemes of intermittent/shortterm BRAFi þ MEKi plus PEM. The primary objective is to explore safety, feasibility and the immune-activating capacity of the different regimens.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Trial design: Stage IV BRAFV600E/K mutation positive melanoma patients, naı¨ve for IT and TT, will start treatment with PEM 200mg q3wk. After 6 wks the patients will be randomized (stratified according their LDH level) to continue PEM for up to 2 years (cohort 1), or to one of the experimental cohorts receiving either dabrafenib 150mg BID þ trametinib 2mg QD two times intermittent for 1 wk (cohort 2), two times intermittent for 2 wks (cohort 3), or continuous for 6 wks (cohort 4). All cohorts continue afterwards with PEM for up to 2 years. Each cohort will consist of 8 patients. Primary endpoints are SUSARs and adherence to the study timeline, the intra-patient alteration in intratumoral CD8þ T cells and the percentage PD1þ CD8þ T cells in the peripheral blood. Tumor biopsies and blood samples including PBMCs are taken at baseline, wk 6, 9, 12, 18 and in case of progression. Secondary endpoints are objective response rate and progression free survival. Enrollment started in May 2016, 11 patients have been included so far. Clinical trial identification: NCT02625337 Legal entity responsible for the study: NKI-AVL Funding: MSD Disclosure: J.V. Thienen: Advisory board: MSD and Bristol-Myers Squib. J.B. Haanen: Advisory role: Bristol-Myers Squib. MSD, Pfizer, Roche, Novartis, Neon Therapeutics Research grants: Bristol-Myers Squib, MSD, GSK. C.U. Blank: Advisory board: BristolMyers Squib, MSD, Novartis, GSK, Pfizer, Lilly, Roche Research grants: Bristol-Myers Squib, Novartis. All other authors have declared no conflicts of interest.

1259TiP

A randomized, double-blind, placebo-controlled, phase III study comparing the combination of PDR001, dabrafenib and trametinib versus placebo, dabrafenib and trametinib in previously untreated patients with unresectable or metastatic BRAF V600–mutant melanoma (COMBI-i)

E. Gasal1, A.M. Arance Fernandez2, P.A. Ascierto3, V. Atkinson4, R. Dummer5, K.T. Flaherty6, J-J. Grob7, J. Hansson8, J. Hassel9, J. Larkin10, C. Lebbe´11, G.V. Long12, P. Lorigan13, W. Miller14, P. Nathan15, A. Ribas16, C. Robert17, D. Schadendorf18, H. Tawbi19, A. Upalawanna20 1 Medical Oncology, Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA, 2 Department of Medical Oncology, Hospital Clınic Barcelona, Barcelona, Spain, 3 Melanoma, Cancer Immunotherapy and Innovative Therapy Unit, Istituto Nazionale Tumori Fondazione Pascale, Naples, Italy, 4Department of Medical Oncology, University of Queensland, Brisbane, Australia, 5Department of Dermatology, University of Zurich Hospital, Zurich, Switzerland, 6Medicine, Dana-Farber Cancer Institute/Harvard Medical School and Massachusetts General Hospital Cancer Center, Boston, MA, USA, 7 Dermatology, Aix-Marseille University and APHM Hospital CHU Timone, Marseille, France, 8Department of Oncology-Pathology, Karolinska Institute, Stockholm, Sweden, 9 Department of Dermatology and NCT, Universit€ at Heidelberg, Heidelberg, Germany, 10 Medicine, Royal Marsden Hospital, London, UK, 11APHP Dermatology and CIC Departments, Hoˆpital Saint-Louis University Paris Diderot, Paris, France, 12Melanoma Medical Oncology, Melanoma Institute Australia, University of Sydney, and Mater Hospital, Sydney, Australia, 13Division of Molecular & Clinical Cancer Sciences, The Christie NHS Foundation Trust, Manchester, UK, 14Departments of Oncology and Medicine, McGill University, Montreal, QC, Canada, 15Department of Medical Oncology, Mount Vernon Cancer Centre, Northwood, UK, 16Department of Medicine, UCLA Medical Center, Los Angeles, CA, USA, 17Department of Medicine, Gustave Roussy Comprehensive Cancer Center, Villejuif–Paris Sud, France, 18Department of Dermatology, University Hospital Essen and German Cancer Consortium, Essen, Germany, 19Department of Melanoma Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA, 20Medical Oncology, Novartis Pharma AG, Basel, Switzerland Background: Checkpoint inhibitor and targeted therapies are both important tools in the management of BRAF V600–mutated unresectable or metastatic melanoma. Although these therapies have improved responses and overall survival, many patients still progress and die from this disease. Thus, additional treatment strategies are needed to improve durability of responses and related long-term outcomes in these patients. Based on preclinical and preliminary clinical data, BRAF and MEK inhibitors can reverse the oncogenic BRAF-induced immune-suppressive phenotype through enhanced melanoma antigen expression and enhanced tumor antigen-specific T-lymphocyte recognition in vivo. These data suggest that there is potential clinical benefit in combining dabrafenib and trametinib with checkpoint inhibitor therapy. Trial design: The 3-part COMBI-i phase 3 study (NCT02967692) will evaluate the safety and efficacy of PDR001, an investigational anti–programmed death 1 antibody, in combination with dabrafenib and trametinib in previously untreated patients with BRAF V600–mutated unresectable or metastatic melanoma. In part 1, a safety run-in will establish the recommended phase 3 regimen (RP3R) for use in part 3 using an adaptive Bayesian logistic regression model. In part 2, tissue and blood samples from the biomarker cohort will be used to characterize baseline immune markers and explore potential immune marker modulation by the triplet therapy. Part 3 is the randomized, double-blind, placebo-controlled portion that will open once the RP3R has been determined. Approximately 500 patients will be randomized 1:1 to receive either PDR001 in combination with dabrafenib and trametinib or placebo in combination with dabrafenib and trametinib, with randomization stratified based on Eastern Cooperative Oncology Group performance status and lactate dehydrogenase level. The primary endpoint will be progression-free survival per investigator’s assessment according to RECIST v1.1. Overall survival will be a key secondary endpoint.

Volume 28 | Supplement 5 | September 2017

Clinical trial identification: NCT02967692 First received: November 16, 2016 Legal entity responsible for the study: Novartis Pharmaceuticals Corporation Funding: Novartis Pharmaceuticals Corporation Disclosure: E. Gasal: Employment: Novartis Stock or Other Ownership: Amgen Inc, Novartis. A.M. Arance Fernandez: Honoraria, Consulting/Advisory Role, and Speakers Bureau: Roche, Bristol-Myers Squibb, MSD, and Novartis Travel/Accommodations/ Expenses: Roche, Bristol-Myers Squibb P.A. Ascierto: Consulting/Advisory Role: Bristol-Myers Squibb, Roche/Genentech, Merck Sharp & Dohme, Novartis, Amgen, Array BioParma, Merck Serono Research Funding: Bristol-Myers Squibb, Roche/ Genentech, Array BioPharma. V. Atkinson: Honoraria and Speakers Bureau: BristolMyers Squibb, MSD, Novartis, Roche Consulting/Advisory Role: Bristol-Myers Squibb, MSD, Novartis, Pierre Fabre Travel/Accommodations/Expenses: BristolMyers Squibb, MSD, Novartis. R. Dummer: Honoraria and Consulting/Advisory Role: Roche, Bristol-Myers Squibb, MSD, Novartis, GlaxoSmithKline, Amgen, Takeda, Pierre Fabre Research Funding: Roche, GlaxoSmithKline, Bristol-Myers Squibb, Novartis, MSD, Takeda, Pierre Fabre. K.T. Flaherty: Consulting/Advisory Role: Novartis. J-J. Grob: Honoraria and Consulting/Advisory Role: Bristol-Myers Squibb, Roche, MSD, Novartis Speakers Bureau: Novartis Travel/Accommodations/Expenses: Bristol-Myers Squibb, MSD, Roche, Novartis. J. Hansson: Consulting/Advisory Role: Roche, Novartis, Merck, Bristol-Myers Squibb Travel/Accommodations/Expenses: Novartis. J. Hassel: Honoraria: Bristol-Myers Squibb, MSD, Roche, GlaxoSmithKline, Novartis, Amgen Research Funding: Bristol-Myers Squibb Travel/Accommodations/ Expenses: Bristol-Myers Squibb, MSD, Amgen, GlaxoSmithKline. J. Larkin: Honoraria: Bristol-Myers Squibb, MSD, Pfizer, Novartis, Eisai, GlaxoSmithKline, Roche Research Funding: Bristol-Myers Squibb, MSD, Pfizer, Novartis. C. Lebbe´: Honoraria, Consulting/Advisory Role: Roche, Bristol-Myers Squib, Novartis, Amgen, MSD Speakers Bureau: Bristol-Myers Squib, Amgen, Roche, Novartis Research Funding: Roche, Bristol-Myers Squib Travel/Accommodations/Expenses: Roche, Bristol-Myers Squib, Amgen. G.V. Long: Consulting/Advisory Role: Bristol-Myers Squibb, Merck, Novartis, Amgen, Roche P. Lorigan: Consulting/Advisory Role: Merck Sharp & Dohme, Bristol-Myers Squibb, Novartis, Amgen, GlaxoSmithKline Speakers Bureau: Merck Sharp & Dohme, Novartis, Bristol-Myers Squibb, Roche Travel/ Accommodations/Expenses: Merck Sharp & Dohme, Bristol-Myers Squibb. W. Miller: Honoraria: Bristol-Myers Squib, Merck, Roche, Novartis, GlaxoSmithKline Consulting/Advisory Role: Bristol-Myers Squib, Merck, Roche, Novartis, Amgen, GlaxoSmithKline Research Funding: Bristol-Myers Squib, Novartis, GlaxoSmithKline, Roche, AstraZeneca, Merck, MethylGene, Bayer, Amgen, MedImmune. P. Nathan: Consulting/Advisory Role: AztraZeneca, Bristol-Myers Squibb, MSD, Immunocore, Pfizer, Pierre Fabre, Novartis, GlaxoSmithKline, Ispen Speakers Bureau: Bristol-Myers Squibb, Novartis Travel/Accommodations/Expenses: Bristol-Myers Squibb, MSD. A. Ribas: Consulting/Advisory Role: Merck, Amgen, Genentech/Roche, Novartis, Lilly Stock or Other Ownership: Kite Pharma, Compugen, FLX Bio, CytomX Therapeutics, Arcus Ventures. C. Robert: Honoraria and Consulting/Advisory Role: Bristol-Myers Squibb, Roche, Merck, Novartis, Amgen, GlaxoSmithKline. D. Schadendorf: Honoraria, Speakers Bureau and Travel/Accommodations/Expenses: Amgen, BristolMyers Squibb, Novartis, MSD, Roche Consulting/Advisory Role: Bristol-Myers Squibb, Novartis, MSD, Roche, Array Research Funding: Bristol-Myers Squibb, Novartis. H. Tawbi: Consulting/Advisory Role: Novartis Research Funding: BristolMyers Squibb, Novartis, Merck. A. Upalawanna: Employment and Stock or Other Ownership: Novartis.

1260TiP

Late physical, psychological and social consequences of ipilimumab treatment in advanced melanoma

A.H. Boekhout1, M. Hauptmann2, A.J.M. van den Eertwegh3, G.A.P. Hospers4, J.W.B. de Groot5, M.J.B. Aarts6, E. Kapiteijn7, A.J. Ten Tije8, D. Piersma9, W.H.J. Kruit10, K.P.M. Suijkerbuijk11, F.W.P.J. Berkmortel12, G. Vreugdenhil13, R. H. Koornstra14, E. Fiets15, M. Lees16, K.J.M. Janssen17, L.V. van den Poll-Franse1, C.U. Blank18 1 Psychosocial Research and Epidemiology, Netherlands Cancer Institute, Amsterdam, Netherlands, 2Epidemiology, Netherlands Cancer Institute, Amsterdam, Netherlands, 3 Medical Oncology, VU University Medical Centre, Amsterdam, Netherlands, 4Medical Oncology, University Medical Centre Groningen, Groningen, Netherlands, 5Medical Oncology, Isala, Zwolle, Netherlands, 6Medical Oncology, Maastricht University Medical Centre, Maastricht, Netherlands, 7Medical Oncology, Leiden University Medical Centre, Leiden, Netherlands, 8Medical Oncology, Amphia hospital, Breda, Netherlands, 9Medical Oncology, Medical Spectrum Twente, Enschede, Netherlands, 10Medical Oncology, Erasmus University Medical Centre, Rotterdam, Netherlands, 11Medical Oncology, University Medical Centre Utrecht, Utrecht, Netherlands, 12Medical Oncology, Atrium Medical Centre, Sittard, Netherlands, 13Medical Oncology, Maxima Medical Centre, Eindhoven, Netherlands, 14Medical Oncology, Radboud Medical Centre, Nijmegen, Netherlands, 15Medical Oncology, Medical Centre Leeuwarden, Leeuwarden, Netherlands, 16Oncology, Bristol-Meyers Squibb, Paris, France, 17Oncology, BristolMeyers Squibb, Utrecht, Netherlands, 18Medical Oncology, Netherlands Cancer Institute, Amsterdam, Netherlands Background: After the introduction of ipilimumab, an anti-CTLA-4 monoclonal antibody, durable, long term survival has become a possibility for a subgroup of advanced melanoma patients. Since ipilimumab is a relatively novel drug there are limited data on the long-term physical, psychological, and social functioning of these patients. This study will evaluate the long-term physical and psychosocial performances and the information needs of advanced melanoma survivors who have been treated with ipilimumab.

doi:10.1093/annonc/mdx377 | 447

abstracts Trial design: This is a prospectively enrolling, multicentre cohort study. Objectives: To assess health-related quality of life (HRQoL), anxiety, depression, fatigue, fear of cancer recurrence, sexual health and generic health status in patients with advanced melanoma who have survived at least 2 years after ipilimumab treatment (without subsequent other systemic therapies) as compared with healthy controls, and to describe the melanoma-specific HRQoL, impact of cancer, social functioning and information needs in patients with advanced melanoma who have survived at least 2 years after ipilimumab treatment. Patients and healthy control population: Patients with advanced (stage IV or unresectable stage III) melanoma who survived at least 2 years and were treated with ipilimumab between 2011 and 2015 in 14 hospitals in the Netherlands are included. The patient population consists of 3 treatment groups based on time since ipilimumab treatment: 24 to < 36 months, 36 to < 48 months and 48 months post-ipilimumab treatment. The healthy control population will be selected from ‘Patient Reported Outcomes Following Initial treatment and Long term Evaluation of

448 | Melanoma and other skin tumours

Annals of Oncology Survivorship (PROFILES)’. PROFILES contains a reference cohort of more than 2000 healthy individuals and is designed to be representative of the Dutch-speaking population in the Netherlands. Measurements: The primary and secondary study outcomes will be measured by questionnaires, at 3 time-points in patients 24 to < 36 months and at 1 time-point in patients 36 months post-ipilimumab treatment. The primary outcome, HRQoL will be assessed with the European Organisation for Research and Treatment of Cancer quality of life questionnaire-C30 (EORTC QLQ-C30). Clinical trial identification: Date of release: November 2016 Legal entity responsible for the study: Netherlands Cancer Institute Funding: Bristol-Myers Squibb Disclosure: A.H. Boekhout: Employee of Bristol-Myers Squibb. M. Lee, KJM Janssen: Employee of and receiving stock from Bristol-Myers Squibb, during the conduct of the study. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

Annals of Oncology 28 (Supplement 5): v449–v452, 2017 doi:10.1093/annonc/mdx378

NEW DIAGNOSTIC TOOLS 1261P

mRNA capture sequencing enabled liquid biopsy screening

J. Vandesompele Research and Development, Biogazelle, Zwijnaarde, Belgium Background: In contrast to general belief, a substantial part of the human protein coding transcriptome is abundantly present in the blood as extracellular mRNA, ready to exploited. It is well known that cancer cells actively and passively release RNA cargo into circulation and their detection may inform on the patient disease status. Methods: We developed and applied a probe based RNA capture sequencing method as a sensitive RNA sequencing workflow to study thousands of transcripts in cell-free RNA from cancer patients’ plasma. The method is based on exome-style enrichment of a randomly primed cDNA library with preservation of strandedness information. More than one million capture probes target 21,000 human messenger RNA and 60,000 human long non-coding RNA genes. Apart from RNA abundance profiling, this type of data can also be used to detect structural RNA variants, such as somatic mutations, fusion genes, and RNA editing events, all known to play an important role in cancer. Results: On average, between 6000-10,000 RNA genes are reproducibly detected in 0.2 ml of plasma. Detection and coverage sensitivity is greatly increased by using larger plasma volume and improved adaptor ligation strategies. We also observed a positive correlation between number of platelets in plasma and detected genes and variants, in line with their tumor-educated nature. Our benchmarked RNA variant pipeline identifies thousands of germline and somatic variants in circulating mRNA. A dedicated titration experiment in which plasma from cancer and healthy individuals were mixed in known ratios demonstrates excellent quantitative performance. Pronounced RNA abundance differences and enriched pathways are observed between cancer types and during treatment. The RNA capture sequencing also works on other body fluids, such as urine and serum, and simultaneous targeting of mRNA and lncRNA provides substantial enrichment of otherwise low-abundant lncRNAs. Conclusions: RNA capture sequencing of liquid biopsies is a promising new application to support precision oncology and is expected to enhance therapy stratification, treatment response monitoring and early detection of relapse. Legal entity responsible for the study: Biogazelle and Ghent University Funding: Biogazelle, Illumina Disclosure: J. Vandesompele: Apart from professorship at Ghent University, cofounder and part-time CSO at Biogazelle, a Ghent University spin-off company.

1262P

Use of droplet digital PCR for quantitative and automatic analysis of the HER2 status in breast cancer patients

K. Otsuji1, T. Sasaki2, A. Tanaka2, A. Kunita2, M. Ikemura2, K. Matsusaka3, K. Tada1, M. Fukayama2, Y. Seto4 1 Department of Breast and Endocrine Surgery, Faculty of Medicine, The University of Tokyo, Tokyo, Japan, 2Department of Pathology, Faculty of Medicine, The University of Tokyo, Tokyo, Japan, 3Department of Molecular Oncology, Graduate School of Medicine, Chiba University, Chiba, Japan, 4Department of Stomach and Esophageal Surgery, Faculty of Medicine, The University of Tokyo, Tokyo, Japan Background: Digital polymerase chain reaction (dPCR) has been used to yield an absolute measure of nucleic acid concentrations. Recently, a new method referred to as droplet digital PCR (ddPCR) has gained attention as a more precise and less subjective assay to quantify DNA amplification. We demonstrated the usefulness of ddPCR to determine HER2 gene amplification of breast cancer. Methods: In this study, we used ddPCR to measure the HER2 gene copy number in clinical formalin-fixed paraffin-embedded samples of 41 primary breast cancer patients. To improve the accuracy of ddPCR analysis, we also estimated the tumour content ratio (TCR), the ratio of tumour cell count per section, for each sample. Results: Our determination method for HER2 gene amplification using the ddPCR ratio (ERBB2:ch17cent copy number ratio) combined with the TCR showed high consistency with the conventionally defined HER2 gene status according to ASCO-CAP (American Society of Clinical Oncology/College of American Pathologists) guidelines (P < 0.0001, Fisher’s exact test). The equivocal area was established by adopting 99% confidence intervals obtained by cell line assays, which made it possible to identify all conventionally HER2-positive cases with our method. In addition, we succeeded in automating a major part of the process from DNA extraction to determination of HER2 gene status. Conclusions: The introduction of ddPCR to determine the HER2 gene status in breast cancer is feasible for use in clinical practice and might complement or even replace conventional methods of examination in the future. Legal entity responsible for the study: The University of Tokyo Funding: None Disclosure: All authors have declared no conflicts of interest.

C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

1263P

BRCA1 large gene rearrangements (LGRs) in Russian breast cancer patients: the development of the droplet digital PCR assay for LGR detection and the identification of recurrent exon 8 deletions

E. Imyanitov1, E. Preobrazhenskaya1, I. Bizin2, E. Kuligina1, E. Anisimova3, S. Laptiyev4, E. Suspitsin1, S. Aleksakhina1, A. Togo1, A. Sokolenko1 1 Department of Tumor Biology, N.N. Petrov Institute of Oncology, St.-Petersburg, Russian Federation, 2Laboratory of Bioinformatics, Polytechnical University, St.Petersburg, Russian Federation, 3Outpatient Clinic, Leningrad Regional Oncology Dispensary, Saint Petersburg, Russian Federation, 4Department of Medical Biology and Genetics, I.P. Pavlov Medical University, St.-Petersburg, Russian Federation Background: Some pathogenic BRCA1 mutations are represented by large gene rearrangements (LGRs). LGRs cannot be detected by conventional Sanger sequencing. Multiplex ligation-dependent probe amplification (MLPA) is a leading method for LGR detection, however it is entirely based on the use of commercial kits, includes relatively lengthy hybridization step and is poorly suitable for the large-scale screening for recurrent deletions. Methods: We developed and validated the droplet digital PCR (ddPCR) assay, which covers the entire coding region of BRCA1 gene and is capable to precisely quantitate the copy number for each exon. Results: 141 breast cancer (BC) patients, who demonstrated evident clinical features of hereditary BC but turned out to be BRCA1/2 mutation-negative upon Sanger sequencing, were subjected to the LGR analysis. 4 patients with LGR were identified, with 3 cases of exon 8 deletion and 1 women carrying the deletion of exons 3-7. Excellent concordance with MLPA was observed. Exon 8 copy number was tested in additional 720 high-risk BC, and another 3 cases with the deletion were revealed; MLPA re-analysis demonstrated that exon 8 loss was a part of a larger genetic alteration in 2 cases, while the remaining patient had isolated defect of exon 8. Long-range PCR and next generation sequencing revealed, that 3 out of 4 samples with isolated exon 8 deletion had an identical rearrangement. Conclusions: Droplet digital PCR is a reliable tool for detection of large gene rearrangements. BRCA1 LGRs are rare in Russian hereditary BC patients, with exon 8 deletion being a recurrent allele in this population. Legal entity responsible for the study: Laboratory of Molecular Oncology, N.N. Petrov Institute of Oncology, St.-Petersburg Funding: Russian Science Foundation (grant 14-25-00111) Disclosure: All authors have declared no conflicts of interest.

1264P

Clinical evaluation of low density array based EGFR mutation detecting kit using tissue samples and liquid biopsies

J. Hernandez-Losa1, Y. Ruano2, R. Somoza1, A. Muro2, S. Ram onY Cajal1, J.L. RodrıguezPeralto2 1 Anatomical Pathology, Vall d’Hebron University Hospital, Barcelona, Spain, 2 Department of Pathology, 12 de Octubre University Hospital, Madrid, Spain Background: The information on EGFR mutations status improves the benefit of targeted therapies for the non-small-cell lung cancer patients. These determinations are already part of standard protocols for cancer treatment. The study of the mutations in EGFR in liquid biopsy also permits a follow-up throughout the treatment and detection of acquired resistance. V Methods: CLART CMA EGFR kit (GENOMICA, Spain) detects the 40 highprevalence mutations associated with sensitivity or resistance to the treatment. These mutations are located in the exons 18, 19, 20 and 21. The kit is based on the multiplex V ARMS-PCR followed by detection on a low-density microarray platform: CLART (Clinical Arrays Technology). The samples were processed in a bench-top semi-autoV mated system, Autoclart. A new version of the kit, CLART CMA EGFR LB, allows using liquid biopsy as a sample with only an additional pre-PCR step. Results: A 107 tissue biopsies from metastatic NSCLC patients were obtained and analyzed in two University Hospitals in Spain: Vall d’Hebron (Barcelona) and 12 de Octubre V (Madrid). The EGFR mutations were detected with CLART CMA EGFR kit with 96.3% concordance with the routine methods used in the hospital practice (Cobas/Therascreen/ Sanger sequencing). The discrepant results were analyzed by Sanger sequencing. The senV sitivity of the CLART CMA EGFR kit is 100% and specificity is 96.5%. The same CLART platform, only with the addition of one pre-PCR amplification step was used for processing the plasma samples (liquid biopsy) from the NSCLC metastatic patients. A total of 8 samples were tested. In six samples the results obtained from tissue samples and liquid biopsies were concordant. In the two discordant samples the exon 19 deletions detected in tissue were not detected in plasma samples. The wt results obtained for these two liquid biopsies were confirmed by Next Generation Sequencing. Conclusions: Given the high specificity and sensitivity and excellent concordance with V the other platforms in hospital practice, CLART CMA EGFR kit is valid for the use in R

R

R

R

R

R

abstracts

Annals of Oncology

the clinical routine. Furthermore, using the same semi-automated platform for tissue samples and liquid biopsy facilitates laboratory work and reduces turnover time per sample. Legal entity responsible for the study: GENOMICA Funding: GENOMICA Disclosure: All authors have declared no conflicts of interest.

1265P

Correlation of somatic genomic alterations between tissue genomics and circulating tumor DNA (ctDNA) employing next generation sequencing (NGS) analysis in lung and gastrointestinal cancers

M.M. Toor1, W. bahaj1, Z. Ahmed1, L. Kujtan1, T.J. Pluard2, M.E. McNally3, L.S. Cummings3, E. Island3, J. Forster3, K.F. Kennedy4, J. Subramanian2, A. Masood2 1 Department of Medicine, UMKC(University of Missouri Kansas City), Kansas City, MO, USA, 2Department of Medical Oncology, Saint Luke’s Cancer Institute, Kansas City, MO, USA, 3Department of Surgery, UMKC(University of Missouri Kansas City), Kansas City, MO, USA, 4Department of Cardiovascular Research, Saint Luke’s Health System, Kansas City, MO, USA Background: Peripheral blood circulating tumor DNA (ctDNA) and tumor tissue next-generation sequencing (NGS) is routinely performed to guide therapy in cancer patients. However, little is known about the concordance or discordance between commercially available tissue genomics testing panels and ctDNA. The aim of our study was to assess concordance between matched cancer tissue genomics and blood based ctDNA in lung and gastrointestinal (GI) cancers. V V Methods: Tissue genomic analysis was performed with Paradigm (n ¼ 17)/Caris V (n ¼ 11) and ctDNA was analyzed with Guardant360 (n ¼ 28). Samples included, non-small cell lung cancer (n ¼ 10), small cell lung cancer (n ¼ 4), colorectal cancer (n ¼ 5), hepatocellular carcinoma (n ¼ 2), intrahepatic cholangiocarcinoma (n ¼ 1), pancreatobiliary adenocarcinoma (n ¼ 3), esophageal adenocarcinoma (n ¼ 2) and gastric adenocarcinoma (n ¼ 1). Results: We identified 6 (21%) patients with at least one gene mutation that was detected by both tissue genomic and ctDNA analysis. Total number of gene mutations identified in 28 patients were 106, but only 8 (7.5%) were detected by both tissue and ctDNA panels. When this testing was done within 90 days the concordance increased to 10.20%. Table. R

reproducibility. Clinically, 43 ctDNA samples from lung, liver, colorectal, breast and gastric cancers were genomically profiled and compared to the known alternations in their matched solid tumors, in terms of single base substitution, insertions/deletions, copy number variations and rearrangement. Results: The analytical validation demonstrated unprecedented accuracy: near 100% specificity (99.6%, 99.9% and 100%) and 95.8%, 100% and 100% sensitivity for 3 reference materials, respectively. The actual detection limit was as low as 0.05%. The reproducibility was assessed as 0.998 (jaccard index) by sequencing 2 replicates of each reference. In clinical validation, compared to matched FFPE results, this ctDNA assay showed overall 99.9% specificity and 89% sensitivity, with > 90% sensitivity when only drug-gable hotspots were concerned. Eight events of gene rearrangements involving known targeted genes of ALK (n ¼ 4), ROS1 (n ¼ 3) and MET (n ¼ 1) were detected from seven patients with 100% sensitivity and 100% specificity, confirmed either by IHC or panel sequencing (depth > 1,000X) over their matched FFPE biopsies. Due to its typically low abundance, CNV from ctDNA was detectable only for those highly amplified genes ( > ¼8 copies) with > ¼4 exons, demonstrating 72.2% sensitivity and 99.5% specificity. Conclusions: Stringent criteria for both analytical and clinical validations are required for clinical utility of ctDNA. Our ctDNA assay has demonstrated high accuracy and reliability in comprehensively genomic profiling of ctDNA, especially in regard to druggable targets, which assures its translational utility in optimizing and monitoring targeted therapies in cancer management. Legal entity responsible for the study: OrigiMed Inc Funding: OrigiMed Inc Disclosure: W. Liu, S. Mu, J. Yao, H. Chen, Z. Hu, J. Hu, G. Chirn, H. Kang, K. Wang, M. Yao: Employee of OrigiMed Inc.

R

R

Table: 1265P Patients with at least one mutation detected by both platforms Mutations detectable by both platforms Mutations detectable by both platforms if interval between tissue and blood collection 5cm), II, IIIA non-small cell lung cancer comparing S-1 versus S-1 with cisplatin

T. Okamoto1, T. Yano2, M. Shimokawa3, S. Takeo4, K. Yamazaki4, K. Sugio5, M. Takenoyama6, A. Nagashima7, T. Tagawa1, Y. Emi8, Y. Maehara1 1 Department of Surgery and Science, Kyushu University, Graduate School of Medical Sciences, Fukuoka, Japan, 2Department of Thoracic Surgery, National Hospital Organization Beppu Medical Center, Beppu, Japan, 3Clinical Research Institute, National Kyushu Cancer Center, Fukuoka, Japan, 4Department of Thoracic Surgery and Clinical Research Institute, National Kyushu Medical Center Hospital, Fukuoka, Japan, 5 Department of Thoracic and Breast Surgery, Oita University, Faculty of Medicine, Yufu, Japan, 6Department of Thoracic Oncology, National Kyushu Cancer Center, Fukuoka, Japan, 7Department of Thoracic Surgery and Clinical Research Institute, Kitakyushu Municipal Medical Center, Kitakyushu, Japan, 8Department of Surgery, Saiseikai Fukuoka General Hospital, Fukuoka, Japan Background: Platinum-based combination chemotherapy is a standard postoperative adjuvant treatment for pathological stage II/III non-small cell lung cancer (NSCLC). C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

abstracts Oral S-1 therapy has been demonstrated to have a good efficacy with less toxicity for advanced NSCLC treatment. S-1 might be also useful in the adjuvant setting for surgical NSCLC. Methods: We performed a phase II randomized open label multi-institutional study in surgical patients with pathological stage IB (T > 5 cm), II or III NSCLC (7th TNM classification), who underwent complete resection from 2009 to 2013. One hundred and thirty-nine patients were randomly assigned to two arms: S-1 80mg/m2 oral, day1 to 14, q3w, 1year (arm A, n ¼ 70) or S-1 80mg/m2 oral day1 to 21, q5w þ cisplatin 60mg/ m2 day8, q5w, 4 courses, followed by S-1 80mg/m2 oral day1 to 14, q3w, total 1year (arm B, n ¼ 69). The primary endpoint was the disease free survival (DFS) rate at 2 years and was evaluated using Bayesian method. Either treatment arm would deserve further study if the Bayesian posterior probability that the 2-year DFS rate would exceed a value of 40% were more than 0.85. The secondary endpoints were overall survival (OS), safety, and feasibility. Results: The clinical characteristics of the patients were well balanced in terms of age, sex and pathological stage between the two arms. The DFS rate at 2 years was 51.4% (95% confidence interval [CI], 0.399–0.628) in arm A and 59.4% (95% CI, 0.476 – 0.702) in arm B. Both treatment arms met the primary endpoint: the probability of a DFS rate 40% at 2 years was over 0.97 in Arm and 1 in arm B. Neither DFS nor OS were significantly different (log-rank test; p ¼ 0.1695 and p ¼ 0.8684, respectively). No treatment-related deaths were observed in either treatment. The main G3/4 adverse events were appetite loss (arm A vs. arm B, 4.3% vs. 11.6%) and anemia (0% vs. 5.8%), which were not statistically different between the two arms. Treatment completion rate did not differ between the two arms (arm A vs. arm B: 45.7% [95% CI, 41.9–66.3%] vs. 43.5% [95% CI, 44.0–68.4%]). Conclusions: Oral S-1 based adjuvant chemotherapy was feasible and promising for patients with completely resected NSCLC. Clinical trial identification: UMIN000001765, 2009/03/12 Legal entity responsible for the study: academic group Funding: Kyushu University Lung Surgery Study Group Disclosure: All authors have declared no conflicts of interest.

Annals of Oncology 1279P

J-J. Hung, W-H. Hsu Department of Surgery, Taipei Veterans General Hospital, Taipei, Taiwan Background: The role of adjuvant chemotherapy for patients with stage I non-small cell lung cancer remains unknown. The prognostic value of histological subtypes in resected node-negative small-sized lung adenocarcinoma has not been widely investigated. This study investigated the prognostic factors in patients with node-negative lung adenocarcinoma of 3cm or smaller to find potential candidates for adjuvant chemotherapy. Methods: A total of 746 patients with completely resected node-negative lung adenocarcinoma of 3cm or smaller were included in the study. Prognostic factors for overall survival or probability of freedom from recurrence (FFR) were investigated. Results: The 5-year overall survival and recurrence-free rates were 86.8% and 84.8%, respectively. During follow-up, 59 (7.9%) patients developed recurrence. Univariate analysis showed that micropapillary/solid predominant group had significantly lower probability of FFR (P ¼ 0.001) in node-negative lung adenocarcinoma of 3cm or smaller. Older age (P ¼ 0.007), greater tumor size (P ¼ 0.006), and micropapillary/solid predominant group (P ¼ 0.031) had significantly lower probability of FFR in multivariate analysis. Conclusions: The new adenocarcinoma classification has significant impact on recurrence in node-negative lung adenocarcinoma of 3cm or smaller. Patients with micropapillary/solid predominant pattern have significant higher risk for recurrence. Legal entity responsible for the study: None Funding: None Disclosure: All authors have declared no conflicts of interest.

1280P 1278PD

Major pathological response after preoperative chemotherapy as a surrogate marker of survival in early-stage non-small cell lung cancer: cohort of NATCH phase III trial

J. Remon1, A. Martinez-Marti1, E. Carcereny Costa2, J. Zeron-Medina Cuairan1, I. Sansano3, J.L. Mate4, N. Pardo5, S. Cedres1, A. Navarro6, A.M. Martinez de castro5, T. Moran7, E. Felip Font8 1 Medical Oncology, Vall d’Hebron University Hospital, Barcelona, Spain, 2Medical Oncology, Catalan Institute of Oncology (ICO Badalona), Hospital Germans Trias i Pujol, Badalona, Spain, 3Pathology, Vall d’Hebron University Hospital, Barcelona, Spain, 4 Pathology, Hospital Universitari Germans Trias i Pujol, Badalona, Spain, 5Medical Oncology, Vall dHebron University Hospital/Vall dHebron Institute Oncology, Barcelona, Spain, 6Deparment of Oncology, University Hospital Vall d’Hebron, Barcelona, Spain, 7 Medical Oncology, Hospital Universitari Germans Trias i Pujol, Badalona, Spain, 8 Medical Oncology Service (Lung Cancer Unit), Vall d’Hebron University Hospital, Barcelona, Spain Background: Randomized phase III NATCH trial in early-stage non-small cell lung cancer (NSCLC) patients (p) reported no statistically differences in disease-free survival (DFS) or overall survival (OS) with the addition of preoperative or adjuvant chemotherapy to surgery. In pre-operative arm, those p who achieved a complete response obtained a benefit in 5-year DFS rate (59% vs. 38%). Recently, major pathological response (MPR) to preoperative chemotherapy (10% or less of residual viable tumor after preoperative therapy) has reported as surrogate marker of OS. The aim of this study is to validate MPR as prognostic factor in a cohort of patients included the NATCH trial. Methods: MPR was analysed in a whole cohort of 57 early-stage NSCLC p treated in the preoperative arm into NATCH trial from 2 institutions. OS according to MPR was analysed (long-rank test) in the whole population and by histologic subtype. Results: In this cohort, median age was 67 years (47-78), 48 p (84%) were males, 26 p (46%) squamous subtype. By stage according to 6th TNM: 9 p (16%) stage IA, 35 p (61%) stage IB, 12 p (21%) stage IIB and 1 p (2%) stage IIIA. 95% p completed 3 cycles of preoperative treatment. Surgical procedures: 81% lobectomies, 14% pneumonectomies, 5% no surgery. 13 out of 57 p (22.8%) had MPR. In the whole population, there was an increase in 5-year OS among those patients with MPR compare to p without MPR (84.6% vs. 58.5%, p ¼ 0.106). According to histological subtype, 5-year OS in squamous NSCLC p with MPR was significantly longer than in p without MPR (100% vs. 47.1%, p ¼ 0.026), but not differences in OS in non-squamous were detected (66.7% vs. 66.7%, p ¼ 0.586). Conclusions: MPR is a prognostic value in squamous NSCLC p who receive preoperative chemotherapy. Validation in extended cohort merits further evaluation. Legal entity responsible for the study: Enriqueta Felip Funding: None Disclosure: All authors have declared no conflicts of interest.

454 | NSCLC, early stage

Factors predicting worse outcomes in patients with N0 lung adenocarcinoma of 3cm or smaller

Regulatory variants in cancer-related pathway genes predict survival of patients with surgically resected non-small cell lung cancer

M.J. Hong1, K.M. Shin2, S.Y. Lee3, J.H. Lee4, J.Y. Park3 Cell and Matrix Research Institute, Kyungpook National University School of Medicine, Daegu, Republic of Korea, 2Department of Radiology, Kyungpook National University, Daegu, Republic of Korea, 3Department of Internal Medicine, Kyungpook National University, Daegu, Republic of Korea, 4Biochemistry and Cell Biology, Kyungpook National University School of Medicine, Daegu, Republic of Korea

1

Background: We conducted this study to identify genetic variants in cancer-related pathway genes which can predict prognosis of NSCLC patients after surgery, using a comprehensive list of regulatory single nucleotide polymorphisms (SNPs) prioritized by RegulomeDB. Methods: A total of 509 potentially functional SNPs in cancer-related pathway genes selected from RegulomeDB were evaluated. These SNPs were analyzed in a discovery set (n ¼ 354), and a replication study was performed in an independent set (n ¼ 772). The association of the SNPs with overall survival (OS) and disease-free survival (DFS) were analyzed. Results: In the discovery set, 76 SNPs were significantly associated with OS or DFS. Among the 76 SNPs, the association was consistently observed for 5 SNPs (ERCC1 rs2298881C>A, BRCA2 rs3092989G>A, NELFE rs440454C>T, PPP2R4 rs2541164G>A, and LTBP4 rs3786527G>A) in the validation set. In combined analysis, ERCC1 rs2298881C>A, BRCA2 rs3092989, NELFE rs440454C>T, and PPP2R4 rs2541164G>A were significantly associated with OS and DFS (adjusted HR aHR for OS ¼ 1.46, 0.62, 078, and 0.76, respectively; P ¼ 0.003, 0.002, 0.007, and 0.003 respectively; and aHR for DFS ¼ 1.27, 0.69, 0.86, and 0.82, respectively; P ¼ 0.02, 0.002, 0.03, and 0.008, respectively). The LTBP4 rs3786527G>A was significantly associated with better OS (aHR ¼ 0.75; P ¼ 0.003). Conclusions: Our results suggest that five SNPs in the cancer-related pathway genes may be useful for the prediction of the prognosis in patients with surgically resected NSCLC. Legal entity responsible for the study: Jae Yong Park Funding: None Disclosure: All authors have declared no conflicts of interest. Keywords: Polymorphism, survival, RegulomeDB, lung cancer

1281P

Relevance between PD-L1 and radiological invasiveness in pathological stage I lung adenocarcinoma

G. Toyokawa, K. Takada, T. Okamoto, Y. Kozuma, T. Matsubara, S. Takamori, T. Akamine, M. Katsura, F. Shoji, Y. Maehara Department of Surgery and Science, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan Background: Programmed death-ligand 1 (PD-L1) was reported to predict the response of immunotherapy; however, the association between PD-L1 expression and radiological/pathological features has yet to be elucidated.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Methods: A total of 292 patients with resected pathological stage I adenocarcinoma were analyzed for PD-L1 expression by immunohistochemistry and evaluated to determine the association between PD-L1 expression and the radiological/pathological invasiveness. Specifically, the radiological invasiveness and noninvasiveness were determined based on the consolidation/tumor (C/T) ratio, with a cut-off value of 0.25 by thin-section computed tomography. Results: Among 292 patients, 47 (16.1%) were positive for PD-L1 expression; the remaining 245 patients (83.9%) were negative for PD-L1 expression. Fisher’s exact test demonstrated that PD-L1 expression was significantly associated with a higher C/T ratio (P¼0.029) and higher maximum standardized uptake value (SUVmax; P¼0.004). The mean values of C/T ratio and SUVmax in patients with and without PD-L1 expression were 0.84560.052 and 7.24160.795, and 0.60760.023 and 3.6060.364, respectively (P2273.3 >2273.3

158.2 172.80 >2273.3 >2273.3 479.3 321.9 225.2 2201.3 1290.8 50.3 30.0 >2273.3

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx381 | 455

abstracts

Annals of Oncology

0.63-0.85) for all lung cancer cases. The CNA score was a significant prognostic factor in univariable (HR 1.22 (95% CI 1.05-1.42 p ¼ 0.008) but not multivariable analyses (HR 0.91 (95% CI 0.73-1.15 p ¼ 0.45). Conclusions: Our preliminary results demonstrate non-invasive detection of tumour derived copy number alterations with low coverage whole genome sequencing. The CNA score is not recommended as a stand-alone test to aid early lung cancer detection. Legal entity responsible for the study: Fiona Taylor Funding: Cancer Research UK and Yorkshire Cancer Research UK and The Roy Castle Foundation UK Disclosure: All authors have declared no conflicts of interest. Keywords: Cell-free DNA, early detection, lung cancer. 1284P

HOXA-related long non-coding RNAs impact prognosis in early stage NSCLC patients

A. Navarro1, S. Santasusagna1, N. Vinolas Segarra2, J.J. Castellano1, J. Moises3, noz1, J. Ramırez4, R.M. Marrades3, L. Molins5, M. Monzo1 S. Morales1, J. Canals1, C. Mu~ 1 Molecular Oncology and Embryology Laboratory, Human Anatomy Unit, University of Barcelona, Barcelona, Spain, 2Medical Oncology, Hospital Clinic y Provincial de Barcelona, Barcelona, Spain, 3Pneumology, Hospital Clinic y Provincial de Barcelona, Barcelona, Spain, 4Pathology, Hospital Clinic y Provincial de Barcelona, Barcelona, Spain, 5Thoracic Surgery, Hospital Clinic y Provincial de Barcelona, Barcelona, Spain Background: HOX genes, grouped in four clusters in humans (HOXA-D), encode for transcription factors that are master regulators of embryonic development and oncogenesis. HOXA cluster, located in chromosome 7, play a critical role in the patterning of tissues with mesodermal components such as lungs. Long non-coding RNAs (lncRNAs) are especially abundant in the HOXA cluster (HOXA10-AS, HOXAS2, HOXAS3, HOXA11AS, HOTAIRM1 and HOTTIP). We aimed to evaluate the prognostic role of HOXArelated lncRNAs in early stage non-small cell lung cancer (NSCLC). Methods: 100 early stage NSCLC patients resected in our center from June 2007 to November 2013 were studied. Patient characteristics: median age, 65 (32-84); 78% males; 74% stage I; 56% ADK; 23% received adjuvant treatment. With a mean followup of 28.7 months, 31% relapsed. As validation data set 200 NSCLC patients from TCGA Research Network were used (RNAseq data). Only stage I-II TCGA samples without prior malignancy or synchronous cancer that not received neoadjuvant treatment were included. Statistical analysis was performed using R and TANRIC. Results: HOTTIP and HOXA11AS impacted prognosis in our cohort of patients. HOTTIP was expressed in all samples and patients with high levels of HOTTIP had shorter TTR (78.3 vs 58 months; p ¼ 0.048) and shorter OS (81.2 vs 61 months; p ¼ 0.023). HOTTIP was overexpressed in SCC (p ¼ 0.007) and in smokers (p ¼ 0.018). HOXA11AS was expressed only in 9% of patients but the patients expressing HOXA11AS had shorter TTR (73.5 vs 32 months; p ¼ 0.002). In the multivariate analysis HOXA11AS emerged as an independent prognostic marker for TTR (OR: 3.13, 95%CI: 1.3-7.3; p ¼ 0.009), while HOTTIP (OR: 2.357, 95%CI: 1.1-5.2; p ¼ 0.036) and age>65 (p ¼ 0.022) for OS. In the validation data set HOXA11AS was validated as prognostic marker (p ¼ 0.019). HOXA11AS expression correlated positively with development genes HOXA11, HOXA13, HOXA10, HOXA9, HOXA3, FOXD1, ZIC5 and miR-196b (HOXA cluster miRNA) and negatively with surfactant metabolism genes SFTPB and NAPSA or let-7a(p < 0.001). Interestingly, the high expression was associated to patients harboring RTN1 mutations (p < 0.0001). Conclusions: HOXA11AS expression in early stage NSCLC patients is a high-risk relapse marker. Legal entity responsible for the study: University of Barcelona, Barcelona, Spain Funding: None Disclosure: All authors have declared no conflicts of interest. Keywords: HOTTIP, NSCLC, lncRNA, HOXA11AS 1285P

Association between polymorphisms in microRNA target sites and survival in early-stage non-small cell lung cancer

J.H. Lee1, S.S. Yoo2, M.J. Hong3, J.E. Choi3, S.Y. Lee2, J.Y. Park2 Biochemistry and Cell Biology, Kyungpook National University School of Medicine, Daegu, Republic of Korea, 2Departments of Internal Medicine, Kyungpook National University, Daegu, Republic of Korea, 3Cell and Matrix Research Institute, Kyungpook National University School of Medicine, Daegu, Republic of Korea

1

Background: MicroRNAs (miRNAs) are small non-coding RNAs that function in regulation of gene expression. Recent studies have also suggested that single nucleotide polymorphisms (SNPs) located in miRNA target sites can influence the prognosis of diverse human cancers, including lung cancer. This study was conducted to evaluate the associations between single nucleotide polymorphisms (SNPs) in miRNA target sites using CLASH data and the survival outcomes of early-stage non-small cell lung cancer (NSCLC) patients. Methods: 100 potentially functional polymorphisms were selected based on cancerrelated miRNA target site in PolymiRTS database 3.0(http://compbio.uthsc.edu/ miRSNP), CLASH data, and CancerGenes database (http://cbio.mskcc.org/cancer V genes). All polymorphisms were genotyped using SEQUENOM’s MassARRAY iPLEX assay according to instructions of the manufacturer. The genotype association with

overall survival (OS) and disease-free survival (DFS) in 782 patients with NSCLC who underwent curative surgical resection were analyzed. Results: Among the 100 SNPs studied, two SNPs showed significant association with survival outcomes. Patients carrying the POLR2A rs2071504TT or CT genotypes showed significantly lower overall survival and disease-free survival than those with the POLR2A rs2071504CC genotype (HR ¼ 1.42, 95% CI ¼ 1.08–1.88, P ¼ 0.01 and HR ¼ 1.34, 95% CI ¼ 1.08–1.67, P ¼ 0.01, respectively). The NR2F6 rs2288539C>T variant was found to be significantly associated with higher overall survival under the recessive model (HR ¼ 0.13, 95% CI ¼ 0.02–0.90, P ¼ 0.04). Conclusions: Our findings suggest that the POLR2A rs2071504C>T and NR2F6 rs2288539C>T can influence the prognosis of early-stage NSCLC patients. Legal entity responsible for the study: Jae Yong Park Funding: None Disclosure: All authors have declared no conflicts of interest. Keywords: miRNA target sites, polymorphisms, survival outcome, non-small cell lung cancer

1286TiP

Neo-adjuvant chemo/immunotherapy for the treatment of resectable stage IIIA non-small cell lung cancer (NSCLC): A phase II multicenter exploratory study” NADIM trial

V. Calvo de Juan1, A. Insa Molla2, M. Cobo Dols3, B. Massuti Sureda4, G. LopezVivanco5, J. Casal Rubio6, J. de Castro7, M. Majem Tarruella8, R. Bernabe Caro9, J. Gonzalez-Larriba10, I.C. Barneto Aranda11, E. Nadal12, A. Martinez Marti13, N. Vinolas Segarra14, M. Guillot Morales15, D. Vicente Baz16, C. Camps Herrero17, M. Domine Gomez18, D. Rodriguez Abreu19, M. Provencio Pulla1 1 Medical Oncology, Hospital Puerta de Hierro Majadahonda, Majadahonda, Spain, 2 Medical Oncology, Hospital Clinico Universitario de Valencia, Valencia, Spain, 3Medical Oncology, Hospital Universitario Carlos Haya, Malaga, Spain, 4Medical Oncology, Hospital General Universitario de Alicante, Alicante, Spain, 5Medical Oncology, Hospital Cruces, Bizcaia, Spain, 6Medical Oncology, Complejo Hospitalario Universitario de Vigo. Alvaro Cunqueiro, Vigo, Spain, 7Department of Translational Oncology, University Hospital, Madrid, Spain, 8Medical Oncology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, 9Medical Oncology, Hospital Universitario Virgen del Rocio, Seville, Spain, 10Medical Oncology, Hospital Clinico Universitario San Carlos, Madrid, Spain, 11 Medical Oncology, University Hospital Reina Soﬁa, Cordoba, Spain, 12Medical Oncology, Catalan Institute of Oncology, L’Hospitalet, Spain, 13Medical Oncology, Vall dHebron University Hospital/Vall dHebron Institute Oncology, Barcelona, Spain, 14 Medical Oncology, Hospital Clinic y Provincial de Barcelona, Barcelona, Spain, 15 Oncology Department, Hospital Universitario Son Espases, Palma de Mallorca, Spain, 16 Medical Oncology, Hospital Universitario Virgen Macarena, Seville, Spain, 17Medical oncology, Hospital General Universitario Valencia, Valencia, Spain, 18Medical Oncology, University Hospital “Fundacion Jimenez Diaz”, Madrid, Spain, 19Medical Oncology, Hospital Universitario Gran Canaria, Las Palmas De Gran Canaria, Spain Background: Lung cancer is the primary cause of cancer mortality in western countries. The cure is unlikely in patients with NSCLC and locally advanced stage who are not surgical candidates, with a 3-year survival rate of 27% in those patients receiving chemotherapy and concomitant radiotherapy. On the contrary, in localized stages (stage I, II, IIIA) with surgical resection and cytostatic therapy, a survival of 5 years of 51% is achieved. Currently, there is no consensus on the best standard treatment: the surgical management of stage IIIA NSCLC remains highly controversial and most patients with stage IIIB disease are generally considered inoperable. Since distant metastases remain the major site of failure, it is likely that more effective cytotoxic or other anti-tumor agents will be required. Chemotherapy stimulates an immune response against tumors, which may facilitate immunotherapy anticancer activity. Evidence of synergy between chemotherapy and immunotherapy was shown in several studies. Trial design: Phase II, single-arm, open-label multicenter study that assesses feasibility, safety and efficacy of combined neoadjuvant therapy with Nivolumab 360 mg þ Paclitaxel 200mg/m2 þ Carboplatin AUC 6 Q3W, three cycles, in resectable stage IIIA NSCLC patients followed by adjuvant treatment for 1 year with Nivolumab 240 mg Q2W for 4 months and Nivolumab 480mg Q4W for 8 months. The primary endpoint will be Progression Free Survival at 24 months from diagnosis and to assess the efficacy of the combination. The secondary endpoints will be time to progression and overall survival at 3 years, response rate, toxicity profile of the combination, the down-staging rate and complete resection rate. Also, surgical outcome and complications will be assessed. Perform correlatives studies with the objectives of exploring the expression of other biomarkers, such as PD-L1, in tumor tissue, free DNA and circulating tumor cells in liquid biopsy. Describe whether PD-L1 expression is a predictive biomarker for ORR, describe PFS in PD-L1 þ (1%) population and report imaging response versus pathological response rate. Clinical trial identification: EudraCT Number: 2016-003732-20 Legal entity responsible for the study: Spanish Lung Cancer Group Funding: Bristol-Myers Squibb Disclosure: All authors have declared no conflicts of interest. Keywords: Neadjuvant chemotherapy, NSCLC.

R

456 | NSCLC, early stage

Volume 28 | Supplement 5 | September 2017

Annals of Oncology 28 (Supplement 5): v457–v459, 2017 doi:10.1093/annonc/mdx379

NSCLC, LOCALLY ADVANCED 1287PD

Preoperative chemotherapy and radiotherapy concomitant to cetuximab in stage IIIB NSCLC: A multicenter phase II SAKK

A. Curioni-Fontecedro1, H-B. Ris2, A. Xyrafas3, H. Bouchaab4, H. Gelpke5, N. Mach6, O. Matzinger7, N. Stojcheva8, M. Frueh9, R. Cathomas10, S. Berardi Vilei8, L. Bubendorf11, M. Pless12, D. Betticher13, S. Peters14 1 Department of Oncology, University Hospital Zurich, Zurich, Switzerland, 2Thoracic Surgery, Centre Hospitalier Universitaire Vaudois - CHUV, Lausanne, Switzerland, 3 Statistics Unit, SAKK - Swiss Group for Clinical Cancer Research, Bern, Switzerland, 4 Radiation Oncology, Centre Hospitalier Universitaire Vaudois - CHUV, Lausanne, Switzerland, 5Surgery, Kantonsspital Winterthur, Winterthur, Switzerland, 6 Oncology, Hoˆpitaux Universitaires de Gene`ve, Geneva, Switzerland, 7Radiation Oncology, Hoˆpital Riviera-Chablais of Vaud-Valais, Vevey, Switzerland, 8Clinical Project Management Unit, SAKK - Swiss Group for Clinical Cancer Research, Bern, Switzerland, 9 Medical Oncologz, Kantonsspital St. Gallen, St. Gallen, Switzerland, 10Medical Oncology/Hematology, Kantonsspital Graubu¨nden, Chur, Switzerland, 11Pathology, Universit€atsspital Basel, Basel, Switzerland, 12Medical Oncology, Kantonsspital Winterthur, Winterthur, Switzerland, 13Oncology, HFR Freiburg – Kantonsspital, Fribourg, Switzerland, 14Oncology, Centre Hospitalier Universitaire Vaudois - CHUV, Lausanne, Switzerland Background: Stage IIIB NSCLC treatment is definitive chemo-radiotherapy (CRT). In our SAKK 16/01 trial, neoadjuvant chemotherapy (CT) followed by neoadjuvant accelerated radiotherapy (RT) and surgery showed a median survival of 28.7 months (mos) in selected stage IIIB patients (pts). These promising results are the rationale for the current trimodality concept, introducing concomitant cetuximab (CET) to neoadjuvant CRT. Methods: Pts with pathologically proven resectable stage IIIB (T4N0-3M0 or T14N3M0, 6th TNM) NSCLC, PS 0-1, and adequate organ function were treated with 3 cycles of neoadjuvant CT (cisplatin 100 mg/m2 and docetaxel 85 mg/m2 d1, q3w) followed by accelerated concomitant boost RT (44 Gy in 22 fractions in 3 weeks), both with concomitant weekly CET (250mg/m2) and subsequent surgery. The primary endpoint was progression-free survival (PFS) at 1 year (yr). Results: 69 pts were treated in 11 Swiss centers. 2/3 were men, median age was 60 yrs. Histology was squamous in 41% and adenocarcinoma in 49%, with T4 disease in 61%, N3 in 46% and both in 7%. A median relative total dose intensity of 99% of CT and 91% of CET was delivered. Per protocol RT was delivered to 95% of pts. 57 (83%) pts underwent surgery, with complete resection (R0) in 74% and a postoperative 30d mortality of 4%. Response rate after CT-immunotherapy was 57% and 64% after CRT-immunotherapy (CRT-I). Major pathologic response was found in 36% of the resected pts. 1-yr PFS based on Kaplan-Meier estimation was 50% (95% CI: 37%-62%). Median PFS was 12 mos (95% CI: 9-16), median OS was 21 mos (95% CI: 14-25), with a 2 and 3-yr survival of 41% and 30%, respectively. Conclusions: This is one of the largest prospective phase II trials to evaluate the role of induction CRT-I and surgery in resectable stage IIIB disease, and the first to associate concurrent CET to the neoadjuvant strategy. Trial treatment is feasible with excellent adherence to the protocol and promising clinical and pathological response rates, PFS and OS, supporting an aggressive approach including surgery in selected IIIB pts. As compared to our previous SAKK 16/01 experience, the addition of CET does not improve the outcome of this group of locally advanced NSCLC pts. Clinical trial identification: SAKK 16/08, NCT01059188, original version dated: 24.08.2009, including amendments: 04.07.2013. Legal entity responsible for the study: Swiss Group for Clinical Cancer Research (SAKK) Funding: Merck Serono Disclosure: All authors have declared no conflicts of interest.

1288PD

Randomized phase ii trial comparing chemoradiotherapy with chemotherapy for completely resected unsuspected N2-positive non-small cell lung cancer

J-M. Sun1, M-J. Ahn1, J.S. Ahn1, H.K. Kim2, Y.C. Ahn3, J.I. Zo2, Y.M. Shim2, K. Park1 Medicine, Samsung Medical Center Sungkyunkwan University School of Medicine, Seoul, Republic of Korea, 2Thoracic and Cardiovascular Surgery, Samsung Medical Center Sungkyunkwan University School of Medicine, Seoul, Republic of Korea, 3 Radiation Oncology, Samsung Medical Center Sungkyunkwan University School of Medicine, Seoul, Republic of Korea

1

Background: We investigated whether concurrent chemoradiotherapy (CCRT) would increase survival in patients with completely resected unsuspected N2-

C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

positive non-small cell lung cancer (NSCLC), compared with adjuvant chemotherapy alone. Methods: Eligible patients were randomly assigned (1:1 ratio) to either the CCRT arm or the chemotherapy arm. In the CCRT arm, patients received concurrent thoracic radiotherapy (50 Gy in 25 fractions) with five cycles of weekly paclitaxel (50 mg/m2) and cisplatin (25 mg/m2), followed by two additional cycles of paclitaxel (175 mg/m 2) plus cisplatin (80 mg/m2) at three-week intervals. In the chemotherapy arm, patients received four cycles of adjuvant paclitaxel (175 mg/m 2) and carboplatin (AUC 5.5) every three weeks. The primary endpoint was disease-free survival. Results: We enrolled and analyzed 101 patients. The median disease-free survival of the CCRT arm was 24.7 months, which was not significantly different from that of the chemotherapy arm (21.9 months; hazard ratio [HR] 0.94, 95% CI: 0.58–1.52, P ¼ 0.40). There was no difference in overall survival (CCRT: 74.3 months, chemotherapy: 83.5 months, HR: 1.33, 95% CI: 0.71–2.49). Subgroup analysis showed chemotherapy alone increased overall survival in never-smokers and multi-station N2-positive patients. The pattern of disease recurrence was similar between the two arms. Conclusions: There was no survival benefit from adjuvant CCRT compared with platinum-based chemotherapy alone for completely resected unsuspected N2-positive NSCLC. However, the role of sequential radiotherapy administered after adjuvant chemotherapy is being evaluated, and further study is needed to evaluate the optimal radiotherapy approach for completely resected N2-positive NSCLC. Clinical trial identification: (NCT01066234) Legal entity responsible for the study: Keunchil Park Funding: None Disclosure: All authors have declared no conflicts of interest.

1289PD

Clinical course in patients with stage III non-small cell lung cancer and interstitial lung disease treated with chemoradiotherapy: a retrospective analysis in a single institute

H. Kobayashi1, T. Naito1, K. Omae2, S. Omori1, K. Nakashima1, K. Wakuda1, A. Ono1, H. Kenmotsu1, H. Murakami1, M. Endo3, H. Harada4, T. Takahashi1 1 Thoracic Oncology, Shizuoka Cancer Center, Shizuoka, Japan, 2Clinical Research Promotion Unit, Clinical Research Center, Shizuoka Cancer Center, Sunto-gun, Japan, 3 Diagnostic Radiology, Shizuoka Cancer Center, Shizuoka, Japan, 4Radiation Oncology, Shizuoka Cancer Center, Shizuoka, Japan Background: Patients with non-small cell lung cancer (NSCLC) and interstitial lung disease (ILD) are often excluded from clinical trials because they are considered to be at high risk for acute exacerbation (AE) of ILD triggered by radiotherapy. Therefore, data on the clinical course in these patients are limited. We examined the relationship between chemoradiotherapy (CRT) and occurrence of AE of ILD as well as the clinical course in these patients at our institution. Methods: A retrospective analysis was performed on 37 patients with stage III NSCLC and ILD treated with first-line CRT in a clinical setting between January 2009 and December 2014 at our institution. Results: Patient characteristics are shown in Table 1. Patients treated with CRT had milder ILD than those treated with chemotherapy alone. Eighteen patients treated with CRT received corticosteroids for the treatment of AE of ILD. Univariate analysis revealed that the risk factors for AE of ILD were ILD classification, smoking history, and V20. In multivariate logistic regression analysis, the independent risk factor for AE of ILD was ILD classification. AE of ILD occurred in nine (82%) patients with usual interstitial pneumonia (UIP) pattern and in nine (35%) with non-UIP pattern. Median overall survival (mOS) was 34.6 months. Univariate analysis and multivariate logistic regression analysis revealed that the prognostic factor for patients with stage III NSCLC and ILD treated with CRT was ILD classification. mOS was 10.9 and 43.0 months in UIP and non-UIP patterns, respectively. Conclusions: In conclusion, this retrospective analysis suggests that ILD classification (UIP or non-UIP) is associated with the occurrence of AE of ILD and prognosis in patients treated with first-line CRT.

abstracts Table: 1289PD First-line treatment Variable

Age, median (range) Sex, male/female PS, 0/1/2 Clinical staging (TNM classiﬁcation, 7th edition), IIIA/IIIB Smoking history, yes/no Histology, squamous/non-squamous ILD classiﬁcation, UIP/non-UIP %VC, median (range) V20, median (range) Lung volume loss or honeycombing

Annals of Oncology

Chemoradiotherapy (N ¼ 37)

Radiotherapy (N ¼ 17)

Chemotherapy (N ¼ 25)

73 (52–85) 32/5 15/22/0 20/17 33/4 19/18 11/26 96 (59–129) 27 (12–35) 4

80 (59–89) 15/2 6/7/4 8/9 16/1 7/10 6/11 86 (62–112) 26 (14–36) 4

69 (58–81) 21/4 10/15/0 10/15 24/1 13/12 16/9 85 (64–121)

Clinical trial identification: The study protocol was approved by the Institutional Review board of Shizuoka Cancer Center (28-J167-28-1-3). Legal entity responsible for the study: Haruki Kobayashi Funding: None Disclosure: All authors have declared no conflicts of interest.

1290P

DNA repair gene expression in bronchial washing fluid as new molecular tool for clinical outcome decision

D. Schveigert1, R. Askinis2, J. Fadejeva1, V. Sapoka3, A. Krasauskas2, S. Cicenas2 Laboratory of Molecular Oncology, National Cancer Institute, Vilnius, Lithuania, 2Department of Thoracic Surgery and Oncology, National Cancer Institute, Vilnius, Lithuania, 3Clinics of Internal Diseases, Vilnius University hospital Santariskiu Klinikos, Vilnius, Lithuania

1

Background: Platinum-based drugs (cisplatin, etc.) are used as a first-line therapy for NSCLC patients. However, such treatment is not effective for all patients. Biomarkers that could predict efficiency of the platinum-based treatment should be identified. Aim: To evaluate whether the response to treatment of NSCLC patients is based on ERCC1 and RRM1 gene expression in bronchial washing fluid. Methods: 70 patients with a first-time diagnosed NSCLC receiving CisplatinþEtoposid were involved in the study. RNA was extracted from bronchial washing fluid using “RNeasy Plus Mini Kit” (QIAGEN, Germany). The analysis of ERCC1 and RRM1 expression was done by qRT-PCR method. A v2 test was used to analyze gene expression in relation to clinicopathological parameters. The survival rates were calculated by the Kaplan-Meier method. The prognostic significance was assessed by the Cox proportional hazards regression model. Results: Statistically significant differences were found between ERCC1 expression and tumour differentiation grade, RRM1 expression and disease stage and lymph node status. ERCC1 expression was associated with NSCLC patient progression-free survival (PFS) rate depending on gender, disease stage, response to treatment. Patients from high ERCC1 expression group had 7.6 months longer survival than patients from low expression group. RRM1 expression was associated with NSCLC patients PFS rates depending on gender, age, tumour histology and differentiation grade. Patients from low RRM1 expression group had 7.9 months longer survival than those from high expression group. Multivariate analysis of factors influencing PFS rate showed that disease stage (p ¼ 0.01), tumor differentiation grade (p ¼ 0.009), response to treatment (p ¼ 0.02) and RRM1 expression (p ¼ 0.001) were independent prognostic factors of NSCLC patients PFS. Conclusions:ERCC1 and RRM1 genes may influence platinum-based chemotherapy treatment of NSCLC patients. In order to improve the effectiveness of treatment it is appropriate to identify RRM1 expression changes in the bronchial washing fluid. Therefore, NSCLC patients with high RRM1 expression should be actively followed-up because of quicker disease progression. Clinical trial identification: Lithuanian Bioethics Committee No. 158200-09-381-104 Legal entity responsible for the study: National Cancer Institute Funding: None Disclosure: All authors have declared no conflicts of interest. Keyword: NSCLC, DNA repair genes, bronchial washing fluid. 1291P

Diagnosis and monitoring of non-small cell ung cancer patients by next generation sequencing and droplet digital PCR on circulating tumor DNA

P. Vannuffel, C. De Rop Molecular Biology, Institut de Pathologie et de Ge´ne´tique, Gosselies, Belgium Background: About 80% to 85% of lung cancers are non-small cell lung cancer (NSCLC). EGFR tyrosine kinase inhibitors as well as several other targeting molecules

458 | NSCLC, locally advanced

13

have been demonstrated to be effective in treating patients with activating mutations. We investigated the use of circulating tumor DNA (ctDNA) and high sensitive detection techniques for mutational profiling to improve the diagnosis and monitoring of NSCLC patients. Methods: ctDNA was extracted from plasma using the QIAamp Circulating Nucleic Acid kit (Qiagen). A custom panel was designed to cover EGFR, KRAS, NRAS, BRAF, PIK3CA, DDR2, AKT1, PTEN, MEK1 and ERBB2 hotspot mutations. Libraries, constructed according to the AmpliSeq protocol, were sequenced on the semiconductor Ion Torrent S5XL platform. The presence of the EGFR T790M mutation was also assessed by a digital PCR assay. Results: A total of 120 patients from 30 Belgian institutions were enrolled in this prospective study. The majority of patients presented with stage IV adenocarcinoma and progression. Forty-six (46) patients had a mutation detected on a former biopsy: EGFR exon 19 (26), EGFR exon 21 (8), KRAS (10), PIK3CA (1) and ERBB2 (1). Among those patients, 28 (61%) harbored the same mutation when their ctDNA was sequenced with our NGS panel: EGFR exon 19 (15), EGFR exon 21 (6), KRAS (5), PIK3CA (1) and ERBB2 (1). For 7 patients, for which no mutation had not been previously detected, 4 EGFR, 2 KRAS and 1 NRAS mutations were found after ctDNA analysis. As far as the ddPCR detection of EGFR T790M was concerned, the mutation was detected on 7 (21%) of the 34 patients presenting EGFR mutations in their prior biopsy (5 in exon 19 and 2 in exon 21). Patients with acquired T790M mutation were previously treated by Afatinib (3), Erlotinib (2) or Gefitinib (1). Conclusions: Our results indicate that ctDNA can be an alternative and noninvasive source of tumor DNA, a surrogate to classical biopsies, particularly when access to tumor tissue is limited. NGS and ddPCR assays are sensitive enough to promote a clinical translation of ctDNA analysis into disease management and therapeutic decision. Supported by a grant from Boehringer Ingelheim. Legal entity responsible for the study: Institut de Pathologie et de Ge´ne´tique Funding: Boehringer Ingelheim Institut de Pathologie et de Ge´ne´tique Disclosure: All authors have declared no conflicts of interest. 1292P

Safety data from randomized phase II study of cisplatin (CDDP)1S-1 versus CDDP1pemetrexed (PEM) combined with thoracic radiotherapy (TRT) for locally advanced non-squamous (non-sq) nonsmall cell lung cancer (NSCLC): SPECTRA study

S. Niho1, T. Yoshida2, T. Akimoto3, K. Sakamaki4, T. Takahashi5, T. Seto6, M. Nishio7, N. Yamamoto8, T. Hida9, H. Okamoto10, T. Kurata11, M. Satouchi12, K. Goto1, T. Yamanaka4, Y. Ohe8 1 Department of Thoracic Oncology, National Cancer Center Hospital East, Kashiwa, Japan, 2Department of Thoracic Oncology, Aichi Cancer Center Hospital, Aichi, Japan, 3Department of Radiation Oncology, National Cancer Center Hospital East, Kashiwa, Japan, 4Department of Biostatistics, Yokohama City University School of Medicine, Yokohama, Japan, 5Division of Thoracic Oncology, Shizuoka Cancer Center, Shizuoka, Japan, 6Department of Thoracic Oncology, National Kyushu Cancer Center, Fukuoka, Japan, 7Department of Thoracic Medical Oncology, The Cancer Institute Hospital of Japanese Foundation for Cancer Research, Tokyo, Japan, 8Department of Thoracic Oncology, National Cancer Center Hospital, Tokyo, Japan, 9Department of Thoracic Oncology, Aichi Cancer Center, Nagoya, Japan, 10Department of Respiratory Medicine and Medical Oncology, Yokohama Municipal Citizen’s Hospital, Yokohama, Japan, 11First Department of Internal Medicine, Kansai Medical University, Hirakata, Japan, 12Department of Thoracic Oncology, Hyogo Cancer Center, Akashi, Japan Background: Both CDDPþS-1 and CDDPþPEM could be given at full systemic doses with TRT in locally advanced NSCLC, and CDDPþPEM is one of the standard chemotherapy regimens in patients with advanced non-sq NSCLC. This multicenter, randomized, open-label, phase II study (SPECTRA) compared the efficacy and safety of the two above-mentioned promising regimens combined with TRT in patients with unresectable locally advanced non-sq NSCLC.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Methods: Patients were randomly assigned to receive CDDPþS-1 (CDDP 60mg/m2, d1, and S-1 80mg/m2, d1-14, q4w, up to 4 cycles) or CDDPþPEM (CDDP 75mg/m2, d1, and PEM 500mg/m2, d1, q3w, up to 4 cycles) combined with TRT 60Gy in 30 fractions. The primary endpoint was 2-year progression-free survival (PFS) rate. The sample size was set at 100 patients. Results: Between Jan 2013 and Oct 2016, 102 patients were enrolled in this study from 9 institutions in Japan. All 102 patients were eligible and assessable, of whom 52 were assigned to CDDPþS-1 and 50 to CDDPþPEM. Baseline characteristics were similar (CDDPþS-1/ CDDPþPEM): median age (range) 64.5 (39-73)/63.5 (32-74) years; women, n ¼ 17 (33%)/n¼17 (34%); stage IIIB, n ¼ 21 (40%)/n¼20 (40%); ECOG PS of 1, n ¼ 14 (27%)/ n¼14 (28%); never smoker, n ¼ 12 (23%)/n¼12 (24%); and adenocarcinoma, n ¼ 47(90%)/n¼45(90%). Completion rate of TRT (60Gy) and chemotherapy (4 cycles) was 92%/98% and 73%/86%, respectively. Response rate was 60%/64%. Grade 3 toxicities included febrile neutropenia (12%/2%), anorexia (8%/16%), diarrhea (8%/0%), esophagitis (6%/8%), pneumonia (4%/4%), neutropenia (35%/50%), anemia (8%/12%), thrombocytopenia (4%/6%), and hyponatremia (12%/12%). Grade 2 radiation pneumonitis was observed in 8 (15%)/2 (4%) patients. No treatment-related death was observed. The data on PFS and overall survival is immature. Conclusions: Response rate was similar between two arms. Toxicities were tolerable and manageable in both arms; however febrile neutropenia was more frequently observed in the CDDPþS-1 arm. Survival data will be analyzed in late 2018. Clinical trial identification: UMIN000009914 (release date: 31/Jan/2013) Legal entity responsible for the study: Yuichiro Ohe Funding: Japan Agency for Medical Research and Development Disclosure: S. Niho, T. Seto: Received honoraria from Taiho and Eli Lilly and research funding from Eli Lilly. K. Sakamaki, T. Yamanaka: Received honoraria from Taiho. T. Takahashi: Received research funding from Taiho and Eli Lilly and honoraria from Eli Lilly. M. Nishio, T. Hida, H. Okamoto, M. Satouchi, K. Goto, Y. Ohe: Received research funding and honoraria from Taiho and Eli Lilly. N. Yamamoto: Received research funding from Taiho. T. Kurata: Received research funding and honoraria from Eli Lilly. All other authors have declared no conflicts of interest.

1293P

Preliminar analysis of the Spanish Lung Cancer Group (SLCG) phase II trial of concurrent chemo-radiotherapy (CT-RT) with cisplatin (P) plus metronomic oral vinorelbine (mOV) for unresectable locally advanced non-small cell lung cancer (LA-NSCLC): NORA trial (GECP 15/02)

M. Majem Tarruella1, L. Isla Casado2, M. Guirado3, B. Massuti Sureda4, A.L. Ortega nas8, M.T. Moran Granados5, R. Marse Fabregat6, M. Domine Gomez7, R. de las Pe~ Bueno9, S. Vazquez Estevez10, M.A. Sala Gonzalez11, J. Coves Sarto12, J.M. SanchezTorres13, D. Vicente Baz14, J. Gonzalez-Larriba15, A. Paredes16, N. Farre´17, L. Fernandez Fornos18, A. Mena19, M. Provencio Pulla20 1 Medical Oncology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, 2Medical Oncology, University Hospital Lozano Blesa, Zaragoza, Spain, 3Medical Oncology, Hospital General Universitario de Elche, Elche, Spain, 4Medical Oncology, Hospital General Universitario de Alicante, Alicante, Spain, 5Oncologıa Me´dica, Complejo Hospitalario de Jae´n, Jae´n, Spain, 6Medical Oncology, Hospital Universitario Son Espases, Palma de Mallorca, Spain, 7Medical Oncology, University Hospital “Fundacion Jimenez Diaz”, Madrid, Spain, 8Medical Oncology, Consorcio Hospitalario Provincial de Castellon, Castell on, Spain, 9Medical Oncology, Catalan Institute of Oncology (ICO Badalona), Hospital Germans Trias i Pujol, Badalona, Spain, 10Medical Oncology, Hospital Universitario Lucus Augusti, Lugo, Spain, 11Medical Oncology Department, atzer, Palma de Hospital de Basurto, Bilbao, Spain, 12Medical Oncology, Hospital Son Ll Mallorca, Spain, 13Medical Oncology, Hospital Universitario de La Princesa, Madrid, Spain, 14Medical Oncology, Hospital Universitario Virgen Macarena, Seville, Spain, 15 Medical Oncology, Hospital Clinico Universitario San Carlos, Madrid, Spain, 16Medical Oncology, Hospital Universitario Donostia, San Sebasti an, Spain, 17Radiation Oncology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, 18Radiation Oncology, Hospital General Universitario de Elche, Elche, Spain, 19Radiation Oncology, Hospital Universitario Son Espases, Palma de Mallorca, Spain, 20Medical Oncology, Hospital Universitario Puerta de Hierro Majadahonda, Madrid, Spain Background: CT-RT is the standard treatment for unresectable LA-NSCLC. P plus vinorelbine is widely used. Metronomic CT is a frequent administration of low doses of CT. mOV has shown good efficacy and improved safety, and could improve the RT effect. Our goal is to evaluate the efficacy and safety of P-mOV with radical RT in patients (pts) with LA-NSCLC.

Volume 28 | Supplement 5 | September 2017

Methods: Pts aged 18-75 years with histologically proven untreated and unresectable LA-NSCLC, adequate bone marrow, hepatic & renal function, ECOG PS0-1, received P 80mg/m2 D1 every 3 weeks combined with mOV 50mg/day on days D1, 3 & 5/weekly, 2 cycles (cy) as induction; patients without progression received 2 more cy of P at the same dose with mOV 30mg/day on D1, 3 & 5/weekly, concurrently with RT (66Gy in 6.5weeks). Primary endpoint was progression-free survival (PFS) by RECIST v1.1; secondary endpoints were: overall response rate, overall survival and safety profile. To guarantee an overall type-1 a error no greater than 0.05 and a type II (b) error 0.1 for PFS, a sample size of 67 pts was planned. Results: Since May 2016, 58 pts have been recruited. Fifty-three pts have been included in the analysis. Pt characteristics: Male 72%; median age 63 (range 33-75); PS 0/1 62/48%; smokers 48%; adenocarcinoma/squamous 43.4/35.9%; stage IIIA/B 35.9/58.5%. Hematological G3-4 toxicities: neutropenia 18.9%; anemia 3.8%; febrile neutropenia 7.5%. Non-hematological G3-4 toxicities: esophagitis 1.9%; infection without neutropenia 1.9%; dyspnea 3.8%; thromboembolism 3.8%. No treatment-related deaths were reported. Conclusions: mOV-P administered with RT has a manageable safety profile. Based on this, accrual is ongoing. Clinical trial identification: EudraCT 2015-003312-21. Legal entity responsible for the study: Spanish Lung Cancer Group (SLCG) Funding: Spanish Lung Cancer Group (SLCG) Disclosure: All authors have declared no conflicts of interest.

1294P

Real world data of practice patterns and outcomes for pemetrexed plus platinum as neoadjuvant chemotherapy in adenocarcinomas of lung from a tertiary cancer center of India: Looking beyond the usual paradigm

A. Kapoor1, S. Zanwar1, A. Joshi1, V. Noronha1, V. Patil1, A. Chougule2, A. Mahajan3, P. Bhargava1, K. Prabhash1 1 Medical Oncology, Tata Memorial Hospital, Mumbai, India, 2Medical Oncology (Molecular Lab), Tata Memorial Hospital, Mumbai, India, 3Radiodiagnosis, Tata Memorial Hospital, Mumbai, India Background: Neoadjuvant chemotherapy (NACT) is the standard of care in non-small cell lung cancers (NSCLC) with locally advanced N2 disease. There is scarcity of data for pemetrexed-platinum regimen as NACT. Also, aside from N2 disease, role of NACT in locally advanced NSCLCs for tumor downstaging is unclear. Methods: Nonmetastatic adenocarcinomas of lung treated with pemetrexed-platinum based NACT were analysed. The patients with locoregionally advanced N2 disease and those who were borderline candidates for upfront definitive treatment were planned for NACT after discussion in a multidisciplinary clinic. Total 4 cycles of 3-weekly pemetrexed and platinum were delivered in combined neoadjuvant and adjuvant setting. Response assessment was done using RECIST criteria. Progression free (PFS) and overall survival (OS) were calculated using Kaplan Meier method. Results: Out of 114 evaluable patients, 99 patients received NACT with pemetrexedplatinum. Most common indication for NACT was N2 disease at baseline (46.4%). Objective response rate was 38.2% (95% CI: 23%-53%) including two complete and 34 partial responses. 12.7% patients had progressive disease on NACT. Median PFS was 15 months (95% CI 11.6-18.4) and median OS was 22 months (95% CI 15.6-28.3) at a median follow-up of 16 months. There was a significant improvement in the OS of patients undergoing definitive therapy versus no definitive therapy (median OS 25 months [95% CI 19.4-30.3] vs 12 months [95% CI 3.2-20.3] respectively; p ¼ 0.047, HR 1.6). Amongst patients who could not undergo definitive CTRT upfront due to dosimetric constraints (n ¼ 36), 26 (72.2%) patients finally underwent CTRT after NACT. Those patients who were not able to undergo definitive CTRT had median PFS of 5 months [95% CI 2.1-7.9] versus 10 months [95% CI 3.8-16.1] in those who were made amenable to definitive CTRT post NACT; p ¼ 0.018. Conclusions: Pemetrexed-platinum based NACT appears to be an effective option and many borderline cases where upfront definitive therapy is not feasible may become amenable to the same after incorporation of NACT. Clinical trial identification: Clinical Trials Registry India (registration number: CTRI/ 2013/01/003335) Legal entity responsible for the study: Tata Memorial Hospital, Mumbai Funding: None Disclosure: All authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx379 | 459

Annals of Oncology 28 (Supplement 5): v460–v496, 2017 doi:10.1093/annonc/mdx380

NSCLC, METASTATIC 1295O

Blood-based biomarkers for cancer immunotherapy: Tumor mutational burden in blood (bTMB) is associated with improved atezolizumab (atezo) efficacy in 2L1 NSCLC (POPLAR and OAK)

D.R. Gandara1, M. Kowanetz2, T.S.K. Mok3, A. Rittmeyer4, L. Fehrenbacher5, D. Fabrizio6, G. Otto7, C. Malboeuf8, D. Lieber7, S.M. Paul2, L. Amler2, T. Riehl9, E. Schleifman10, C.A. Cummings2, P.S. Hegde2, W. Zou11, A. Sandler12, M. Ballinger12, D.S. Shames10 1 Thoracic Oncology, UC Davis Comprehensive Cancer Center, Sacramento, CA, USA, 2 Oncology Biomarker Development, Genentech, South San Francisco, CA, USA, 3Clinical Oncology, Chinese University of Hong Kong, Hong Kong, China, 4Thoracic Oncology, Lungenfachklinik Immenhausen, Immenhausen, Germany, 5Oncology, Kaiser Permanente Medical Center, Vallejo, CA, USA, 6Cancer Immunotherapy, Foundation Medicine Inc., Cambridge, MA, USA, 7Product Development, Foundation Medicine Inc., Cambridge, MA, USA, 8Molecular Biology and Sequencing, Foundation Medicine Inc., Cambridge, MA, USA, 9Product Development Clinical (Oncology), Genentech, South San Francisco, CA, USA, 10Oncology Biology Development, Genentech, South San Francisco, CA, USA, 11Biostatistics, Genentech, South San Francisco, CA, USA, 12Product Development Oncology, Genentech, South San Francisco, CA, USA Background: Atezo (anti–PD-L1) was FDA approved for 2Lþ NSCLC based on results from the randomized OAK and POPLAR trials, with atezo showing superior efficacy vs docetaxel (doc). We previously showed that TMB in tissue correlates with atezo efficacy in 1Lþ NSCLC. To address the significant challenge of consistently obtaining sufficient tumor tissue for molecular testing, we developed a novel blood-based assay to measure bTMB. Here we analyzed plasma samples from OAK and POPLAR with the bTMB assay to correlate bTMB with atezo clinical activity. Methods: The biomarker evaluable population (BEP) included 211 pts in POPLAR (ITT¼287) and 583 pts in OAK (excludes pts with known EGFR/ALK mutations; ITT¼850), with blood samples available for targeted genomic sequencing. The bTMB assay interrogates single nucleotide variants (SNVs) in 394 genes from cell-free DNA in plasma and reports a score based on the number of high-confidence SNVs identified. The BEP was grouped by bTMB cut points based on the minimum number of SNVs present. Results: In POPLAR, improved PFS and OS HRs with atezo vs doc were observed at a range of bTMB cut points compared with the ITT and BEP. In OAK, PFS benefit with atezo vs doc was observed at bTMB cut points 10 compared with BEP. (Table) Importantly, bTMB did not correlate with PD-L1 expression (SP142 or 22C3). Conclusions: These exploratory analyses represent the first demonstration of a novel blood-based assay measuring bTMB that may predict atezo clinical efficacy in 2Lþ NSCLC. Thus, the bTMB assay may provide a non-invasive biomarker to identify pts who may derive clinical benefit from single agent checkpoint inhibition. Prospective studies using bTMB are currently ongoing in pts with 1L NSCLC (B-F1RST/BFAST).

Table: 1295O Clinical efficacy of atezo vs doc in bTMB subgroups POPLAR study ITT (N 5 287) OS HR (95% CI) PFS HR (95% CI) bTMB subgroup No. of patients OS HR PFS HR

OS HR (95% CI) PFS HR (95% CI) bTMB subgroup No. of patients OS HR PFS HR

0.73 (0.53, 0.99) 0.94 (0.72, 1.23) 10 16 96 63 0.59 0.56 0.68 0.57 OAK study ITT (N 5 850) 0.73 (0.62, 0.87) 0.95 (0.82, 1.10) 10 16 251 158 0.69 0.64 0.73 0.65

BEP (N 5 211) 0.68 (0.50, 0.93) 0.90 (0.68, 1.20) 20 42 0.51 0.58 BEP (N 5 583) 0.64 (0.53, 0.77) 0.87 (0.73, 1.04) 20 105 0.65 0.61

BEP, biomarker-evaluable population; bTMB, tumor mutational burden in blood; ITT, intention to treat.

C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

Clinical trial identification: NCT02008227; NCT01903993 Legal entity responsible for the study: F. Hoffmann - La Roche Ltd. Funding: F. Hoffmann - La Roche Ltd. Disclosure: D.R. Gandara: Consultant: Genentech Clinical trial grant: Genentech. M. Kowanetz, D.S. Shames: Genentech employee and Roche stock. T.S.K. Mok: Spt: AZ, BI, PFE, NV, SFJ, ROG, MSD, CLVS, BMS, Eisai, Taiho ROG/GNE, LLY, NV; Stock Sanomics AB: AZ, ROG/GNE, PFE, LLY, BI, CLVS, MSD, NV, SFJ, ACEA Bio, VRTX, BMS, GeneDecode, OGX, CELG, RXDX Bod: IASLC, CLCRF, CSCO, HKCTS. A. Rittmeyer: Grants as an advisor or speaker by: Astra Zeneca, BMS, Boehringer Ingelheim, Eli Lilly, Pfizer and Roche Genentech. D. Fabrizio, C. Malboeuf: Employee and stockholder of Foundation Medicine. G. Otto, D. Lieber: Employee and shareholder of Foundation Medicine. S.M. Paul: Genentech employee, Roche Stock, Travel, Accommodations, Expenses from Genentech. L. Amler, T. Riehl, E. Schleifman, C.A. Cummings, P.S. Hegde: Genentech employee. W. Zou: Genentech employee, Roche research funding. A. Sandler, M. Ballinger: Employee of Genentech, Roche stock. All other authors have declared no conflicts of interest.

1296O

Clinical efficacy of atezolizumab (Atezo) in PD-L1 subgroups defined by SP142 and 22C3 IHC assays in 2L1 NSCLC: Results from the randomized OAK study

S. Gadgeel1, M. Kowanetz2, W. Zou3, F.R. Hirsch4, K.M. Kerr5, D.R. Gandara6, F. Barlesi7, K. Park8, M. McCleland9, H. Koeppen10, M. Ballinger11, A. Sandler11, P.S. Hegde9, A. Rittmeyer12 1 Medical Oncoloy, University of Michigan, Ann Arbor, MI, USA, 2Oncology Biomarker Development, Genentech, South San Francisco, CA, USA, 3Biostatistics, Genentech, South San Francisco, CA, USA, 4Division of Medical Oncology, University of Colorado Anschutz Medical Campus, Aurora, CO, USA, 5Pathology, Aberdeen Royal Inﬁrmary/Aberdeen University Medical School, Aberdeen, UK, 6Thoracic Oncology, UC Davis Comprehensive Cancer Center, Sacramento, CA, USA, 7Multidisciplinary Oncology & Therapeutic Innovations, Aix Marseille University, Marseille, France, 8 Division of Heamatology/Oncology, Department of Medicine, Samsung Medical Center, Seoul, Republic of Korea, 9Oncology Biomarker Development, Genentech, South San Francisco, CA, USA, 10Research Pathology, Genentech, South San Francisco, CA, USA, 11Product Development Oncology, Genentech, South San Francisco, CA, USA, 12Thoracic Oncology, Lungenfachklinik Immenhausen, Immenhausen, Germany Background: In the Phase III OAK trial, patients (pts) with previously treated advanced NSCLC had improved median overall survival (OS) with atezo vs docetaxel (doc), regardless of PD-L1 expression (per VENTANA PD-L1 SP142 IHC assay). Although efficacy correlated with PD-L1 expression on tumor cells (TC) and tumorinfiltrating immune cells (IC), an OS benefit was also observed in pts with PD-L1– negative tumors (i.e., TC0 and IC0; HR, 0.75 [95% CI: 0.59, 0.96]). To determine whether these results were consistent across PD-L1 IHC assays, we assessed atezo efficacy in PD-L1 subgroups as defined by SP142 and Dako 22C3 pharmDx PD-L1 IHC assays. Methods: PD-L1 expression was assessed prospectively with SP142 and retrospectively with 22C3. The SP142 assay measured PD-L1 expression on TC and IC, while the 22C3 assay gave a tumor proportion score (TPS) based on TC membrane staining. Results: Among the primary population of 850 pts (ITT850), 400 had results from the 22C3 assay (biomarker-evaluable population [BEP]). Clinical outcomes in the BEP vs ITT850, and prevalence in PD-L1 subgroups are summarized (Table). Among pts with tumors negative by SP142 (TC0 and IC0), most (77%) were also negative by 22C3 (TPS < 1%). Comparable OS benefit with atezo was seen in PD-L1–negative subgroups defined by both assays. Improved clinical benefit was observed in pts with the highest PD-L1 expression by either assay (TC3 or IC3 by SP142, or TPS 50% by 22C3; Table). Conclusions: Prevalence of PD-L1 subgroups in the BEP was consistent with previous reports for both assays. Most tumors considered negative by SP142 were also negative by 22C3. An OS benefit (atezo vs doc) was observed in PD-L1–negative subgroups defined by either assay and was consistent with the overall population results from OAK. These data provide evidence of atezo OS benefit in pts with PD-L1–negative tumors irrespective of the PD-L1 IHC assay used.

abstracts

Annals of Oncology

Table: 1296O Clinical efficacy in OAK ITT850 and BEP populations ITT850 (N ¼ 850)

BEP (N ¼ 400)

OS HR (atezo vs doc) (95% CI) 0.73 (0.62, 0.87) 0.56 (0.44, 0.71) PFS HR (atezo vs doc) (95% CI) 0.95 (0.82, 1.10) 0.75 (0.61, 0.93) Prevalence of PD-L1 subgroups in OAK BEP (n 5 400) SP142 22C3 PD-L1 negative TC0 and 38% 55% IC0, or TPS < 1% PD-L1 positive TC1/2/3 or 62% 46% IC1/2/3, or TPS 1% PD-L1 high TC3 or IC3, or TPS 50% 18% 25% OS HR (atezo vs doc) in PD-L1 subgroups in OAK BEP (n 5 400) (95% CI) SP142 22C3 PD-L1 negative 0.55 0.61 TC0 and IC0, or TPS < 1% (0.37, 0.80) (0.45, 0.84) PD-L1 positive 0.58 0.51 TC1/2/3 or IC1/2/3, or TPS 1% (0.42, 0.78) (0.36, 0.73) PD-L1 high 0.37 0.49 TC3 or IC3, or TPS 50% (0.20, 0.66) (0.29, 0.80)

Clinical trial identification: NCT02008227 Legal entity responsible for the study: F. Hoffmann - La Roche Ltd. Funding: F. Hoffmann - La Roche Ltd. Disclosure: S. Gadgeel: Speaker’s bureau- Astra-Zeneca, Genentech/Roche Advisory Boards- Astra-Zeneca, Ariad, Pfizer, Bristol Myers- Squibb and Genentech/Roche. M. Kowanetz: Genentech employee and Roche stock. W. Zou: Genentech employee, Roche research funding. F.R. Hirsch: Scientific Advisory Boards: Genentech/Roche, AstraZeneca, Lilly, BMS HTG Research Grant (through University of Colorado): Bayer, BMS, Genentech/Roche. K.M. Kerr: Lecture honoraria and/or consultancy fees from Roche, AZ, BI, Lilly, Pfizer, Merck Serono, MSD, BMS. D.R. Gandara: Consultant: Genentech Clinical trial grant: Genentech. F. Barlesi: Honorarium from Roche. K. Park: Consulting/Advisory Role: Astellas Pharma, AstraZeneca, Boehringer Ingelheim, Clovis Oncology, Lilly, Hanmi, Kyowa Hakko Kirin, Novartis, Ono Pharmaceutical, Roche Speakers Bureau: Boehringer Ingelheim Research Funding: AstraZeneca. M. McCleland, P.S. Hegde, H. Koeppen: Genentech employee. M. Ballinger, A. Sandler: Employee of Genentech, Roche stock. A. Rittmeyer: Grants as an advisor or speaker by: Astra Zeneca, BMS, Boehringer Ingelheim, Eli Lilly, Pfizer and Roche Genentech.

1297O

Randomized results of fixed-duration (1-yr) vs continuous nivolumab in patients (pts) with advanced non-small cell lung cancer (NSCLC)

D.R. Spigel1, M. McLeod2, M.A. Hussein3, D.M. Waterhouse4, L. Einhorn5, L. Horn6, B. Creelan7, S. Babu8, N.B. Leighl9, F. Couture10, J. Chandler11, G. Goss12, G. Keogh13, E.B. Garon14, K.B. Blankstein15, D.B. Daniel16, M. Mohamed17, A. Li18, N. Aanur18, R. Jotte19 1 Medical Oncology, Sarah Cannon Research Institute/Tennessee Oncology, PLLC, Nashville, TN, USA, 2Medical Oncology, Florida Cancer Specialists, Leesburg, FL, USA, 3Oncology, Florida Cancer Specialists, Leesburg, FL, USA, 4Oncology/Hematology, OHC (Oncology Hematology Care), Cincinnati, OH, USA, 5Hematology/Oncology, Indiana University, Indianapolis, IN, USA, 6Thoracic Oncology, Vanderbilt Ingram Cancer Center, Nashville, TN, USA, 7Thoracic Oncology, H. Lee Mofﬁtt Cancer Center, Tampa, FL, USA, 8Oncology/ Hematology, Fort Wayne Medical Oncology and Hematology, Fort Wayne, IN, USA, 9 Department of Medical Oncology, The Princess Margaret Cancer Centre, Toronto, ON, Canada, 10Oncology/Hematology, CISSS Chaudie´ere-Appalaches, Levis, QC, Canada, 11 Oncology/Hematology, The West Clinic, P.C., Memphis, TN, USA, 12Medical Oncology, The Ottawa Hospital, University of Ottawa, Ottawa, ON, Canada, 13Oncology/Hematology, Charleston Hematology Oncology Associates, Charleston, USA, 14Oncology/Hematology, UCLA Medical Center, Santa Monica, CA, USA, 15Oncology/Hematology, Hunterdon Healthcare, Flemington, NJ, USA, 16Oncology, Tennessee Oncology, Chattanooga, TN, USA, 17 Oncology/Hematology, Cone Health Cancer Center at Wesley Long, Greensboro, NC, USA, 18 Oncology, Bristol-Myers Squibb, Princeton, NJ, USA, 19Oncology/Hematology, The US Oncology Network/Rocky Mountain Cancer Centers, Denver, CO, USA Background: Nivolumab, the anti-programmed death (PD)-1 antibody, has demonstrated durable responses and survival benefit in pts with advanced NSCLC, with some

Volume 28 | Supplement 5 | September 2017

pts continuing to derive benefit even after discontinuation of nivolumab (due to adverse events [AEs] or a stopping rule). This raises the question of whether continuous nivolumab treatment is necessary for long-term benefit. CheckMate 153 (NCT02066636), an ongoing phase IIIB/IV study conducted primarily in the community setting, is evaluating the clinical benefit of a fixed-duration (1 yr) of nivolumab treatment vs continuous treatment in pts with previously treated advanced NSCLC. Pts who remained on nivolumab treatment for 1 yr were randomized to either continue receiving treatment or to stop treatment. Methods: Pts with stage IIIB/IV NSCLC and 1 prior systemic therapy were enrolled and treated with nivolumab 3 mg/kg IV Q2W. The primary objective of the study overall was the incidence of high-grade (grade 3–5) select treatment-related AEs. Pts still on treatment at 1 yr were randomized 1:1 either to continue nivolumab until progressive disease, unacceptable toxicity, or withdrawal of consent (continuous-treatment arm), or to discontinue treatment, with the possibility of resuming treatment upon disease progression (fixed-duration arm). Prespecified exploratory objectives included safety and efficacy in the 2 randomized arms. Results: As of April 2016, 1375 pts were enrolled and treated; 218 pts were randomized after 1 yr of treatment to the continuous-treatment arm (n ¼ 111) or the fixed-duration arm (n ¼ 107). Of these 218 pts, 133 (61%) had received 2 prior therapies and 10 (5%) had baseline ECOG PS 2. Data from an upcoming database lock (at which time, the expected post-randomization follow-up 10.7 mo) will be presented for randomized pts and will include overall survival, progression-free survival, and safety. In addition, data from pts who were re-treated in the fixed-duration arm will be presented. Conclusions: The results from CheckMate 153 represent the first insights from a randomized trial evaluating the impact of stopping treatment with a PD-1/PD-L1 inhibitor at 1 yr vs continuing treatment in pts with advanced, previously treated NSCLC. Clinical trial identification: NCT02066636 Legal entity responsible for the study: Bristol-Myers Squibb Funding: Bristol-Myers Squibb Disclosure: D.R. Spigel: Served as a consultant or advisor for Genentech/Roche (Inst), Novartis (Inst), Celgene (Inst), Bristol-Myers Squibb (Inst), Lilly (Inst), AstraZeneca (Inst), Pfizer (Inst), Clovis Oncology (Inst), Boehringer Ingelheim (Inst); travel funding from Genentech/Roche, Novartis, Celgene, Bristol-Myers Squibb, Lilly, AstraZeneca, Pfizer, Clovis Oncology, Biodesix, Boehringer Ingelheim, Peregrine Pharmaceuticals; owns stock in Foundation Medicine, Illumina; received institutional research funding from Genentech/ Roche, Novartis, Celgene, Bristol-Myers Squibb, Lilly, AstraZeneca, Pfizer, Clovis Oncology, Boehringer Ingelheim, Peregrine Pharmaceuticals, Oncogenex, OncoMed, Amgen, Verastem, Daiichi Sankyo, University of Southwestern Medical Center – Simmons Cancer Center, Merck. D.M. Waterhouse: Served as consultant for BMS and Lilly; participated in speakers’ bureau for BMS, Celgene, Genentech/Roche, and Lilly. L. Einhorn: Served as a consultant for Celgene and ZIOPHARM Oncology; owns stock or other ownership interests with Amgen and Biogen Idec. L. Horn: Served as consultant for BMS, Merck, Bayer, Xcovery, GNE, BI, and Lilly; received honoraria for Biodesix; institution received research funding for AstraZeneca. B. Creelan: Participated on a speakers’ bureau for AstraZeneca and BMS; received research funding for Boehringer Ingelheim; travel funding from AstraZeneca, Merck Sharp & Dohme. S. Babu: Received research funding, consultancy fees, and travel honoraria from Alexion Pharmaceuticals. N.B. Leighl: Received honoraria from Pfizer; received institutional research funding from Novartis; travel funding from AstraZeneca and Merck, Sharp & Dohme. J. Chandler: Served as consultant for BMS; participated in speakers’ bureau for Janssen; received research funding from BMS, EMD Serono, GNE/Roche, GSK, Lilly, and Onyx; travel funding from BMS and Janssen. G. Goss: Received honoraria from AZ, BI, BMS, Lilly, Pfizer; served as a consultant or advisor for AZ, BI, BMS; and received travel funding from AZ, BI, BMS, Pfizer. E.B. Garon: Institution received research funding from Merck, Genentech, AstraZeneca, Novartis, Pfizer, Lilly, Bristol-Myers Squibb, and Boehringer Ingelheim. A. Li, N. Aanur: Employed by BMS and owns stock in BMS. R. Jotte: Received honoraria from Bristol-Myers Squibb and Lilly; participated in speakers’ bureau for Bristol-Myers Squibb and Lilly. All other authors have declared no conflicts of interest.

1300PD

Impact of co-occurring genomic alterations on overall survival of BRAF V600E and non-V600E mutated NSCLC patients: Results of the Network Genomic Medicine

A. Kron1, R. Riedel1, S. Michels1, J. Fassunke2, S. Merkelbach-Bruse2, M. Scheffler1, L. Nogova1, R. Fischer1, F. Ueckeroth2, D. Abdulla1, F. Kron1, B. Pauli1, B. Kaminsky3, J. Braess4, U. Graeven5, C. Grohe6, S. Krueger7, R. Bu¨ttner2, J. Wolf1 1 Department I of Internal Medicine, Lung Cancer Group Cologne, University Hospital of Cologne, Cologne, Germany, 2Institute of Pathology, University Hospital of Cologne, Cologne, Germany, 3Department of Pneumology, Bethanien Hospital, Solingen, Germany, 4Department of Oncology and Haematology, St. John of God’s Hospital, Regensburg, Germany, 5Department of Oncology and Haematology, Kliniken Maria Hilf GmbH, Mo¨nchengladbach, Germany, 6Department of Respiratory Diseases, Universit€ atsmedizin Charite, Berlin, Germany, 7Department of Pneumology, Florence Nightingale-Krankenhaus, Du¨sseldorf, Germany Background: Activating BRAF mutations are found in 1-3% of lung adenocarcinomas. Treatment with a combination of a BRAF- and a MEK-inhibitor is now the approved standard therapy for V600E mutated patients. The molecular co-alterations that drive the heterogeneity of BRAF mutated lung cancer patients (pts) are poorly characterized by the lack of multiplex diagnostics results. Frequently used single gene tests are unable to detect co-occurring mutations and their mutual impact on overall survival. The Network Genomic Medicine (NGM) performs high sensitive next generation

doi:10.1093/annonc/mdx380 | 461

abstracts

Annals of Oncology

sequencing (NGS) based diagnostics on a central platform in Cologne for inoperable lung cancer pts in Germany. Methods: The NGS panel used in NGM consists of 17 genes to cover potentially targetable aberrations and is run on Illumina (MySeq) platform. In 2016, we have started the retrospective evaluation of BRAF mutated pts with available clinical data and given consent who had received NGS-based molecular diagnostics. In particular, we have focused on BRAF V600E and non-V600E mutated lung cancer pts with and without cooccurring mutations: their frequency, significance and impact on overall survival. Results: We have analyzed 174 pts (V600E¼55 pts, non-V600E¼119 pts) with eligible clinical data. Co-occurring mutations were detected in 121 BRAF mutated pts (70%). The most frequent co-alteration was found in TP53 for 89 pts (74%). Regardless of treatment regime, BRAF mutated lung cancer pts without co-alterations seem to have a better overall survival (OS) with 15 versus 13 month (p ¼ 0.463), same data for the TP53 co-mutated pts (p ¼ 0.449). Likewise, non-targeted treatment of V600E mutation seems to be a negative prognostic factor with OS ¼ 15 month versus 22 month in non-V600E mutated pts (p ¼ 0.957). Conclusions: We report for the first time to our knowledge the heterogeneity of BRAF mutated lung cancer pts in the largest cohort. This work provides evidence that cooccurring genomic alterations influence the overall survival of these pts and stresses the relevance of the multiplex genotyping. Further data including therapies, co-alterations in V600E and other clinicopathologic parameters will be provided. Legal entity responsible for the study: University Hospital of Cologne for the Network Genomic Medicine Funding: None Disclosure: All authors have declared no conflicts of interest.

1301PD

Three-year follow-up from CheckMate 017/057: Nivolumab versus docetaxel in patients with previously treated advanced non-small cell lung cancer (NSCLC)

E. Felip Font1, S.N. Gettinger2, M.A. Burgio3, S.J. Antonia4, E. Holgado5, D.R. Spigel6, O. Arrieta7, M. Domine Gomez8, O. Aren Frontera9, J. Brahmer10, L.Q. Chow11, L. Crin o12, C. Butts13, B. Coudert14, L. Horn15, M. Steins16, W.J. Geese17, A. Li17, D. Healey17, E.E. Vokes18 1 Thoracic Tumors Group, Hospital Universitari Vall d’Hebron, Barcelona, Spain, 2 Oncology, Yale Cancer Center, New Haven, CT, USA, 3Oncology, IRCCS Istituto Scientiﬁco Romagnolo per lo Studio e la Cura dei Tumori, Meldola, Italy, 4Oncology, H. Lee Mofﬁtt Cancer Center & Research Institute, Tampa, FL, USA, 5Oncology, Hospital de Madrid, Norte Sanchinarro, Madrid, Spain, 6Medical Oncology, Sarah Cannon Research Institute/Tennessee Oncology, PLLC, Nashville, TN, USA, 7Oncology, Instituto Nacional on Jime´nez Dıaz, Madrid, de Cancerologıa, Mexico City, Mexico, 8Oncology, Fundaci Spain, 9Oncology, Centro Internacional de Estudios Clinicos, Santiago, Chile, 10 Oncology, The Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Baltimore, MD, USA, 11Oncology, University of Washington, Seattle, WA, USA, 12 Oncology, Ospedale di Perugia, Perugia, Italy, 13Oncology, Cross Cancer Institute, Edmonton, AB, Canada, 14Oncology, Centre Georges Franc¸ois Leclerc, Dijon, France, 15 Thoracic Oncology, Vanderbilt-Ingram Cancer Center, Nashville, TN, USA, 16Oncology, Thoraxklinik, Heidelberg University Hospital, Heidelberg, Germany, 17Immuno-oncology, Bristol-Myers Squibb, Princeton, NJ, USA, 18Oncology, University Of Chicago Medicine & Biological Sciences, Chicago, IL, USA Background: Long-term data comparing outcomes with immune checkpoint inhibitors versus chemotherapy in NSCLC are limited. The phase 3 trials CheckMate 017 and 057 demonstrated improved overall survival (OS), objective response rates (ORR), and quality of life, as well as a favorable safety profile, with the anti-programmed death (PD)-1 antibody nivolumab versus docetaxel in patients with previously treated

advanced squamous and non-squamous NSCLC, respectively. Updated results based on a minimum follow-up of 3 y are reported. Methods: Patients were randomized 1:1 to receive nivolumab 3 mg/kg Q2W (with option to change to 480 mg Q4W after Sep 2016) or docetaxel 75 mg/m2 Q3W until progression or discontinuation. After completion of the primary analyses, patients who ended treatment with docetaxel could cross over to receive nivolumab. The primary endpoint of each study was OS; other endpoints were ORR, progression-free survival, and efficacy by PD ligand 1 (PD-L1) expression. Results: After a minimum follow-up of 36.6 mo in each study (Feb 2017 database locks), 6% of the 427 total patients randomized to the 2 nivolumab arms remained on treatment; no patients remained on docetaxel. Nivolumab continued to show an OS benefit versus docetaxel, with 3-y OS rates of 16% versus 6% in CheckMate 017 and 18% versus 9% in CheckMate 057. Similar to prior reports, an OS benefit was observed in squamous NSCLC regardless of PD-L1 expression, and in non-squamous NSCLC was enhanced at higher PD-L1 expression levels (Table). Of 427 patients in the combined nivolumab arms, 71 (17%) had OS 3 y. Additional 3-y data across trial endpoints will be presented. Conclusions: With 3 y of follow-up from 2 randomized phase 3 studies, nivolumab continued to demonstrate an OS benefit versus docetaxel in patients with advanced squamous and non-squamous NSCLC. Overall, 3-y survival was achieved in 17% of nivolumab-treated patients. Clinical trial identification: NCT01673867; NCT01642004 Legal entity responsible for the study: Bristol-Myers Squibb Funding: Bristol-Myers Squibb Disclosure: E. Felip Font: Member of advisory boards fro Eli Lilly, Pfizer, Roche, Merck Sharp & Dohme, Boehringre Ingelheim. Speaker’s bureau/lecture fees from Astra Zeneca, Bristol-Myers Squibb and Novartis. S.N. Gettinger: Research funding: ARIAD, AstraZeneca, Boehringer-Ingelheim, Bristol-Myers Squibb, Genentech, Incyte, Pfizer; Consulting fees: ARIAD, Bristol-Myers Squibb, Janssen. S.J. Antonia: Other from Bristol-Meyers Squibb, other from Novartis, other from Merck, other from CBMG, other from Boehringer Ingelheim, other from Genentech, other from AstraZeneca/ MedImmune, other from Memgen, outside the submitted work. D.R. Spigel: Consulting/advisory roles & research funding from Genentech/Roche, Novartis, Celgene, BMS, Lilly, AstraZeneca, Pfizer, Clovis Oncology, BI research funding from Peregrine Pharma, Oncogenex, OncoMed, Amgen, Verastem, Daiichi Sankyo, Merck. O. Arrieta: Personal fees from Bristol-Myers Squibb. O. Aren Frontera: Personal fees for advisory boards with Bristol-Myers Squibb and Roche. J. Brahmer: Advisor/consultant for BMS (uncompensated), Celgene, Eli Lilly, Merck & Co, Syndax and has received grant/trial funding from BMS, Merck & Co, MedImmune/AstraZeneca, Johnson & Johnson, Incyte, Five Prime Therapeutics. L.Q. Chow: Grants and personal fees from BMS, during the conduct of the study; grants and/or personal fees from Novartis, BMS, Merck, Eli Lilly/Imclone, Genentech, Pfizer, AstraZeneca/ MedImmune, Incyte, Seattle Genetics, Sanofi Genzyme, Amgen. L. Crin o: Personal fees from an advisory board for Bristol-Myers Squibb. C. Butts: Honoraria for participation in Advisory Boards from BMS, AstraZeneca, Pfizer and Merck. Participated in clinical trials sponsored by these, as well as Roche, BI, Novartis, and Lilly. No compensation for the participation. L. Horn: Personal fees from AbbVie, Genentech, Merck, Lilly, nonfinancial support from Bristol-Myers Squibb, non-financial support from Xcovery, non-financial support from Bayer. W.J. Geese, A. Li: Employee of Bristol-Myers Squibb. D. Healey: Other from Bristol-Myers Squibb, during the conduct of the study; other from Bristol-Myers Squibb, outside the submitted work. E.E. Vokes: Consultant/ advisory role for AbbVie, Amgen, AstraZeneca, BMS, BI, Celegene, Eli Lilly, Genentech, Leidos, Merck, Regeneron, Serono, Takeda, VentiRx. All other authors have declared no conflicts of interest.

Table: 1301PD OSa overall and by PD-L1 expression level

Overall, n 3-y OS rate, % PD-L1 14.7). 324 patients (40%) were continuing treatment at database lock. 403 patients (50%) had TRAEs. 95 (12%) had grade 3–4 TRAEs, most frequently asthenia (12 [2%]) and fatigue (10 [1%]). Of the 5 cases (1%) of TR grade 3–4 pneumonitis, 3 had a documented resolution, and in these patients, resolution occurred within 5 wk. TRAEs led to treatment discontinuation in 45 patients (6%), most commonly pneumonitis, asthenia, and fatigue (7, 5, and 5 patients each). 2 deaths were deemed TR. Median OS was 9.9 mo (95% CI: 8.7, 13.1). In the subgroup aged 70 yr (n ¼ 279), 155 patients (56%) had TRAEs and 16 (6%) discontinued due to TRAEs. In the subgroup with ECOG PS 2 (n ¼ 98), 45 patients (46%) had TRAEs and 5 (5%) discontinued due to TRAEs. Additional data including outcomes in the age 70 yr and ECOG PS 2 subgroups will be presented. Conclusions: The safety of nivolumab in this study was consistent with prior studies of nivolumab in previously treated SQ NSCLC, with no new safety signals. Tolerability in patients aged 70 yr or with ECOG PS 2 was comparable to the overall population. Clinical trial identification: NCT02409368 Legal entity responsible for the study: Bristol-Myers Squibb Funding: Bristol-Myers Squibb and Ono Disclosure: S. Popat: Honoraria from Merck, Pfizer; served as a consultant/advisor for BI, Eli Lilly, Novartis, Roche, Pfizer (Inst), BI (Inst), BMS (Inst), MSD (Inst); received institutional research funding from BI, Roche, BMS, Clovis; travel: BI, MSD, BMS. A. Ardizzoni: Honoraria from Eli Lilly, BMS, MSD, BI; served as a consultant/advisor for Eli Lilly, BMS, MSD, BI, GSK. T. Ciuleanu: Advisor for Amgen, Astellas, AZ, BI, BMS, Eli Lilly, Ipsen, Janssen, Merck, Novartis, Pfizer, Roche, Sandoz, Sanoi, Serono, Servier, Teva. R. Califano: Consultant/advisor for AstraZeneca, Roche, Clovis Oncology, Novartis, and Pfizer. R. Griffiths: Teaching honoraria from Bristol-Myers Squibb. W. Appel: Consultant or advisor for Amgen, AstraZeneca, and Boehringer Ingelheim; travel funding from Amgen. J. Wolf: Personal fees from University Hospital of Cologne; grants and personal fees from AZ, Novartis, Roche, Pfizer, BI, BMS, Clovis, and nonfinancial support from Novartis, Roche, BI, outside of the submitted work. J. Jiang, L.R. Molife: Employed by BMS and owns stock in BMS. E. Felip Font: Honoraria from BI, MSD, Eli Lilly, Roche, Pfizer, Novartis, BMS, Celgene; Consultant for BI, MSD, Eli Lilly, Roche, Pfizer, Novartis, BMS, Celgene; participated on speakers’ bureau for BMS, Novartis, and Roche. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx380 | 463

abstracts 1304PD

Correlation of radiation pneumonitis history before nivolumab and onset risk of interstitial lung disease or progression free survival of nivolumab in patients with non-small cell lung cancer

A. Tamiya1, M. Tamiya2, K. Nakahama1, Y. Taniguchi1, T. Shiroyama3, S-I. Isa4, T. Inoue2, K. Nishino2, T. Kumagai2, H. Suzuki3, T. Hirashima5, F. Imamura2, S. Atagi4 1 Internal Medicine, Kinki-chuo Chest Medical Center, Sakai, Japan, 2Thoracic Oncology, Osaka International Cancer Institute, Osaka, Japan, 3Thoracic Malignancy, Osaka Prefectural Habikino Hospital, Habikino, Japan, 4Clinical Research Center, Kinki-chuo Chest Medical Center, Sakai, Japan, 5Thoracic Oncology, Osaka Prefectural Habikino Hospital, Habikino, Japan Background: Nivolumab (Nivo) demonstrated the promising efficacy for patients (pts) with non-small cell lung cancer (NSCLC) as second or later line treatment. And, abscopal effect of the immune checkpoint inhibitor after the radiotherapy (RT) attracts attention. However, it has not clarified the correlation of radiation pneumonitis history (RPH) before Nivo and onset risk of interstitial lung disease (ILD) or progression free survival (PFS) of Nivo. So we retrospectively analyzed the correlation of RPH before Nivo and onset risk of ILD or PFS of Nivo treatments in patients with NSCLC. Methods: 201 pts treated with Nivo from December 2015 to July 2016 were retrospectively reviewed. This study was multicenter study conducted by the three respiratory medical centers in Japan. We collected clinical data including age, sex, smoking history, histological types, performance status (PS), RPH, and history of RT to chest field, at the time of starting Nivo. And we evaluate the ILD and efficacy. We investigated relationship between RPH and ILD or PFS. The data cut off was on the end of November 2016. Results: Median age was 68 years old, 135 pts were male, 157 pts had smoking history, 153 pts were PS 0 or 1, 34 pts experienced radiation pneumonitis before Nivo, and 50 pts received the RT to chest field (31 pts were curative RT). For all participants, median PFS was 2.8 months (M), overall ILD rate was 12.4%. In the incidence of ILD, no RPH vs RPH; 9.6% vs 26.5% (relative risk ratio (RRR): 2.76, 95% confidence interval (CI): 1.33-5.73), non-RT to chest field vs RT to chest field; 8.6% vs 22.0% (RRR: 2.37, 95% CI: 1.15-4.88). Furthermore, median PFS was no RPH vs RPH; 2.3 M vs 3.6 M, non-RT to chest field vs RT to chest field; 2.2 M vs 3.3 M, and in univariate analysis, RPH had a trend with PFS (hazard ratio (HR): 0.71, 95% CI: 0.44-1.10), however RT to chest field did not correlate with PFS (HR: 1.02, 95% CI: 0.69-1.47). In multivariate analysis, RPH significantly correlated with PFS (HR: 0.58, 95% CI: 0.35-0.93). Conclusions: The RPH before Nivo not only gives onset risk of ILD but also contributes to the prolongation of PFS of Nivo. Clinical trial identification: Protoco; number: UMIN000025908 release date: 31th January, 2017 Legal entity responsible for the study: Fumio Imamura Funding: Ono pharmaceutical Co., Ltd and Bristol-Myers Squibb Co., Ltd. Disclosure: A. Tamiya: Grants from Ono Pharmaceutical, Bristol-Myers Squibb, and received the personal fees from Eli Lilly, Ono Pharmaceutical, Chugai Pharmaceutical, Boehringer Ingelheim, AstraZeneca, Bristol-Myers Squibb. M. Tamiya: Grants from Ono Pharmaceutical and Bristol-Myers Squibb, and personal fees from Boehringer Ingelheim, Chugai Pharmaceutical, Pfizer, AstraZeneca, Taiho Pharmaceutical, Eli Lilly, Asahi Kasei Pharmaceutical, Daichi Sankyo CO. LTD. Area Medical. K. Nakahama: Grants from Ono Pharmaceutical and Bristol-Myers Squibb. Y. Taniguchi: Grants from Ono Pharmaceutical, grants from Bristol-Myers Squibb, and personal fees from Chugai Pharmaceutical. T. Shiroyama: Grants and personal fees from Ono Pharmaceutical, personal fees from Taiho Pharmaceutical, Boehringer Ingelheim, AstraZeneca. T. Inoue: Grants and personal fees from Ono Pharmaceutical, Bristol-Myers Squibb, personal fees from Chugai Pharmaceutical. K. Nishino: Personal fees from Chyugai, Boehringer Ingelheim, Eli Lilly, AstraZeneca. T. Kumagai: Personal fees from Ono Pharmaceutical, Astra Zeneca, Boehringer Ingelheim. H. Suzuki: Grants from Ono Pharmacetical, personal fees from Taiho Pharmacetical, Boehringer Ingelheim, Pfizer, Eli-Lilly. T. Hirashima: Grants and personal fees from Ono Pharma, Bristol-Myers Squibb, MSD, AstraZeneca, Chugai, Lilly Japan, Boehringer Ingelheim, and grants from Eisai, Daiichi Sankyo, Merck Serono, Taiho, Kyowa Hakko Kirin, Takeda, and personal fees from Bayer. F. Imamura: Grants and personal fees from Ono Pharmaceutical; personal fees from Pfizer, AstraZeneca, Novartis Pharma, Kyowa Hakko Kirin, Boehinger Ingelheim, Taiho Pharmaceutical, Eli Lilly Japan, Chugai Pharceutical, Bristol-Myers Squibb. S. Atagi: Personal fees from Bristol-Myers Squibb, Ono Pharmac, Taiho, Chugai, AstraZeneca, Eli Lilly, Boehringer Ingelheim, and grants from Ono Pharmac, Pfizer, Chugai, AstraZeneca, MSD, Taiho, Yakult Pharmaceutical Industry, Eli Lilly, Boehringer Ingelheim. All other authors have declared no conflicts of interest.

1305PD

Whole body PD-1 and PD-L1 PET in pts with NSCLC

A-L. Niemeijer1, E. Smit1, I. Bahce1, O. Hoekstra2, M. Huisman2, G.A. van Dongen2, B. Windhorst2, N. Hendrikse3, A.J. Poot2, D. Vugts2, D.K. Leung4, W. Hayes4, R.A. Smith4, L.M. Wilson4, E. Thunnissen5, J. de Langen1 1 Pulmonary Diseases, Vrije University Medical Centre (VUMC), Amsterdam, Netherlands, 2 Radiology and Nuclear Medicine, Vrije University Medical Centre (VUMC), Amsterdam, Netherlands, 3Clinical Pharmacology & Pharmacy & Radiology and Nuclear Medicine, Vrije University Medical Centre (VUMC), Amsterdam, Netherlands, 4Imaging, BristolMyers Squibb, Princeton, NJ, USA, 5Pathology, Vrije University Medical Centre (VUMC), Amsterdam, Netherlands Background: Tumor PD-L1 IHC relates moderately with treatment outcome following anti-PD1 therapy in pts with NSCLC and single biopsies do not account for tumor

464 | NSCLC, metastatic

Annals of Oncology heterogeneity. Aim: 1. Assess safety of the PET procedures. 2. Quantify 89Zirconiumlabeled nivolumab (89Zr-nivo) and 18F-labeled BMS-986192 (18F-PD-L1) uptake. 3. Assess tracer uptake heterogeneity. 4. Correlate tracer uptake with PD-1/PD-L1 IHC in tumor, stroma and with treatment outcome. Methods: NSCLC pts eligible for treatment with nivolumab were included. Pts received whole body 18F-PD-L1 and 89Zr-nivo PET scans. Baseline tumor biopsy was required to assess PD-(L)1 IHC status (28.8 assay). SUVpeak was calculated for delineable lesions and correlated to PD-(L)1 IHC and response after 12 wks of nivolumab treatment. Results: 10 pts (3 50%, 5 1%, 5 negative by PD-L1 IHC) were enrolled and 37 lesions analysed. No toxicity related to radiotracer was observed. Tumor uptake of both tracers was visualized in all pts, but not in all lesions. Tracer uptake varied among pts with mean 18F-PD-L1 SUVpeak 4.6, range 0.5 - 14.4 and mean 89Zr-nivo SUVpeak 5.0, range 1.6 – 11 (p ¼ 0.03) and within pts with mean SUVpeak difference 3.6-fold (62.1) and 2.4-fold (60.77) between lesions for 18F-PD-L1 and 89Zr-nivo, respectively. For lesions with 50% PD-L1 IHC, mean 18F-PD-L1 SUVpeak was 8.0 (64.7) as compared to 3.5 (61.6) for lesions with 0) in 36% (not regressing tumors). 40 pts (16%) had HPD. Only 3 pts (1,2%) had confirmed pseudoprogression, 2 of them were initially

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology qualified as HPD. Tumor burden baseline, clinical, molecular, pathological characteristics, PD-L1 status and RR to treatment before IO were not different between HPD and not-HPD pts. Compared to not-HPD population, HPD pts had significantly lower mPFS (1.4 vs 4.9m, p < 0.001) and mOS (3.4 vs 17m, p < 0.001). Conclusions: HPD occurs in 16% of 242 advanced NSCLC pts treated with IO, leading to decreased survival. Further work is needed to better characterize this population. Legal entity responsible for the study: Benjamin Besse Funding: Gustave Roussy Disclosure: C. Audigier Valette: Bristol Myers Squibb V. Westeel: BMS, MSD, Merck, AZ B. Duchemann: Bristol-Myers Squibb, Roche D. Planchard: AstraZeneca Boehringer Ingelheim BMS Lilly MSD Pfizer Roche Novartis Chugai J-C. Soria: AstraZeneca, Astex, Clovis, GSK, Gammamabs, Lilly, MSD, Mission Therapeutics, Merus, Pfizer, Pharmamar Pierre Fabre, Roche-Genentech, Sanofi, Servier, Symphogen, Takeda. All other authors have declared no conflicts of interest.

The resistance gene EGFR T790M was detected in 7 of 9 patients with extracranial lesions, but was only detected in 1 of 14 CSFCTCs samples. Other potential resistant mutations such as MET amplification and ERBB2 mutation were also identified in CSFCTCs. Conclusions: CellSearch could be a more sensitive method for detecting tumor cells in CSF, and potentially provides earlier diagnosis of LM. More importantly, CSFCTCs could be an important and new way of “liquid biopsy” for genetic profiles of metastatic tumor cells in LM patients of NSCLC. Legal entity responsible for the study: Yi-Long Wu Funding: Geneseeq Biotechnology, Inc., Nanjing, China Disclosure: All authors have declared no conflicts of interest.

1308PD

Preliminary efficacy and safety of lorlatinib in patients (Pts) with ROS1-positive non-small cell lung cancer (NSCLC)

Y-S. Li1, B-Y. Jiang1, W-B. Guo1, X-C. Zhang1, J-J. Yang1, Y. Shao2, B. Huang3, Y-H. Liu4, Y-L. Wu1 1 Guangdong General Hospital & Guangdong Academy of Medical Sciences, Guangdong Lung Cancer Institute, Guangzhou, China, 2Geneseeq Biotechnology, Geneseeq, Nanjing, China, 3Guangdong General Hospital & Guangdong Academy of Medical Sciences, Department of Radiology, Guangzhou, China, 4Guangdong General Hospital & Guangdong Academy of Medical Sciences, Department of Pathology, Guangzhou, China

B. Besse1, A.T. Shaw2, B.J. Solomon3, T.M. Bauer4, R. Chiari5, C-C. Lin6, M. Satouchi7, J.S. Clancy8, L.P. James9, A. Abbattista10, E. Felip Font11 1 Department of Medical Oncology, Gustave Roussy Cancer Campus, Villejuif, France, 2 Cancer Center, Massachusetts General Hospital, Boston, MA, USA, 3Medical Oncology, Peter MacCallum Cancer Center, Melbourne, Australia, 4Medical Oncology, Sarah Cannon Research Institute, Nashville, TN, USA, 5Medical Oncology, Santa Maria della Misericordia Hospital, Azienda Ospedaliera di Perugia, Perugia, Italy, 6Department of Oncology, National Taiwan University Hospital, Taipei, Taiwan, 7Department of Thoracic Oncology, Hyogo Cancer Center, Akashi, Japan, 8Clinical, InVentiv Clinical, Princeton, NJ, USA, 9Global Product Development, Pﬁzer Oncology, New York, NY, USA, 10 Global Biometrics and Data Management, Pﬁzer Oncology, Milan, Italy, 11Oncology, Vall d’Hebron Institute of Oncology, Barcelona, Spain

Background: Leptomeningeal metastases (LM) are more common in non-small cell lung cancer (NSCLC) with EGFR mutations. The diagnosis is difficult by traditional imaging only, and leads to poor understanding of resistance mechanisms of LM. Methods: We compared the CellSearch AssayTM, the Thinprep cytologic test (TCT), and brain magnetic resonance imaging (MRI) in 21 NSCLC patients with suspected LM. Next-Generation sequencing that included 416 cancer-associated genes was also performed on cerebrospinal fluid circulating tumor cells (CSFCTCs) of 19 patients. Results: Twenty-one patients were diagnosed with LM, and CSFCTCs were captured by CellSearch in 20 patients (median, 969 CSFCTCs/7.5 mL; range, 27–14,888). CellSearch had a sensitivity of 95.2% for LM diagnosis, which was higher than that of TCT (12/21, 57.1%), MRI (10/21, 47.6%), and MRI plus TCT (19/21, 90.5%), respectively. CTCs were found only in 5 of 14 patients (median, 2 CTCs/7.5 mL; range, 2–4), which was a much lower ratio than CSFCTCs. Genetic profiles of CSFCTCs were highly concordant with molecular mutations identified in the primary tumor (17/19, 89.5%).

Background: Most pts with ROS1-rearranged NSCLC achieve initial benefit with crizotinib but subsequently develop resistance, and further treatment options are limited. In Phase 1 of this study, lorlatinib, a potent and brain-penetrant TKI, was associated with a response rate of 50% in ROS1-positive NSCLC pts, most of whom had central nervous system (CNS) metastases (mets) and a majority of whom had received prior crizotinib. Antitumor activity and safety were explored at the recommended Phase 2 dose. Methods: In this ongoing, Phase 2 study (NCT01970865), pts with ROS1þ NSCLC, 6 asymptomatic untreated or treated CNS mets were enrolled with no restriction on the type or number of prior lines of therapy. Pts received lorlatinib 100 mg QD. The primary objective was overall and intracranial (IC) antitumor activity, measured as overall and IC response by independent central review. Results: 23 pts with ROS1þ NSCLC were treated; 12 had CNS involvement, 16 had received prior crizotinib and 1 had received prior crizotinib and ceritinib. The overall response rate (ORR) by investigator assessment for this cohort was 8/23 (34.8%; 95%

1307PD

Detection of driver and resistance mutations in leptomeningeal metastases of NSCLC by next-generation sequencing of cerebrospinal fluid circulating tumor cells

Table: 1307PD Patient No.

Primary gene profile

Targeted therapy before LM

1

EGFR L858R (lung)

Erlotinib

2b 3 4 5 6 7 9

EGFR del19 (lymph node) EML4-ALK (lung) EGFR 20INS (lung) EGFR del19 (pleura) EGFR L858R (lung) EGFR, ALK (WT) EGFR del19 (pleura)

Icotinib Crizotinib None Geﬁtinib Erlotinib None Geﬁtinib

10 11 12 13 14 15 16 17 18c 19 21

EGFR L858R (lung) EGFR L858R (lung) EGFR L858R (lung) EML4-ALK (lung) EGFR del19 (lung) EGFR del19 (lung) EGFR L858R (lung) EGFR L861Q (lymph node) EGFR L858R (lung) EGFR L858R (lung) EGFR L858R (lung)

Geﬁtini, erlotinib Geﬁtini, erlotinib Geﬁtinib Crizotinib Erlotinib, afatinib Icotinib None Erlotinib Geﬁtinib Geﬁtinib Geﬁtinib

Volume 28 | Supplement 5 | September 2017

Rebiopsy gene profile

EGFR L858R (lung, ARMS); MET(IHC):80% (3þ) UA EML4-ALK (pleural effusion) UA EGFR del19 and T790M (lung) EGFR L858R and T790M (lung) Snapshot, MET, KIT (WT) EGFR del19 and T790M (liver);MET: 100% (3þ) (liver) EGFR del19 and T790M (lung) UA EGFR L858R and T790M (plasma) UA UA UA EGFR L858R (lymph node) UA EGFR L858R and T790M (lung) EGFR L858R and T790M (lung) UA

CSFCTCs gene profile (NGS)

EGFR L858R EGFR del19; EGFR T790M EML4-ALK; MET ampliﬁcation EGFR 20INS EGFR del19; EGFR ampliﬁcation EGFR L858R Common driver gene (WT) EGFR del19; ERBB2 exon8 T328fs; ROS1 exon7 W215X EGFR del19 EGFR L858R Common driver gene (WT) EML4-ALK EGFR del19; MET ampliﬁcation EGFR del19 EGFR and ALK (WT); RET exon4 V253E EGFR L861Q; EGFR del19 EGFR L858R; PIK3CA exon2 N107S; MET exon11 F839L EGFR L858R EGFR L858R

doi:10.1093/annonc/mdx380 | 465

abstracts CI: 16.4, 57.3). The best overall response was SD in 11 pts (47.8%) and progressive disease (PD) in 2 pts (8.7%). The disease control rate at 12 weeks was 17/23 (73.9%; 95% CI: 51.6, 89.8). Of 8 pts with a confirmed response, 3 have progressed (after a range of 3.5–7.0 mos) and the remainder are in follow-up (range: 5.6–8.3 mos). The IC-ORR was 3/12 (25%) (95% CI: 5.5, 57.2). The most commonly reported treatment-related adverse events (TRAEs) were hypercholesterolemia (95.7%) and hypertriglyceridemia (69.6%) both managed with statins or other lipid lowering agents. Additional TRAEs included peripheral edema (34.8%) and cognitive effects (30.4%). Treatment-related dose interruptions and dose reductions occurred in 26.1% and 17.4% of pts, respectively; there were no treatment-related discontinuations or deaths. Most pts remain on treatment. Conclusions: Lorlatinib has shown clinical activity in ROS1þ NSCLC pts, a majority of whom had CNS involvement and most of whom had received prior crizotinib. Overall, lorlatinib was well-tolerated with lipid elevations being the most common TRAEs. Clinical trial identification: NCT01970865 Legal entity responsible for the study: Pfizer Funding: Pfizer Disclosure: B. Besse: Corporate sponsored research: Pfizer. A.T. Shaw: Advisory board or board of directors: Blueprint medicines, KSQ Therapeutics (scientific advisory board). Consulting/Honoria Pfizer, Novartis, Ariad, Genentech/Roche, Ignyta, Daiichi-sankyo, Taiho, LOXO, Blueprint medicines, EMD Serono, Foundation Medicine. B.J. Solomon: Membership of an advisory board or board of directors: Advisory Boards: Pfizer, Novartis, Roche-Genentech, AstraZeneca, Merck, Bristol Myers Squibb. M. Satouchi: Corporate sponsored research: Pfizer Honoraria: Pfizer. J.S. Clancy: Employed by inVentiv clinical but under contract position with Pfizer. L.P. James, A. Abbattista: Employee and stock owner of Pfizer. E. Felip Font: Advisory boards: Eli Lilly, Pfizer, Roche, MSD, Boehringer Ingelheim Speaker’s bureau/lectures fees: AstraZeneca, BMS, Novartis. All other authors have declared no conflicts of interest.

1309P

Efficacy and safety of necitumumab and pembrolizumab combination therapy in patients with stage IV non-small cell lung cancer (NSCLC)

B. Besse1, P. Garrido Lopez2, J. Puente3, A. Cortot4, M.E. Olmedo Garcia5, M. Perol6, M. Gil7, G. Chao8, J. Shahidi9, J. Bennouna10 1 Cancer Medicine, Institut de Cance´rologie Gustave Roussy, Villejuif, France, 2Medical Oncology, Hospital Universitario Ram on y Cajal, Madrid, Spain, 3Medical Oncology, Hospital Clinico Universitario San Carlos, Madrid, Spain, 4Pneumology-Oncology, Hospital Albert Calmette, Lille, France, 5Medical Oncology, Hospital Universitario Ramon y Cajal, Madrid, Spain, 6Oncologie, Centre Le´on Be´rard, Lyon, France, 7Medical Oncology, Eli Lilly Poland, Warsaw, Poland, 8Biometrics, Eli Lilly and Company, Bridgewater, NJ, USA, 9Medical Oncology, Eli Lilly and Company, Bridgewater, NJ, USA, 10 Medical Oncology, CHU de Nantes, Nantes, France Background: Studies of EGFR-directed monoclonal antibody (mAb) necitumumab (neci) and anti-PD1 pembrolizumab (pembro) demonstrate activity of each agent in NSCLC. Methods: This phase 1b, multicenter, single arm study of neci and pembro examined the safety, efficacy, and tolerability in pretreated patients with Stage IV NSCLC (NCT02451930). PDL1 was centrally assessed retrospectively using IHC 22C3 pharmDx assay (negative, weak positive, strong positive if < 1%, 1-49%, 50% of tumor cells were stained, respectively). Escalating doses of neci 600 – 800 mg IV (days 1 and 8 every 3 weeks [Q3W]) were administered with pembro (200 mg IV) on Day 1 Q3W (Part A). Established dose from Part A was used in the expansion cohort (Part B). Study objectives were to determine the dose-limiting toxicity (DLT) and evaluate tolerability and overall response rate (ORR) by RECIST 1.1. Secondary objectives included progression-free survival (PFS) and overall survival (OS). Results: for 64 patients are reported. Part A completed without DLTs (9 patients; 2 squamous, 7 nonsquamous). Overall, 3 patients received neci 600 mg and 61 patients received neci 800 mg; all patients received pembro 200 mg. ORR (95% CI) was 23.4% (13.8, 35.7). Median PFS (95% CI) was 4.1 m (2.4, 6.9). Six-month OS rate (95% CI) was 74.7% (61.5, 83.9). Treatment-emergent adverse events occurring in 20% of all patients (n (%)) included dermatitis acneiform: 43 (67.2); asthenia: 24 (37.5); dry skin: 23 (35.9); hypomagnesemia: 21 (32.8); fatigue: 20 (31.3); decreased appetite and diarrhea (each): 17 (26.6); headache, pruritus, and stomatitis (each): 14 (21.9). Two neci 800 mg þ pembro patients experienced grade 5 respiratory tract infection. Conclusions: The results suggest modest activity of the combination in a NSCLC patient population with a relatively high proportion of PDL1 negative tumors (Table).

466 | NSCLC, metastatic

Annals of Oncology

Table: 1309P Necitumumab 600 mg/800 mg 1 Pembrolizumab 200 mg Overall

Squamous

Nonsquamous

(N 5 64)

(n 5 30)

(n 5 34)

Age, median (range), y

65 (43, 81)

67.5 (48, 81)

61 (43, 75)

Male, n (%)

46 (71.9)

23 (76.7)

23 (67.6)

1 line

36 (56.3)

22 (73.3)

14 (41.2)

2 lines

15 (23.4)

5 (16.7)

10 (29.4)

3 lines

13 (20.3)

3 (10.0)

10 (29.4)

Prior systemic therapy, n (%)

Baseline ECOG PS, n (%)

64 (100)

30 (100)

34 (100)

0

17 (26.6)

3 (10.0)

14 (41.2)

1

46 (71.9)

26 (86.7)

20 (58.8)

2

1 (1.6)

1 (3.3)

0

Tobacco use, n (%)

64 (100)

30 (100)

34 (100)

Former

41 (64.1)

21 (70.0)

20 (58.8)

Current

14 (21.9)

7 (23.3)

7 (20.6)

Never

9 (14.1)

2 (6.7)

7 (20.6)

ORR n (%) (95% CI)

15 (23.4) (13.8, 35.7)

6 (20.0) (7.7, 38.6)

9 (26.5) (12.9, 44.4)

mPFS (months) (95% CI)

4.1 (2.4, 6.9)

2.8 (1.4, 5.5)

6.9 (2.7, 12.3)

6-month OS rate (%) (95% CI)

74.7 (61.5, 83.9)

63.6 (42.8, 78.6)

84.2 (66.0, 93.1)

PDL1 Status

64 (100)

30 (100)

34 (100)

PDL1 negative

32 (50.0)

13 (43.3)

19 (55.9)

ORR n (%) (95% CI)

4 (12.5) (3.5, 29.0)

1 (7.7) (0.2, 36.0)

3 (15.8) (3.4, 39.6)

mPFS (m) (95% CI)

2.69 (1.4, 4.1)

6-month OS rate (%) (95% CI)

68.2 (47.7, 82.0)

Ef˚cacy

PDL1 Weak positive

12 (18.8)

7 (23.3)

5 (14.7)

ORR n (%) (95% CI)

3 (25.0) (5.5, 57.2)

1 (14.3) (0.4, 57.9)

2 (40.0) (5.3, 85.3)

mPFS (m) (95% CI)

5.4 (0.8, –)

6-month OS rate (%) (95% CI)

83.3 (48.2, 95.6)

PDL1 Strong positive

10 (15.6)

5 (16.7)

5 (14.7)

ORR n (%) (95% CI)

4 (40.0) (12.2, 73.8)

2 (40.0) (5.3, 85.3)

2 (40.0) (5.3, 85.3)

mPFS (m) (95% CI)

7.6 (1.0, 12.3)

6-month OS rate (%) (95% CI)

80.0 (40.9, 94.6)

Unknown

10 (15.6)

5 (16.7)

5 (14.7)

ORR n (%) (95% CI)

4 (40.0) (12.2, 73.8)

2 (40.0) (5.3, 85.3)

2 (40.0) (5.3, 85.3)

mPFS (m) (95% CI)

– (0.82, –)

6-month OS rate (%) (95% CI)

78.8 (38.1, 94.3)

ECOG PS, Eastern Cooperative Oncology Group performance status; PDL1, programmed death ligand 1; ORR, overall response rate; mPFS, median progression-free survival; OS, overall survival; CI, conﬁdence interval.

Clinical trial identification: NCT02451930 Legal entity responsible for the study: Eli Lilly and Company Funding: Eli Lilly and Company Disclosure: B. Besse: Research grants from Eli Lilly and Company and Merck, Sharp & Dohme. P. Garrido Lopez: Personal fees from BMS, Novartis, Pfizer, Roche, MSD, Guardant, and Abbvie; and grants and personal fees from BI. A. Cortot: Personal fees from Lilly and MSD; personal fees and non-financial support from Roche, Novartis, Astra-Zeneca, and Pfizer; and grants, personal fees, and non-financial support from Boehringer-Ingelheim. M. Perol: Personal fees from Eli Lilly and MSD. M. Gil: Employee of Eli Lilly and Company and owns Eli Lilly stock. Dr. Gil has a patent Combination Therapy for cancer pending to Eli Lilly. G. Chao: Employee of Eli Lilly and Company. J. Shahidi: Employee and stock owner of Eli Lilly and Company. J. Bennouna: Personal fees from Roche, Boehringer-Ingelheim, BMS, and Astra-Zeneca. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 1310P

Survival and safety of atezolizumab by best overall response (BOR) in the phase III NSCLC OAK study

F. de Marinis1, F. Barlesi2, A. Rittmeyer3, J. von Pawel4, J-Y. Han5, M. Kozloff6, A. Spira7, L. Fehrenbacher8, D.R. Gandara9, W. Yu10, P. He11, C. Yun12, M. Ballinger13, M. Gandhi12, S. Gadgeel14 1 Thoracic Oncology, European Institute of Oncology, Milan, Italy, 2Multidisciplinary Oncology & Therapeutic Innovations, Aix Marseille University, Marseille, France, 3 Thoracic Oncology, Lungenfachklinik Immenhausen, Immenhausen, Germany, 4 Thoracic Oncology, Asklepios-Fachkliniken Mu¨nchen-Gauting, Gauting, Germany, 5 Center for Lung Cancer, National Cancer Center, Goyang, Republic of Korea, 6 Medicine, Ingalls Memorial Hospital, Harvey, USA, 7Oncology, Virginia Cancer Specialists Research Institute, Fairfax, USA, 8Oncology, Kaiser Permanente Medical Center, Vallejo, CA, USA, 9Thoracic Oncology, UC Davis Comprehensive Cancer Center, Sacramento, CA, USA, 10USMA Biometrics, Genentech, South San Francisco, CA, USA, 11 PD Biostatistics, Genentech, Inc., South San Francisco, CA, USA, 12USMA, Genentech, South San Francisco, CA, USA, 13Product Development Oncology, Genentech, South San Francisco, CA, USA, 14Medical Oncology, University of Michigan, Ann Arbor, MI, USA Background: Atezolizumab (atezo; anti–PD-L1) inhibits binding of PD-L1 to PD-1 and B7.1, thereby restoring tumor-specific T-cell immunity. OAK, the first randomized Ph III study of atezo vs docetaxel (doc) in 2L/3L NSCLC demonstrated a superior OS benefit of atezo (HR 0.73; 95% CI: 0.62, 0.87; P ¼ 0 .0003) in patients (pts) regardless of PD-L1 expression levels on tumor cells (TC) or tumor-infiltrating immune cells (IC). Here we present efficacy and safety analyses in the OAK primary population (n ¼ 850) by BOR subgroups. Methods: Previously treated pts were randomized 1:1 to atezo (1200 mg) or doc (75 mg/m2) IV q3w. Co-primary endpoints were OS in ITT and PD-L1 expression subgroup ( 1% PD-L1 on TC or IC). Secondary endpoints included ORR and safety. BOR subgroups were defined based on RECIST v1.1 response determined by investigators. Time to response (TTR) was based on tumor assessment every 6 weeks. Data cutoff, July 7, 2016. Results: Baseline demographics were generally similar across BOR subgroups except for PD-L1 expression status. TC1/2/3 or IC1/2/3 prevalence was 74% in CR/PR, 53% in SD and 55% in PD subgroups, respectively. The survival benefit of atezo vs doc was observed across BOR subgroups with greatest benefit occurring in the CR/PR subgroup (HR, 0.32; 95% CI 0.16, 0.63; see Table). Among pts in the CR/PR subgroup, the median TTR was comparable between atezo and doc arms (2.8 mo each). The median duration of response was longer in the atezo arm pts (16.3 mo, 95% CI: 10.0, NE) vs doc arm pts (6.2 mo, 95% CI: 4.9, 7.6). OS benefit with atezo was also observed in patients with SD and PD as BOR (see Table). No new safety findings were observed among BOR subgroups. Conclusions: Atezo responses were durable. Atezo responders had more than twothirds reduction in the risk of death compared with doc responders. In addition, the improved OS with atezo vs doc seen in pts with SD and PD suggests that clinical benefit also extended to patients who did not have a radiographic response.

Table: 1310P Efficacy of atezolizumab vs docetaxel by BOR subgroups Patient Population

Atezolizumab n

ITT (N ¼ 850) 425 BOR subgroups CR/PR 58 SD 150 PD 187

mOS (95% CI), mo

Docetaxel n

HR (95% CI)a

mOS, (95% CI) mo

13.8 (11.8, 15.7) 425 9.6 (8.6, 11.2)

0.73 (0.62, 0.87)

NE (NE, NE) 57 20.0 (15.9, NE) 0.32 (0.16, 0.63) 17.6 (15.7, 20.2) 177 13.0 (11.5, 14.7) 0.70 (0.53, 0.92) 7.3 (6.7, 9.4) 117 6.4 (5.6, 7.3) 0.72 (0.56, 0.93)

a Stratiﬁed HR for ITT and unstratiﬁed HR for subgroups. 95% CI for HR were estimated using Cox regression. NE, not estimable.

Clinical trial identification: NCT02008227 Legal entity responsible for the study: F. Hoffmann - La Roche Ltd. Funding: F. Hoffmann - La Roche Ltd. Disclosure: F. de Marinis: Consultation fees received from Roche/BMS/Boehringer/ Novartis/Pfizer/MSD/Astrazeneca. F. Barlesi: Honarium from Roche. A. Rittmeyer: Grants as an advisor or speaker by: Astra Zeneca, BMS, Boehringer Ingelheim, Eli Lilly, Pfizer and Roche Genentech. J. von Pawel: Adboard: AbbVie, Pfizer, Bristol Myers Squibb, Novartis money paid to the institution. A. Spira: Research sponsored by Roche/Genentech (payable to institution), Speakers bureau. D.R. Gandara: Consultant: Genentech Clinical trial grant: Genentech. W. Yu: Genentech Employee. P. He: Employee of Roche/Genentech, and have stocks for Roche, Amgen. Husband has stocks for Allergan and Gilead. C. Yun: Employee of Genentech, Roche stock, Research

Volume 28 | Supplement 5 | September 2017

funding from Genentech M. Ballinger: Employee of Genentech, Roche stock. M. Gandhi: Employee of Genentech. S. Gadgeel: Speaker’s bureau: Astra-Zeneca, Genentech/Roche Advisory Boards- Astra-Zeneca, Ariad, Pfizer, Bristol Myers- Squibb and Genentech/Roche. All other authors have declared no conflicts of interest.

1311P

LKB1 loss is a novel genomic predictor of de novo resistance to PD-1/ PD-L1 axis blockade in KRAS-mutant lung adenocarcinoma

F. Skoulidis1, M.D. Hellman2, M. Awad3, J.F. Gainor4, H. Rizvi2, B. Carter5, W.L. Denning1, P. Villalobos6, E.R. Parra6, Y.Y. Elamin1, J. Zhang1, G.C. Leonardi7, D.F. Halpenny8, V. Papadimitrakopoulou1, I.I. Wistuba6, J.D. Wolchok9, A.T. Shaw10, P.A. J€anne11, C.M. Rudin2, J.V. Heymach1 1 Thoracic and Head and Neck Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA, 2Thoracic Oncology, Memorial Sloan-Kettering Cancer Center, New York, NY, USA, 3Thoracic Oncology, Dana-Farber Cancer Institute, Boston, MA, USA, 4Termeer Center for Targeted Therapies, Massachusetts General Hospital, Boston, MA, USA, 5Diagnostic Imaging, The University of Texas MD Anderson Cancer Center, Houston, USA, 6Translational Molecular Pathology, MD Anderson Cancer Center, Houston, TX, USA, 7Thoracic Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA, 8Radiology, Memorial Sloan-Kettering Cancer Center, New York, NY, USA, 9Oncology, Memorial Sloan Kettering Cancer Center and Weill Cornell Medical College, New York, NY, USA, 10Cancer Center, Massachusetts General Hospital, Boston, MA, USA, 11Lowe Center for Thoracic Oncology, Dana Farber Cancer Institute, Boston, MA, USA Background: We previously reported that KRAS-mutant lung adenocarcinomas (LUAC) with co-occurring genetic events in STK11/LKB1 (KL) or TP53 (KP) define subgroups with marked differences in immune contexture, including paucity of infiltrating CD8þ lymphocytes in KL LUACs. Here, we present updated data on the clinical efficacy of PD-1/PD-L1 inhibitors in co-mutation defined KRAS-mutant LUAC subsets using data assembled by members of the SU2C/ACS Lung Cancer Dream Team. Methods: Patients (pts) with metastatic KRAS-mutant LUAC who received at least one cycle of PD-1/PD-L1 inhibitor therapy, were alive for 14 days thereafter, and had available molecular profiling were identified retrospectively. Efficacy assessment was based on RECIST v1.1. PD-L1 expression was tested using 22C3 pharmDx or E1L3N IHC assays. Isogenic derivatives of the LKR10 KrasLA1/þ murine LUAC cell line with CRISPR/Cas9-mediated Lkb1 knockout were used in preclinical experiments. Results: 192 immunotherapy-treated (82% nivolumab, 12% pembrolizumab, 5% antiPD1/PD-L1 plus anti-CTLA-4) pts with KRAS-mutant LUAC were included in the analysis. The ORR differed significantly between the KL (8.9%), KP (37.9%) and K-only sub-groups (25.8%) (P ¼ 0.00069, Fisher’s exact test) and was concordant for each genotype across patient cohorts [ORR for KL: 8.3% in the MDA cohort, 8.7% in the MSKCC cohort and 9.5% in the DFCI/MGH cohort). KL LUAC exhibited significantly shorter PFS (mPFS 1.8m vs 3m, HR ¼ 0.47, 95% CI 0.32-0.7, P ¼ 0.0002, log-rank test) and OS (mOS 6.8m vs 16.1m, HR 0.48, 95% CI 0.3 to 0.76, P ¼ 0.0018, log rank test) compared to KRAS-mutant LUAC with wild-type LKB1. 11/14 KL tumors with available IHC data were negative for PD-L1 expression. Among 7 PD-L1-negative KP tumors, 3 PRs and 2SDs were recorded. In syngeneic murine models loss of Lkb1 promoted resistance to PD-1 inhibitor monotherapy, suggesting a causative role. Conclusions: Inactivation of LKB1 represents a novel genomic predictor of de novo resistance to PD-1/PD-L1 blockade in KRAS-mutant LUAC. In addition to tumor PD-L1 status and tumor mutational burden precision immunotherapy approaches should take into consideration the LKB1 status of individual tumors. Legal entity responsible for the study: SU2C/ACS Lung Cancer Dream Team Funding: SU2C/ACS Lung Cancer Dream Team, MD Anderson Lung Cancer Moonshot, Cancer Prevention Research Institute of Texas, Andrew Sabin Family Foundation. Disclosure: M.D. Hellman: Consultant/Advisory Board: Genentech, BMS, Merck, AstraZeneca, Janssen, Novartis Research support- Genentech, BMS. J.V. Heymach: Scientific Advisory Board: Genentech, BMS, AstraZeneca, Eli Lilly, EMD Serono, Boehringer Ingelheim. All other authors have declared no conflicts of interest.

1312P

Prediction of survival with durvalumab in locally advanced or metastatic NSCLC using early tumor assessments

X. Zhang1, K. Park2, N.A. Rizvi3, P.A. Dennis4, R. Narwal5, Y. Huang6, R. Arani7, P. Mukhopadhyay7 1 Immuno-Oncology GMD, AstraZeneca, Gaithersburg, MD, USA, 2Division of Hematology/Oncology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea, 3Department of Medicine, Columbia University Medical Center, New York, NY, USA, 4Global Medicines Development, AstraZeneca USA, Gaithersburg, MD, USA, 5Clinical Pharmacology, Pharmacometrics, & DMPK (CPD), MedImmune, Gaithersburg, MD, USA, 6GMD Oncology B&I, AstraZeneca, Gaithersburg, MD, USA, 7Biometrics & Information Sciences, AstraZeneca, Gaithersburg, MD, USA Background: The analysis objective was to assess if limited tumor assessments can predict long-term overall survival (OS) in patients (pts) with locally advanced or metastatic (Stage IIIB-IV) non-small cell lung cancer (NSCLC) treated with durvalumab. Methods: We used data from a Phase II, non-comparative, open-label multicenter study of durvalumab in NSCLC pts with 2 prior systemic treatment regimens,

doi:10.1093/annonc/mdx380 | 467

abstracts

Annals of Oncology

Table: 1312P ATLANTIC (model building) Bad Group (n 5 157) Median OS (95% CI), days 6-month OS rate (95% CI) 1-year OS rate (95% CI) HR [Good vs. Bad] (95% CI)

340 (292, 403) 0.742 (0.665, 0.804) 0.478 (0.397, 0.554) 0.2059 (0.0569, 0.4437)

Good Group (n 5 34) NE (557, NE) 0.941 (0.785, 0.985) 0.882 (0.716, 0.954)

Study 1108 (validation) Bad Group (n 5 117)

Good Group (n 5 58)

265 (194, 315) 0.605 (0.517, 0.682) 0.371 (0.282, 0.460) 0.2637 (0.1661, 0.4187)

739 (616, NE) 0.937 (0.855, 0.973) 0.819 (0.708, 0.891)

NE, not estimable

including 1 platinum-based (ATLANTIC). Per exploratory analysis, the first 2 postbaseline assessments were used to develop the model. Using an elastic net statistical method, combined with cross validation, we identified important baseline variables, built a scoring system (defined as 0.28*sex þ 0.188*histology group þ 0.034*smoker group – 0.176*line of therapy – 0.041*tumor assessment) in which assessments are represented as a single variable (interpreted as a weighted average), and identified the optimal score thresholds to segment pts into 2 groups (‘good’ vs. ‘bad’) with significant differences in long-term OS. Results: As of June 3, 2016, 444 pts had received treatment; 191 from cohort 2 (EGFR/ALK wild-type pts) with sufficient assessments (baseline and 1 follow-up) were used to develop the model. Median age was 64.0 years, 61.8% had WHO PS 1, 18.8% had squamous histology, mean number of prior anticancer regimens was 4.0, and 83.7% were current/exsmokers; PD-L1 expression was high (25% of tumor cells stained) in 57.1%, low/negative in 35.6%, and unknown in 7.3%. OS results are summarized in the table. The model was validated using data from a Phase I/II open-label trial of durvalumab (1108). Conclusions: We developed an algorithm based on baseline characteristics and tumor assessments to segment NSCLC pts treated with durvalumab into 2 groups with distinct OS. The scoring system was independently validated. However, the predictive versus prognostic value of this algorithm needs further evaluation using data from randomized trials. Clinical trial identification: NCT02087423 (release date: March 4, 2014) Legal entity responsible for the study: AstraZeneca PLC Funding: AstraZeneca Disclosure: X. Zhang: Full time employee of AstraZeneca. K. Park: Consulting: Astellas, AZ, Boehringer Ingelheim, Clovis, Lilly, Hanmi, Kyowa Hakko Kirin, Novartis, Ono Pharma. Speaker Bureau: Boehringer Ingelheim, Research funding: AZ. N.A. Rizvi: Advisory Board: Merck, AZ, Roche, BMS, Novartis, Pfizer, Lilly, Novartis, Abbvie Co-founder and shareholder: Gritstone Oncology Scientific Advisory Board: Nilogen Oncosystems. P.A. Dennis, Y. Huang, P. Mukhopadhyay: Employee and shareholder AstraZeneca. R. Narwal: Employee and shareholder MedImmune. R. Arani: Employee B&I, AstraZeneca, Shareholder AstraZeneca.

1313P

Immune-related adverse events (irAEs) in advanced NSCLC patients treated with atezolizumab: Safety population analyses from the Ph III study OAK

D. Cortinovis1, J. von Pawel2, K. Syrigos3, J. Mazieres4, R. Dziadziuszko5, L. Fehrenbacher6, P. Conkling7, J. Goldschmidt8, C.A. Thomas9, R. Bordoni10, M. Kosty11, F.S. Braiteh12, P. He13, M. Ballinger14, M. Gandhi15, H. Patel16, D.R. Gandara17 1 Medical Oncology Unit, AOU San Gerardo, Monza, Italy, 2Thoracic Oncology, Asklepios-Fachkliniken Mu¨nchen-Gauting, Gauting, Germany, 33rd Department of Medicine, Athens School of Medicine, Athens, Greece, 4Thoracic Oncology Department, Toulouse University Hospital, Toulouse, France, 5Department of Oncology and Radiotherapy, Medical University of Gdansk, Gdansk, Poland, 6Oncology, Kaiser Permanente Medical Center, Vallejo, CA, USA, 7Medical Oncoloy, Virginia Oncology Associates, Norfolk, VA, USA, 8Clinical Research, Blue Ridge Cancer Care, Blacksburg, USA, 9Medical Oncoloy, New England Cancer Specialists, Scarborough, ME, USA, 10 Medicine, Georgia Cancer Specialists; Northside Hospital Cancer Institute, Atlanta, GA, USA, 11Division of Hematology/Oncology, Scripps Clinic, La Jolla, USA, 12Medical Oncology, Comprehensive Cancer Centers of Nevada, Las Vegas, USA, 13Biostatistics, Genentech, Inc., South San Francisco, CA, USA, 14Product Development Oncology, Genentech, South San Francisco, CA, USA, 15USMA, Genentech, Inc., South San Francisco, CA, USA, 16Product Development Safety Science, Genentech, South San Francisco, CA, USA, 17Thoracic Oncology, UC Davis Comprehensive Cancer Center, Sacramento, CA, USA

randomized Ph III study of atezo in NSCLC patients (pts) who had failed prior platinum therapy. In the primary efficacy population (n ¼ 850), atezo benefit was seen regardless of PD-L1 expression levels on tumor cells (TC) or tumor-infiltrating immune cells (IC). Here, we present the analyses of irAEs in the safety population (N ¼ 1225) of OAK. Methods: Pts were randomized 1:1 to atezo (1200 mg) or doc (75 mg/m2) IV q3w. Coprimary endpoints were OS in ITT and in PD-L1 expression subgroups. Secondary endpoints included ORR and safety. irAEs were defined using MedDRA Preferred Terms that included both diagnosed immune conditions and signs and symptoms potentially representative of immune related events, regardless of investigator-assessed causality. Safety analyses conducted were incidence, nature and severity of irAEs, corticosteroid use and irAEs leading to atezo interruption/discontinuation. Data cutoff: July 7, 2016. Results: In the atezo arm, 6.2% of pts had grade 3-4 irAEs and 25.0% of pts had grade 1-2 irAEs. No grade 5 irAEs were reported. Low rates of any-grade hypothyroidism (3.9%), pneumonitis (1.5%), hepatitis, (1.1%), and colitis (0.3%) were observed. Grade 3-4 irAEs included pneumonitis (0.7%) and hepatitis (0.7%); no pts developed Grade 3-4 colitis. 36 (5.9%) atezo arm pts experienced irAEs requiring corticosteroid treatment. Majority of irAEs in the atezo arm were manageable; 13 pts (2.1%) discontinued atezo. Meningoencephalitis (0.7%) and AST/ALT elevation (0.3%/0.2%) were the most frequently reported irAEs leading to atezo discontinuation. 26 pts (4.3%) had dose interruptions due to irAEs. AST/ALT elevation (0.8%/0.8%) and diarrhea (0.8%) were the most frequently reported irAEs leading to dose interruption. Conclusions: The irAEs occurring in atezo-treated pts were mostly low grade and manageable, with few pts requiring dose interruption/discontinuation of atezo and corticosteroid treatment. Efficacy data based on irAE subgroups of OAK are presented separately. Clinical trial identification: NCT02008227 Legal entity responsible for the study: F. Hoffmann - La Roche Ltd. Funding: F. Hoffmann - La Roche Ltd. Disclosure: D. Cortinovis: Membership for AB for Roche, Novartis, MSD, BI. J. von Pawel: Adboard: AbbVie, Pfizer, Bristol Myers Squibb, Novartis money paid to the institution. R. Dziadziuszko: Honoraria or consulting fees from Roche, Pfizer, Boehringer-Ingelheim, Clovis Oncology, Novartis, Astra-Zeneca, Tesaro. P. Conkling: research funding is US Oncology Research. J. Goldschmidt: Honoraria: Amgen Consulting/Advisory Role: Amgen Speakers Bureau: Bristol Myers-Squibb, Celgene. M. Kosty: Support limited to Institutional support for the reported trial and other clinical trials. No direct compensation to investigator. No other conflicts to report. F.S. Braiteh: COI: speaking and consulting fees received from Genentech. P. He: Employee of Roche/Genentech, and have stocks for Roche, Amgen. Husband has stocks for Allergan and Gilead. M. Ballinger: Employee of Genentech, Roche stock. M. Gandhi: Employee of Genentech. H. Patel: Genentech Employee. D.R. Gandara: Consultant: Genentech Clinical trial grant: Genentech. All other authors have declared no conflicts of interest.

Background: A superior survival benefit with atezolizumab (atezo; anti–PD-L1) vs docetaxel (doc; HR 0.73; 95% CI: 0.62, 0.87) has been demonstrated in OAK, the first

468 | NSCLC, metastatic

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 1314P

Association between immune-related adverse events (irAEs) and atezolizumab efficacy in advanced NSCLC: analyses from the phase III study OAK

J. von Pawel1, K. Syrigos2, J. Mazieres3, D. Cortinovis4, R. Dziadziuszko5, D.R. Gandara6, P. Conkling7, J. Goldschmidt8, C.A. Thomas9, R. Bordoni10, M. Kosty11, F.S. Braiteh12, S. Hu13, M. Ballinger14, H. Patel15, M. Gandhi16, L. Fehrenbacher17 1 Medical Oncoloy, Asklepios-Fachklinikum, Gauting, Germany, 23rd Department of Medicine, Athens School of Medicine, Athens, Greece, 3Thoracic Oncology, Toulouse University Hospital, Toulouse, France, 4Medical Oncology Unit, AOU San Gerardo, Monza, Italy, 5Department of Oncology and Radiotherapy, Medical University of Gdansk, Gdansk, Poland, 6Thoracic Oncology, UC Davis Comprehensive Cancer Center, Sacramento, CA, USA, 7Medical Oncoloy, Virginia Oncology Associates, Norfolk, VA, USA, 8Clinical Research, Blue Ridge Cancer Care, Blacksburg, USA, 9Medical Oncoloy, New England Cancer Specialists, Scarborough, ME, USA, 10Medicine, Georgia Cancer Specialists; Northside Hospital Cancer Institute, Atlanta, GA, USA, 11Division of Hematology/Oncology, Scripps Clinic, La Jolla, USA, 12Medical Oncology, Comprehensive Cancer Centers of Nevada, Las Vegas, USA, 13Biostatistics, Genentech, Inc., South San Francisco, CA, USA, 14Product Development Oncology, Genentech, South San Francisco, CA, USA, 15Product Development Safety Science, Genentech, South San Francisco, CA, USA, 16USMA, Genentech, South San Francisco, CA, USA, 17 Oncology, Kaiser Permanente Medical Center, Vallejo, CA, USA Background: OAK, the first randomized Ph III study of atezolizumab (atezo; anti–PDL1), demonstrated a superior survival benefit with atezo vs docetaxel (doc) in NSCLC patients (pts) who had failed prior platinum therapy (HR 0.73; 95% CI: 0.62, 0.87). This analysis evaluates the benefit of atezo in pts with and without irAEs from the primary efficacy population (n ¼ 850) of OAK. Safety data are reported separately. Methods: Pts were randomized 1:1 to atezo (1200 mg) or doc (75 mg/m2) IV q3w. irAEs were defined using MedDRA Preferred Terms that included both diagnosed immune conditions and signs and symptoms potentially representative of immune related events, regardless of investigator-assessed causality. For this analysis, efficacy was evaluated in pts with and without irAEs in the atezo and doc arms; efficacy endpoints were OS, PFS and ORR. To overcome the inherent survivor bias between irAE subgroups (pts surviving longer may be more likely to have irAEs), OS was evaluated using a timedependent (TD) Cox model. Additional exploratory analyses included atezo efficacy in pts who did vs did not receive systemic steroids for irAEs. Data cutoff: July 7, 2016. Results: The incidence of irAEs in the atezo arm was 31% (25.0% grade 1-2, 6.2% grade 3-4, no grade 5). Baseline characteristics including PD-L1 expression on tumor cells or tumor-infiltrating immune cells were generally similar between irAE subgroups. OS per TD Cox model was in favor of atezo arm pts with irAEs vs those without irAEs (HR 0.79; 95% CI: 0.60, 1.05). Median time to onset of first irAE was 1.6 mo; post irAE mOS was 17.3 mo (95% CI: 11.7, 21.2). 24 atezo arm pts (6%) required corticosteroid treatment. Median OS in pts who did vs did not receive corticosteroids was 16.0 mo (95% CI: 10.3, 23.5; n ¼ 24) vs 21.9 mo (95% CI: 16.6, NE; n ¼ 106), respectively. Median PFS was 5.9 mo (95% CI: 2.6, 14.5) vs 5.4 mo (95% CI: 4.2, 8.8) and ORR was 29% (95% CI: 13, 51) vs 21% (95% CI: 13, 30) in pts who did vs did not receive corticosteroids. Conclusions: In this analysis, irAEs did not negatively impact the survival benefit of atezo. Further investigation on the impact of corticosteroids on atezo efficacy in randomized trials is needed. Clinical trial identification: NCT02008227 Legal entity responsible for the study: F. Hoffmann - La Roche Ltd. Funding: F. Hoffmann - La Roche Ltd. Disclosure: J. von Pawel: Adboard: AbbVie, Pfizer, Bristol Myers Squibb, Novartis money paid to the institution. D. Cortinovis: Membership for AB for Roche, Novartis, MSD, BI. R. Dziadziuszko: Honoraria or consulting fees from Roche, Pfizer, Boehringer-Ingelheim, Clovis Oncology, Novartis, Astra-Zeneca, Tesaro. D.R. Gandara: Consultant: Genentech Clinical trial grant: Genentech. P. Conkling: Research funding is US Oncology Research. J. Goldschmidt: Honoraria: Amgen Consulting/ Advisory Role: Amgen Speakers Bureau: Bristol Myers-Squibb, Celgene. M. Kosty: Support limited to Institutional support for the reported trial and other clinical trials. No direct compensation to investigator. F.S. Braiteh: COI: speaking and consulting fees received from Genentech. S. Hu: Genentech employee and Roche stock. M. Ballinger: Employee of Genentech, Roche stock. H. Patel: Genentech Employee. M. Gandhi: Employee of Genentech. All other authors have declared no conflicts of interest.

1315P

Efficacy and safety data from patients with advanced non-squamous NSCLC and brain metastases from the nivolumab expanded access programme (EAP) in Italy

L. Crin o1, P. Bidoli2, F. Roila3, E. Cortesi4, M.C. Garassino5, F. Cappuzzo6, F. Grossi7, G. Tonini8, G. Sarobba9, G. Pinotti10, G. Numico11, R. Samaritani12, L. Ciuffreda13, A. Frassoldati14, M. Bregni15, A. Santo16, F.V. Piantedosi17, I. Alfonso18, F. De Marinis19, A. Delmonte20 1 Medical Oncology, Istituto Scientiﬁco Romagnolo per lo Studio e la Cura dei Tumori (IRST)- IRCCS, Meldola, Italy, 2Medical Oncology Unit, San Gerardo Hospital, Monza, Italy, 3Medical Oncology, Azienda Ospedaliera Sta Maria, Terni, Italy, 4Medical Oncology Unit, Policlinico Umberto I, Rome, Italy, 5Clinical Oncology, Isituto Nazionale dei Tumori, Milan, Italy, 6Oncology and Hematology, Ospedale Civile di Ravenna - S.ta Maria delle Croci, Ravenna, Italy, 7Medical Oncology Unit, IRCCS AOU San Martino IST-Istituto Nazionale per la Ricerca sul Cancro, Genoa, Italy, 8Oncology, Campus BioMedico di Roma, Rome, Italy, 9Medical Oncology, AUSL3, Nuoro, Italy, 10Oncology, Ospedale di Circolo Fondazione Macchi, Varese, Italy, 11Medical Oncology, Ospedale SS. Antonio e Biagio, Alessandria, Italy, 12Oncology, PTP Nuovo Regina Margherita, Rome, a della Salute e della Scienza di Torino, Turin, Italy, 14Medical Italy, 13Oncology, Citt Oncology, Azienda Ospedaliera di Ferrara St. Anna, Ferrara, Italy, 15Oncology, Ospedale di Busto Arsizio, Busto Arsizio, Italy, 16Medical Oncology, University of Verona, Verona, Italy, 17Oncology, Azienda Ospedaliera Dei Colli-Monaldi, Naples, Italy, 18AO del Colli, Ospedale Monaldi-Cotugno-CTO, Naples, Italy, 19Oncologia, Istituto Europeo di Oncologia, Milan, Italy, 20Oncology, Istituto Tumori della Romagna I.R.S.T., Meldola, Italy Background: Brain metastases are a very common secondary localization of disease in patients (pts) with lung cancer. The prognosis of these pts is still poor and they are usually excluded from clinical trials. The EAP offered an opportunity to evaluate nivolumab treatment in these patients outside of a controlled clinical trial in Italy. Methods: Nivolumab was available upon physician request for pts aged 18 years with a diagnosis of non-squamous non-small cell lung cancer (non-Sq-NSCLC) who had relapsed after a minimum of one prior systemic treatment for stage IIIB/stage IV nonSq-NSCLC. Nivolumab 3 mg/kg was administered intravenously every 2 weeks to a maximum of 24 months. Pts included in the analysis had received 1 dose of nivolumab and were monitored for adverse events using Common Terminology Criteria for Adverse Events. Pts with brain metastasis were eligible if asymptomatic, neurologically stable and either off corticosteroids or on a stable dose or decreasing dose of 10 mg daily prednisone. Results: Of 1588 patients with non-Sq-NSCLC participating in the EAP in Italy, 409 (26%) had asymptomatic and controlled brain metastases. With a median number of 7 doses (1-45) and a median follow-up of 6.1 months (0.1-21.9), the disease control rate was 40%, including 3 pts with a complete response, 65 patients with a partial response and 96 with stable disease. As of March 2017, median overall survival among patients with brain metastases was 8.1 (6.2-10.1) months. Among these pts, 117 were receiving steroid therapy at baseline and 74 received concomitant radiotherapy. Overall, safety profile and discontinuations for drug-related toxicity were consistent with what observed in the general population. Conclusions: These preliminary data showed efficacy of nivolumab in patients with non-Sq-NSCLC with brain metastases, with safety results consistent to what already reported in controlled clinical trials, thus supporting the use of nivolumab in this population with poor prognosis. Clinical trial identification: CA209966 Legal entity responsible for the study: Lucio Crin o Funding: None Disclosure: All authors have declared no conflicts of interest.

1316P

Efficacy and immune activation with PEGylated human IL-10 (AM0010) in combination with an anti-PD1 in advanced NSCLC: Update

D.J. Wong1, J.G. Schneider2, R. Aljumaily3, M.W. Korn4, K. Autio5, J.R. Infante6, M.R. Patel7, K. Papadopoulos8, A. Naing9, N. Gabrail10, P. Munster11, J. Goldman1, P. Van Vlasselaer12, A. Hung13, G. Brown13, M. Oft13, E.B. Garon14 1 Department of Medicine, University of California, Los Angeles, CA, USA, 2Oncology, Winthrop University Hospital, Mineola, USA, 3Hematology/Oncology, Oklahoma University Medical Centre, Oklahoma City, OK, USA, 4Cancer Center, UCSF, San Francisco, CA, USA, 5Genitourinary Oncology, MSKCC, New York, NY, USA, 6Medical Oncology, Sarah Cannon Research Institute/Tennessee Oncology, Nashville, TN, USA, 7 Oncology, Sarah Cannon Research Institute (Nashville, TN); Florida Cancer Specialists & Research Institute, Sarasota, FL, USA, 8Oncology, South Texas Accelerated Research Therapeutics (START), San Antonio, TX, USA, 9Medical Oncology, MD Anderson Cancer Center, Houston, TX, USA, 10Oncology, Gabrail Cancer Center, Ohio, USA, 11Cancer Center, UCSF Helen Diller Family Comprehensive Cancer Center, San Francisco, CA, USA, 12Corporate, ARMO BioSciences, Redwood City, USA, 13Clinical Development, ARMO BioSciences, Redwood City, USA, 14Oncology/Hematology, David Geffen School of Medicine at UCLA, Los Angeles, CA, USA Background: At therapeutic concentrations, AM0010 stimulates the cytotoxicity, survival and proliferation of intratumoral antigen activated CD8þ T cells in pre-clinical cancer models and in patients. AM0010 activates antigen stimulated CD8 T cells while

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx380 | 469

abstracts

Annals of Oncology

PD-1 inhibits them, providing a rationale for combining AM0010 with PD-1 inhibitors. Methods: 34 NSCLC pts. received AM0010 (10-20mg/kg QD, SC) with pembrolizumab (2mg/kg, q3wk IV; n ¼ 5) or nivolumab (3mg/kg, q2wk IV; n ¼ 29). Tumor responses were assessed by irRC. Immune responses were measured by analysis of serum cytokines (Luminex), activation of blood derived T cells (FACS) and peripheral T cell clonality (TCR sequencing). Results: Pts had a median of 2 prior therapies. Median follow-up is 12.9 mo (range 3.726.9). AMO010 plus anti-PD-1 was well tolerated. All TrAEs were reversible. G3/4 TrAEs included thrombocytopenia (8), anemia (7), fatigue (6), rash (4), pyrexia (2), hypertriglyceridemia (3) and pneumonitis (1). As of May 5 2017, 26 pts had at least 1 tumor assessment, and partial responses (PRs) were observed in 10 pts (38.5%). 12 patients had stable disease (SD: 46.1%). mPFS and mOS were not reached. Updated efficacy data will be available by Aug. 31 2017.

Table: 1316P Preliminary response data strati˚ed for PD-L1 (Study in progress) NSCLC (n ¼ 26)

PR, n (%) SD, n (%)

PD-L1 (22C3 IHC) (n ¼ 20) 75 years. Methods: Nivolumab was available upon physician request for pts aged 18 years who had relapsed after a minimum of one prior systemic treatment for stage IIIB/IV Non-Sq-NSCLC. Nivolumab 3 mg/kg was administered intravenously every 2 weeks to a maximum of 24 months. Pts included in the analysis had received 1 dose of nivolumab and were monitored for adverse events (AEs) using Common Terminology Criteria for Adverse Events. Results: Of 1588 Italian pts with advanced non-Sq-NSCLC participating in the EAP in Italy, 522 (33%) were 70 years and 232 (15%) were 75 years. The median follow up was 7.6 months (0.1-20.8) and 8.3 (0.1-20), respectively. For pts aged 70, median number of doses was 9 (1-44) and the disease control rate (DCR) was 48%, including 2 pts with a complete response (CR), 106pts with a partial response (PR) and 145 with stable disease (SD). For pts aged 75, median number of doses was 11 (1-39) and the DCR was 53%, including 58pts with a PR and 64 with SD. Among pts aged >70, 403 discontinued treatment for any reason, with only 25 (5%) discontinuation due to related AEs vs 177 discontinuation for any reason, with only 13 (6%) due to related AEs among pts aged >75. As of March 2017, median overall survival was 11.5 months (10.0-13.0) and 12.0 months (9.2-14.8) respectively in ptsaged 70 and 75. The efficacy, safety and drug-related discontinuation results are in line with what observed in the general population. Conclusions: These results suggest that elderly population receive similar benefit from nivolumab treatment both in term of efficacy and safety than younger patients, supporting the use of nivolumab in this subpopulation. Legal entity responsible for the study: Bristol-Myers Squibb Funding: Bristol-Myers Squibb Disclosure: All authors have declared no conflicts of interest.

1322P

Real life experience with nivolumab in patients (pts) with advanced non-squamous NSCLC (nSq-NSCLC) exhibiting KRAS mutations: The Italian Expanded Access Program (EAP)

F. Cappuzzo1, P. Bidoli2, R. Chiari3, A. Chirco4, D. Turci1, A. Ardizzoni5, A. Santoro6, C. Natoli7, G. Francini8, M. Giordano9, G. Borra10, C. Defferrari11, L. Livi12, A. Berruti13, M. Minelli14, E. Ricevuto15, A. Illiano16, G. Puppo17, A. Delmonte18, A. Misino19 1 Oncology and Hematology, Ospedale Civile di Ravenna - S.ta Maria delle Croci, Ravenna, Italy, 2Oncology and Hematology, Azienda Ospedaliera S. Gerardo U.O. Oncologia Medica, Monza, Italy, 3Medical Oncology, Azienda Ospedaliera di Perugia, Perugia, Italy, 4Oncology and Hematology, Ospedale Papa Giovanni XXIII, Bergamo, Italy, 5Medical Oncology, University Hospital Sant’Orsola, Bologna, Italy, 6Oncology and Hematology, Istituto Clinico Humanitas, Rozzano, Italy, 7Oncology and Hematology, a P.O. Clinicizzato Ss. Annunziata Universita’ Degli Studi, Chieti, Italy, 8Oncology, Unit Sanitaria Locale 7 Siena, Siena, Italy, 9Oncology, Presidio Ospedaliero Ospedale 10 a di Sant’Anna, San Fermo Della Battaglia, Italy, Oncology, AOU Maggiore della Carit Novara, Novara, Italy, 11Oncology, Ospedali Galliera, Genoa, Italy, 12Radioterapia, 13 Azienda Ospedaliera Universitaria Careggi, Florence, Italy, Medical Oncology, University of Brescia, Spedali Civili, Brescia, Italy, 14Oncology, Ospedale San Giovanni Addolorata, Rome, Italy, 15Oncology, Rete Oncologica ASL1 Abruzzo, L’Aquila, Italy, 16 Oncology, AO del Colli, Monaldi-Cotugno-CTO, Naples, Italy, 17Oncology, AOU Pisana, Pisa, Italy, 18Department of Medical Oncology, Istituto Scientiﬁco Romagnolo per lo Studio e la Cura dei Tumori (IRST) IRCCS, Meldola, Italy, 19Oncology, Istituto Tumori Giovanni Paolo II, Bari, Italy Background: Nivolumab has been approved by different regulatory agencies worldwide for the treatment of nSq-NSCLC based on its superiority in Overall Survival (OS)

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx380 | 471

abstracts versus docetaxel in CheckMate 057 trial. In a pre-specified subgroup analysis of the same trial, this advantage was confirmed also in KRAS-mutationþ pts but the small number precluded any definitive conclusion. The Italian nivolumab EAP for non-sqNSCLC might represent an important source of information in that respect. The current analysis describes results of the use of nivolumab in the group of EAP pts with KRAS mutations. Methods: Nivolumab was provided upon physicians’ request for pts aged 18 years who had relapsed after a minimum of one prior systemic treatment for stage IIIB/stage IV non-Sq-NSCLC. Nivolumab 3 mg/kg was administered intravenously every 2 weeks for 24 months. Pts included in the analysis received 1 dose of nivolumab and were monitored for adverse events (AEs) using Common Terminology Criteria for Adverse Events. Results: In total, 1588 Italian pts with advanced nSq-NSCLC received at least one dose of nivolumab in the EAP across 168 sites. Among pts evaluated for KRAS mutation, 206 (39%) resulted positive. In this subgroup of pts, with a median follow-up of 7.7 months (0.1-21.2) and a median number of 8 doses (1-45), the best overall response rate (BORR) was 20%, including 2 pts with complete response and 39 pts with partial response, and the median OS was 10.7 months (8.6-12.8). These results were in line with those ones showed in the overall population (18% BORR and 11 months median OS, respectively). Conclusions: This analysis confirms, in a real word setting and in a much larger number of pts, the efficacy of nivolumab in KRAS-positive pts in CheckMate 057. Nivolumab represents a potentially effective therapeutic option for KRAS mutation, a molecular alteration for which there is currently no direct targeted therapy. Clinical trial identification: CA209-966 Legal entity responsible for the study: Lucio Crin o Funding: None Disclosure: F. Cappuzzo: Consultant and participation in advisory boards for BMS, Roche, Pfizer, AZ. All other authors have declared no conflicts of interest.

1323P

Baseline corticosteroids (CS) could be associated with absence of benefit to immune checkpoint inhibitors (ICI) in advanced non-small cell lung cancer (NSCLC) patients

G. Martınez Bernal1, L. Mezquita2, E. Auclin3, R. Ferrara2, D. Planchard2, J. Remon Masip2, J. Lahmar2, M-E. Boucher2, C. Caramella4, J. Adam5, A. Gazzah6, J-C. Soria6, B. Besse2 1 Medical Oncology, Hospital Universitario Virgen del Rocio, Seville, Spain, 2Medical Oncology, Institut Gustave Roussy, Villejuif, France, 3Medical and Gastrointestinal Oncology, Hopital European George Pompidou, Paris, France, 4Radiology, Institut Gustave Roussy, Villejuif, France, 5Pathology, Institut Gustave Roussy, Villejuif, France, 6 Drug Development Department (DITEP), Institut Gustave Roussy, Villejuif, France Background: Concomitant use of corticosteroids (CS) during immune checkpoint inhibitors (ICI) therapy are not recommended, but their real impact on ICI efficacy remains unknown. The aim of this study was to assess the impact of CS on ICI outcomes in NSCLC pts. Methods: Baseline CS intake and dose, patient characteristics and outcome were retrospectively collected in patients treated with PD1/PDL1 inhibitors from Nov. 2012 to Mar. 2017 in our Institute. Primary endopoints were overall survival (OS) and disease control rate (DCR: complete response þ partial response þ stable disease) and secondary endpoint was progression free survival (PFS). Results: We enrolled 244 pts. Median age was 63 years (30-85), 158 (65%) were males, 212 (87%) smokers, 196 (80%) PS 0-1; 155 (64%) had non-squamous cells carcinoma. The median of prior lines was 1 (0-11). KRASmut and EGFRmut was present in 62 (25%) and 14 (6%) of NSCLC, 3 (1%) were ALKþ, 64 (26%) PDL1 þ (cut-off 1% of tumor cells), 24 (10%) PD-L1- and 156 (64%) PD-L1 unknown. In the whole population, the overall response rate (ORR) was 20% and DCR 50%. Median OS and PFS were 9 months (m) [6-12] and 2m [2-3], respectively. The median follow-up was 10m [7-12]. Sixty-six patients (27%) received CS at baseline. Main reasons for taking CS were dyspnea (48%) and brain metastasis (15%). The median dose of daily prednisone was 16.25 mg [5;32.75] and >20mg in 19 (29%). CS dose >20mg was an independent factor for poor OS [HR 1,013, 95% CI 1,006; 1,02, p < 0,0001). For patients taking CS > 20mg, the median OS was 3m [2-12] vs. 10m [7-15] for 20mg was 1m [1-4] vs. 3m [2-4] for < 20mg (p ¼ 0.002). CS > 20mg was also significantly associated with progressive disease (p ¼ 0.011). Conclusions: Baseline daily prednisone intake of at least 20mg is associated with poor outcomes in advanced NSCLC treated with ICI. Further prospective studies are awaited for validating the real impact of CS in ICI efficacy. Legal entity responsible for the study: Dr Benjamin Besse Funding: Institut Gustave Roussy Disclosure: D. Planchard: AstraZeneca Boehringer Ingelheim BMS Lilly MSD Pfizer Roche Novartis Chugai. J-C. Soria: AstraZeneca, Astex, Clovis, GSK, Gammamabs, Lilly, MSD, Mission Therapeutics, Merus, Pfizer, Pharmamar Pierre Fabre, RocheGenentech, Sanofi, Servier, Symphogen, Takeda. All other authors have declared no conflicts of interest.

472 | NSCLC, metastatic

Annals of Oncology 1324P

Practical effectiveness efficacy and safety of nivolumab for advanced non-small cell lung cancer: A retrospective multicenter analysis

K. Kobayashi1, I. Nakachi2, K. Naoki1, Y. Oyamada3, M. Nakamura4, T. Inoue5, H. Tateno6, F. Sakamaki7, K. Sayama8, T. Terashima9, H. Koh10, D. Arai2, H. Yasuda1, I. Kawada1, K. Soejima1, T. Betsuyaku1 1 Division of Pulmonary Medicine, Keio University, Tokyo, Japan, 2Pulmonary Division, Internal Medicine, Saiseikai Utsunomiya Hospital, Utsunomiya, Japan, 3Pulmonary Division, Internal Medicine, National Hospital Organization Tokyo Medical Center, Tokyo, Japan, 4Pulmonary Division, Internal Medicine, Tokyo Saiseikai Central Hospital, Tokyo, Japan, 5Pulmonary Division, Internal Medicine, Sano Kousei General Hospital, Sano, Japan, 6Pulmonary Division, Internal Medicine, Saitama City Hospital, Saitama, Japan, 7Pulmonary Division, Internal Medicine, Tokai University Hachioji Hospital, Hachiouji, Japan, 8Pulmonary Division, Internal Medicine, Kawasaki Municipal Hospital, Kawasaki, Japan, 9Pulmonary Division, Internal Medicine, Tokyo Dental College Ichikawa General Hospital, Ichikawa, Japan, 10Pulmonary Division, Internal Medicine, Tachikawa Hospital, Tachikawa, Japan Background: Nivolumab, an immune checkpoint inhibitor, is now a standard treatment for previously treated advanced non-small cell lung cancer. We aimed to evaluate the real-world efficacy and safety of nivolumab in a non-selective population and identify clinical characteristics that influence efficacy. Methods: One hundred and forty-two patients with previously treated advanced nonsmall cell lung cancer who were administered nivolumab at Keio University and affiliated hospitals (in Japan) between January and July 2016 were enrolled. Treatment responses and adverse events were retrospectively reviewed and clinical characteristics associated with nivolumab responses were evaluated using univariate and stratified analyses and the Cochran-Mantel-Haenszel test. Results: The objective response rate was 17.0%, while the proportion of patients with adverse events of any grade was 45.0%. Clinical characteristics such as age, sex, Eastern Cooperative Oncology Group performance status, histological type, presence or absence of central nervous system metastases, smoking status, and the number of previous lines of treatment, were not related to efficacy. However, EGFR/ALK mutation status was inversely associated with treatment response (P < 0.05). Prior radiotherapy also exhibited a positive association with treatment response (P ¼ 0.012). Although no significant difference was observed between current/ex-smokers and non-smokers (P ¼ > 0.05), a subgroup analysis revealed smoking status in pack-years to be significantly higher in responders than non-responders (P < 0.05). Nivolumab is effective and safe regardless of age or the number of previous lines of treatment. Conclusions: The objective response rate and adverse event profiles were comparable to those observed in previous clinical trials. An EGFR/ALK mutation-negative status and prior radiotherapy are key clinical characteristics that are statistically associated with a positive treatment response. Our findings may aid in the efficient immunotherapeutic management of advanced non-small cell lung cancer. We will add other measurements of the treatment activity such as progression-free survival by the congress. Legal entity responsible for the study: Keio University Funding: None Disclosure: All authors have declared no conflicts of interest.

1325P

Generalization and representativeness of phase III immune checkpoint inhibitor trials in NSCLC

S.H. Yoo, M. Kim, B. Keam, T.M. Kim, D-W. Kim, D.S. Heo Internal Medicine, Seoul National University Hospital, Seoul, Republic of Korea Background: Immune checkpoint inhibitors (ICBs) have become standard treatment in platinum-failed non-small cell lung cancer (NSCLC) based on several phase III studies. Recent randomized phase III trials have led to the approval of ICB. However, strict criteria for patient enrollment of phase III trials raise questions regarding generalization in the real world. The aim of this study was to evaluate whether pivotal phase III trials using ICB represent the real world NSCLC patients. Methods: We reviewed the inclusion/exclusion criteria of 3 practice changing phase III trials (CheckMate057, CheckMate017, KEYNOTE-010). Availability of tumor tissue and other exclusion criteria for KEYNOTE-010 were additionally checked. We retrospectively analyzed the database of stage IIIB or IV NSCLC patients diagnosed from 2011 to 2013 at Seoul National University Hospital (cohort 1). We also analyzed the criteria in 53 NSCLC patients who have treated with nivolumab or pembrolizumab as a routine practice (cohort 2). Results: Among the 715 NSCLC patients in cohort 1, 499 (69.9%) were ineligible for 3 trials. Reasons for ineligibility were as follows: no platinum doublet 23.6%, lack of tissue 22.7%, the Eastern Cooperative Oncology Group performance status > 1 14.1%, steroid use 18.2%, active central nervous system metastasis 8.3%, hepatitis B or C virus/ human immunodeficiency virus 8.0% and no measurable lesion 7.3%. EGFR mutation was more common in ineligible group than eligible group (44.7% vs 19.7%, P < 0.001) In cohort 2 which comprise 53 patients who received ICB as a routine practice, 67.9% were classified as ineligible group. Treatment outcomes of ICB in cohort 2 seems to be inferior than those of 3 trials: response rate of 11.3%, disease control rate of 26.4%, and median progression-free survival of 1.67 months.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Conclusions: Only 30.2% of NSCLC patients were eligible for ICB phase III trial. Although ICB have approved for all platinum-failed NSCLC, the real world efficacy of 60% ineligible patients were still unknown. These findings suggest that a huge gap between the practice changing phase III trials and real world NSCLC patients. Legal entity responsible for the study: Bhumsuk Keam Funding: None Disclosure: All authors have declared no conflicts of interest.

1326P

Long-term follow-up results of stage III-IV non-small-cell lung cancer (NSCLC) patients treated with an epitope derived from Indoleamine 2,3 Dioxygenase (IDO) in a phase I study

J.W. Kjeldsen1, T.Z. Iversen2, L.E. Noerregaard2, A. Mellemgaard2, M.H. Andersen1, I-M. Svane1 1 Department of Hematology and Department of Oncology, Herlev Hospital, University of Copenhagen, Center for Cancer Immune Therapy, Herlev Hospital, Herlev, Denmark, 2 Department of Oncology, Herlev and Gentofte Hospital, Herlev, Denmark Background: A long-term follow up on an earlier published clinical trial of 15 stage IIIIV NSCLC patients treated with and IDO peptide vaccine.1 Methods: 15 patients with stage III-IV NSCLC in disease stabilization after standard chemotherapy were treated with subcutaneous vaccinations (100 mg IDO5 peptide, seq ALLEIASCL, in 900 ml Montanide). Patients were enrolled from 2010 to 2012 and treated biweekly for 2.5 months and thereafter monthly until progression or up to 5 years. As published in Clin Cancer Res 2013, the vaccine was well tolerated and a longlasting PR þ SD (>8.5 months) was seen in 47% of the patients. A long-term follow-up has been made, investigating the long term clinical benefit and immunity. Results: 3 of the 15 patients are still alive (May 2017) corresponding to a 5-year overall survival of 20%. One was excluded due to progression after 11 months; the other two have continued vaccination for 5 years and have both received 56 vaccines in total. One of the two patients developed a partial response of target lesions in the liver 15 months after the first vaccine and has been in stable disease ever since. The other patient had a solitary metastasis in a retroperitoneal gland at baseline and at the 1st evaluation scan the patient had no sign of malignancy and has been tumour free ever since. The vaccine was well tolerated for all 5 years. Analyses of PBMCs every 3rd to 6th month during treatment of the two long-term responders demonstrated a stable CD8þ T-cell population. The percentage of NK cells, MDSCs and Tregs were stable over time, demonstrating no sign of toxicity of the immune cells in the blood. Presence of IDO-specific CD8þ T cells were demonstrated by IFN-c Elispot and could be detected in both patients at several time points during vaccinations. Conclusions: The vaccine was well tolerated with no severe toxicity for administration up to five years. Two of 15 patients are long-term survivors with ongoing clinical response five years after 1. vaccination. 1. Iversen, T. Z. et al. Long-lasting disease stabilization in the absence of toxicity in metastatic lung cancer patients vaccinated with an epitope derived from indoleamine 2,3 dioxygenase. Clin. Cancer Res. 20, 221–232 (2014). Clinical trial identification: NCT01219348 Legal entity responsible for the study: Herlev and Gentofte Hospital, Center for Cancer Immune Therapy, Department of Hematology and Oncology, Denmark Funding: IO Biotech, Denmark Disclosure: All authors have declared no conflicts of interest.

1327P

The Lung Immune Prognostic Index (LIPI), a predictive score for immune checkpoint inhibitors in advanced non-small cell lung cancer (NSCLC) patients

L. Mezquita1, E. Auclin2, M. Charrier3, R. Ferrara1, J. Remon Masip1, D. Planchard1, S. Ponce Aix1, L. Paz-Ares4, J. Lahmar1, L. Leroy5, C. Audigier-Valette6, J. Zeron-Medina Cuairan7, P. Garrido Lopez8, S. Brosseau9, J. Mazieres10, C. Caramella11, J. Adam12, N. Chaput3, J-C. Soria13, B. Besse1 1 Medical Oncology, Institut Gustave Roussy, Villejuif, France, 2Medical Oncology, Hopital European George Pompidou, Paris, France, 3Laboratory of Immunomonitoring in Oncology, Institut Gustave Roussy, Villejuif, France, 4Medical Oncology Department, University Hospital 12 de Octubre, Madrid, Spain, 5Medical Oncology, Bergonie´, Bordeaux, France, 6Pneumologie, CH Toulon, Toulon, France, 7Medical Oncology, Vall d’Hebron University Hospital, Barcelona, Spain, 8Medical Oncology, Hospital Universitario Ram on y Cajal, Madrid, Spain, 9Thoracic Oncology Department, Hopital Bichat Claude Bernard, Paris, France, 10Thoracic Oncology Department, Toulouse University Hospital, Toulouse, France, 11Radiology, Institut Gustave Roussy, Villejuif, France, 12Pathology, Institut Gustave Roussy, Villejuif, France, 13Department of Drug Development, Gustave Roussy Cancer Campus, Villejuif, France Background: Derived NLR (neutrophils/(leucocytes-neutrophils) ratio) and lactate dehydrogenase (LDH) have been correlated to immune checkpoint inhibitors (ICI) benefit in melanoma. We tested whether baseline dNLR and LDH, as a score, could have the same role in advanced NSCLC patients. Methods: Baseline dNLR and LDH were retrospectively collected in 466 patients treated with ICI from Nov. 2013 to Jan. 2017, in a training cohort (N ¼ 161) and a

Volume 28 | Supplement 5 | September 2017

validation cohort (N ¼ 305) from 8 European centers. As a control, a cohort (N ¼ 162) treated only with chemotherapy between Nov. 2012 and Jul. 2016 from 2 centers. The primary endpoint was overall survival (OS), and secondary endpoints were progression free survival (PFS) and disease control rate (DCR). Results: In the immunotherapy cohort, 301 patients (65%) were males, 422 (90%) smokers and 401 (87%) with PS 1, with median age 63 years; 270 (58%) had adenocarcinoma and 159 (34%) squamous; 85 (18%) were KRASmut, 19 (4%) EGFRmut and 6 (1%) ALK positive. PDL1 was 1% by immunohistochemistry in 96 (74%), negative in 33 (26%) and unknown in 337 patients. The median of prior lines was 1 (011). In the training cohort, the median PFS and OS were 3 months (m) [2-4] and 10m [8-13]. dNLR>3 and LDH> Upper Limit of Normal (ULN) were independent factors for OS (HR 2.22, 95% CI 1.23-4.01; HR 2,51, 1.32-4.76, respectively). According to dNLR>3 and LDH>ULN, LIPI categorized 3 groups (good: 0 factor, intermediate: 1 factor, poor: 2 factors), which correlated with outcome in both cohorts. Median OS for good, intermediate, and poor was 34m, 10m and 3m, respectively (p ¼ 0.0001). The PFS and DCR were also correlated (p ¼ 0.001, p ¼ 0.005). Same results were observed in the validation cohort for OS (p ¼ 0.004), PFS and DCR (both p ¼ 0.005), but not in the chemotherapy cohort for both factors analyzed. Conclusions: Baseline LIPI, combining dNLR>3 and LDH>ULN, predicts the resistance to ICI acording to OS, PFS and DCR, but it is not correlated with the efficacy of chemotherapy, suggesting LIPI as a predictive score in ICI treatment. Legal entity responsible for the study: Prof Benjamin Besse Funding: None Disclosure: J-C. Soria: Consultancy fees from AstraZeneca, Astex, Clovis, GSK, Gammamabs, Lilly, MSD, Mission Therapeutics, Merus, Pfizer, Pharmamar, Pierre Fabre, Roche-Genentech, Sanofi, Servier, Symphogen, Takeda. All other authors have declared no conflicts of interest.

1328P

Association of single nucleotide polymorphisms with efficacy in nivolumab-treated NSCLC patients

E.A. Basak1, S. Bins1, S. El Bouazzaoui2, S.L.W. Koolen1, E. Oomen-De Hoop1, C. van der Leest3, A.A.M. van der Veldt1, S. Sleijfer1, R. Debets1, R.H.N. van Schaik2, R.H.J. Mathijssen1, J.G.J.V. Aerts4 1 Medical Oncology, Erasmus MC Cancer Institute, Rotterdam, Netherlands, 2Clinical Chemistry, Erasmus University Medical Center, Rotterdam, Netherlands, 3Pulmonology, Amphia Hospital, Breda, Netherlands, 4Pulmonology, Erasmus University Medical Center, Rotterdam, Netherlands Background: Proper patient selection for PD-1 checkpoint inhibitors is crucial given its limited efficacy in the majority of patients. We previously showed that a single nucleotide polymorphism (SNP) in the auto-immunity (AI) related PTPN11 gene (rs2301756) is associated with increased toxicity on nivolumab (Bins et al, ESMO immuno-oncology symp, 2016). The objective of the current analysis was to assess whether SNPs in PTPN11 and other genes associated with AI, which are putatively considered important in PD-1 influenced T-cell responses, are correlated with treatment efficacy of nivolumab in NSCLC patients. Methods: The association between 5 SNPs and efficacy was evaluated in 161 (chemotherapy pretreated) advanced NSCLC patients. Efficacy measures included early progressive disease (PD; 90 days after start), tumor response according to RECISTv1.1, PFS and OS. The SNPs were located on 4 genes, being PDCD1, PTPN11, ZAP70 and IFNG, respectively, encoding for the proteins PD-1, SHP-2 and ZAP70 and IFN-c. The best model for every SNP, being either a dominant, recessive, multiplicative or additive model, was selected. SNPs were analyzed in multivariable logistic regression models if they showed a p-value 0.0001).

474 | NSCLC, metastatic

Annals of Oncology Conclusions: While the first NGM evaluation in 2013 already showed a survival benefit of pts with activating genetic aberrations in EGFR and ALK, our current evaluation shows the heterogeneity of ALK-positive lung cancer pts and, for the first time to our knowledge, the impact of co-occurring mutations in these pts cohort. This work provides evidence for the efficacy of sequential ALK inhibitor treatment using next generation inhibitors and underlines the relevance of multiplex genotyping. Legal entity responsible for the study: University Hospital of Cologne for the Network Genomic Medicine Funding: None Disclosure: All authors have declared no conflicts of interest.

1331P

Detection of EGFR T790M in Asia-Pacific patients (pts) with EGFR mutation-positive advanced non-small cell lung cancer (NSCLC): Circulating tumour (ct) DNA analysis across 3 platforms

C. Zhou1, M. Wang2, Y. Cheng3, Y. Chen4, X. Ye5, Y. Sun5, X. Huang6, S. Patel6, Y. Chen5, Y-L. Wu7 1 Shanghai Pulmonary Hospital, Tongji University, Shanghai, China, 2Department of Respiratory Medicine, Peking Union Medical College Hospital, Beijing, China, 3 Department of Oncology, Jilin Cancer Hospital, Jilin, China, 4Cancer Centre of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China, 5China Development Unit, AstraZeneca, Shanghai, China, 6Biostatistics and Informatics, AstraZeneca, Cambridge, UK, 7Guangdong Lung Cancer Institute, Guangdong General Hospital and Guangdong Academy of Medical Sciences, Guangzhou, China Background: Osimertinib is an oral, potent, CNS active, irreversible EGFR-TKI approved to treat pts with T790M-positive NSCLC. Non-invasive methods to confirm presence of T790M are needed to identify pts who could benefit. Methods: AURA17 (NCT02442349) is a Phase II, single arm study investigating the safety and efficacy of osimertinib 80 mg once daily in an Asia-Pacific pt population with T790M positive advanced NSCLC, who had disease progression following EGFRV TKI therapy. Tumour tissue T790M status was centrally confirmed by cobas EGFR Mutation Test (Roche Molecular Systems). Where possible, matched plasma ctDNA V samples collected at screening were analysed for EGFR mutations using 3 tests: cobas EGFR Mutation Test v2.0 (cobas plasma), AmoyDx SuperARMS EGFR T790M Mutation Detection Kit (SuperARMS) and droplet digital PCR (ddPCR; in-house research assay). Results: Table summarises concordance data. R

R

Table: 1331P Sensitivity and specificity of plasma tests using cobas tissue test as the reference % (95% CI)

T790M L858R Exon 19 deletions

PPA NPA PPA NPA PPA NPA

cobas plasma (n ¼ 240)

SuperARMS (n ¼ 249)

ddPCR (n ¼ 249)

42 (34, 50) 83 (72, 91) 65 (54, 75) 100 (98, 100) 86 (80, 92) 97 (91, 99)

49 (41, 57) 78 (67, 86) NA* NA* NA* NA*

56 (48, 64) 73 (62, 83) 62 (51, 72) 99 (96, 100) 66 (58, 74) 98 (93, 100)

NPA, negative percent agreement (speciﬁcity); PPA, positive percent agreement (sensitivity); *SuperARMS used solely for detection of T790M Number of patients tested with cobas tissue test: 277 In the evaluable for response (EFR; 4 March 2016 data cut-off) set, pts with T790M-positive status by both tumour and plasma analysis had conﬁrmed objective response rates (ORR) with osimertinib (RECIST 1.1 by blinded independent central review) of 56% (95% CI 43, 69; 36/64 pts) using cobas plasma, 64% (52, 74; 49/77 pts) using SuperARMS, and 56% (45, 67; 49/87 pts) using ddPCR. ORR in the overall EFR tumour T790M-positive population was 60% (52, 68; 100/166 pts).

Conclusions: Using cobas tissue test as the reference, sensitivity for plasma T790M detection slightly increased with superARMS and ddPCR compared to cobas plasma test. Conversely, specificity slightly decreased. In pts with tumour T790M positive status, ORR with osimertinib was consistent across plasma tests, and with the overall tumour T790M-positive population. Biopsy is recommended for pts with a plasma T790Mnegative test result, where feasible. Clinical trial identification: NCT02442349 Legal entity responsible for the study: AstraZeneca Funding: AstraZeneca

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Disclosure: C. Zhou: Lecture honorarium: Eli Lily, AstraZeneca, Roche, Pfizer, Sanofi, Boehringer Iingelheim, Henrui Advisory Board: Roche, Boehringer Ingelheim, AstraZeneca. Y. Cheng: No financial interest in products or processes involved in our research. X. Ye, Y. Sun, X. Huang: Employee of AstraZeneca. S. Patel: Employee of, and shareholder in, AstraZeneca. Y-L. Wu: Speaker fees from AstraZeneca, Roche, Eli Lilly, Sanofi, Pfizer. All other authors have declared no conflicts of interest.

1332P

Influence of plasma T790M mutation on clinical decision after 1st generation EGFR-TKI resistance in a Real-world study

S. Zhang1, L. Zhu2, B. Xia2, X. Chen3, S. Ma2 Hangzhou Translational Medicine Research Center, Hangzhou First People’s Hospital, Nanjing medical university, Hangzhou, China, 2Radiation Oncology, Hangzhou First People’s Hospital, Nanjing Medical University, Hangzhou, China, 3Medical Oncology, Hangzhou First People’s Hospital, Nanjing Medical University, Hangzhou, China 1

Background: T790M mutation detection in circulating tumor DNA (ctDNA) has shown great potential in clinical application. However, few studies reported the influence of plasma T790M mutation on selection of clinical treatment and survival time after first generation TKI resistance. Methods: 307 patients with advanced or recurrent NSCLC who had progressed during EGFR-TKIs treatment were enrolled prospectively (NCT02418234) from March 2015 to March 2016. Blood samples were drawn within two weeks from PD occurred. T790M mutations were evaluated by droplet digital PCR. We undertook follow-up every 3 months by phone till April 2017. The median follow-up time was 11 months (range, 2 to 22 months). Results: Our results showed that the median survival time after TKI progression was 17.5 months (95%CI, 15-20 months) and six kinds of treatments were used in these patients, including continuation of TKI (27.0%), AZD9291 (27.7%), chemotherapy with or without radiotherapy (15.3%), TKI combined with chemo/RT (6.5%), switch to another TKI (2.6%) and best supportive care (11.4%). 88.6% of patients received subsequent treatment. T790M- patients were likely to receive continuation of original TKIs, which accords for 29.5% (52/176), the percentage of switch to another TKI is the lowest (2.8%, 5/176). In T790Mþ patients, AZD9291 is the first choice as the subsequent treatment, which accords for 38.9% (51/131). Switch to another TKI (2.2%, 3/131) and TKI combined with chemo/RT (6.1%, 8/131) is the least selection. Although most T790M- patients received continuation of original TKIs, combination of TKI and chemotherapy/radiotherapy seems to be a better choice, which got the longest survival than other treatment. For T790Mþ patients, patients who choose AZD9291 had the longest survival. Conclusions: T790M status in ctDNA have the great influence on clinical decision of the subsequent treatment, AZD9291 is the most frequent choice for the plasma T790Mþ patients, which contributed the longest survival after 1st generation EGFRTKI resistance. Clinical trial identification: NCT02418234 Legal entity responsible for the study: Shenglin Ma Funding: Projects of Medical and Health Technology in Zhejiang Province (WKJ-2J1532) Disclosure: All authors have declared no conflicts of interest.

1333P

Detectability of RET fusions by amplicon-based next generation sequencing in nationwide lung cancer genomic screening project: LCSCRUM-Japan

Y. Usui1, K. Kirita1, S. Matsumoto1, Y. Ohe2, M. Nishio3, T. Seto4, M. Kodani5, K. Taima6, Y. Hattori7, T. Kohno8, K. Yoh1, K. Goto1 1 Thoracic Oncology, National Cancer Center Hospital East, Kashiwa, Japan, 2 Department of Thoracic Oncology, National Cancer Center Hospital, Tokyo, Japan, 3 Department of Thoracic Medical Oncology, The Cancer Institute Hospital of Japanese Foundation for Cancer Research, Tokyo, Japan, 4Department of Thoracic Oncology, National Kyushu Cancer Center, Fukuoka, Japan, 53rd Internal Medicine, Tottori University Hospital, Yonago, Japan, 6Department of Respiratory Medicine, Hirosaki University, Hirosaki, Japan, 7Department of Thoracic Oncology, Hyogo Cancer Center, Akashi, Japan, 8Research Institute, National Cancer Center, Tokyo, Japan Background: RET fusions have been identified as a therapeutic target in non-small cell lung cancer (NSCLC). The establishment of molecular diagnostics for RET fusions is required for the development of molecular-targeted therapies for fusion-positive patients. We have operated molecular testing for various actionable gene alterations, including RET fusions, in our nationwide lung cancer genomic screening project in Japan (LC-SCRUM-Japan). Methods: Since 2013, RET fusions were analyzed by multiplex RT-PCR, and fusionpositive samples were confirmed by break-apart FISH. Since 2015, an amplicon-based next generation sequencing (NGS) system, OncomineTM Comprehensive Assay was also applied in our genomic screening. Results: As of February 2017, 244 institutions across Japan were participating, and a total of 4015 lung cancer patients including 3392 non-squamous (Non-Sq) NSCLCs were enrolled into the LC-SCRUM-Japan. Of the Non-Sq NSCLCs, available samples for multiplex RT-PCR and NGS were 94% (3186/3392) and 90% (1671/1852),

Volume 28 | Supplement 5 | September 2017

respectively. RET fusions were detected in 89 samples (2.8%), including 54 detected by NGS. The types of variants found were KIF5B-RET (n ¼ 50, 56%), CCDC6-RET (n ¼ 20, 22%) and undetermined (n ¼ 19, 21%). Of the 54 RET fusion-positive samples by NGS, the positivity was confirmed using break-apart FISH in 35 (92%) of 38 available samples. Among the 1335 samples available for both RT-PCR and NGS, the concordance rate between the two assays for the detection of RET fusions were 0.99. The detection sensitivity and specificity of the fusions in NGS assay were 0.98 and 0.99. Conclusions: Our nation-wide screening results revealed that this amplicon-based NGS assay showed high sensitivity and specificity for the detection of RET fusions and is clinically applicable for diagnosis of RET fusions in lung cancer. Legal entity responsible for the study: National Cancer Center Funding: Astra Zeneka, Eisai, Chugai Pharmaceutical, Pfizer, DAIICHI SANKYO, Astellas Pharma, Taiho Pharmaceutical, Ono Pharmaceutical, Kyowa Hakko Kirin, Eli Lilly Japan, Takeda Pharmaceutical, Novartis Pharma, Amgen Astellas BioPharma, MSD, Merck Serono. Disclosure: S. Matsumoto: Astra Zeneka, Eisai, Chugai Pharmaceutical, Pfizer, Daiichi Sankyo, Astellas Pharma, Taiho Pharmaceutical, Ono Pharmaceutical, Kyowa Hakko Kirin, Eli Lilly Japan, Takeda Pharmaceutical, Novartis Pharma, Amgen Astellas BioPharma, MSD, Merck Serono. Y. Ohe: Honorarium/Consultant/Expert Testimony/ Research Funding (Institution); AstraZeneca, Chugai, Lilly, ONO, BMS, DaiichiSankyo, Nipponkayaku, Boehringer, Bayer, Pfizer, MSD, Taiho, Clovis, Sanofi, Novartis, Kyorin, Dainippon-Sumitomo, Merck. M. Nishio: Novartis, Ono Pharmaceutical, Bristol-Myers Squibb, TAIHO Pharmaceutical, Eli Lilly, Pfizer, Astellas Pharma, AstraZeneca and Honoraria, Pfizer, Bristol-Myers Squibb, Chugai Pharmaceutical, Taiho Pharmaceutical, AstraZeneca. T. Seto: AstraZeneca, Astellas Pharma, Chugai, Daiichi Sankyo, Eisai, Eli Lilly, Merck Serono, MSD, Nippon Boehringer Ingelheim, Novartis Pharma, Pfizer Japan, Bristol-Myers Squibb, Kissei, Kyowa Hakko Kirin, Nippon Kayaku, Ono, Roche Singapore Pte, Taiho, Yakult. K. Yoh: AstraZeneca, Chugai Pharma, Boehringer Ingelheim, Lilly Japan, Bristol-Myers Squibb, Taiho Pharmaceutical, Ono Pharmaceutical, Pfizer. K. Goto: Astellas Pharma, Astra Zeneka, Amgen Astellas BioPharma, Eisai, Ono, Kyowa Hakko Kirin, Daiichi Sankyo, Taiho, Takeda, Chugai, Novartis, Pfizer, MSD, Eli Lilly, Merck Serono, BristolMyers Squibb, Life Technologies. All other authors have declared no conflicts of interest.

1334P

A comparison of Bronchial Washing Fluid (BWF) and histologic samples in the analysis of EGFR mutation in NSCLC patients

X. Zhang1, M. Ye1, C. Li1, Q. Hu2, M. Du2, X. Wang2, Z. Gong3, D. Zhang3, Q. Hong1, J. Hu1, L. Mei1, Y. Hou2, X. Zhang1 1 Department of Pulmonary Disease, Shanghai Zhongshan Hospital, Shanghai, China, 2 Department of Pathology, Shanghai Zhongshan Hospital, Shanghai, China, 3 Shanghai Yunying Medical Technology, Co. Ltd, Shanghai, China Background: The aim of the current study was to examine the clinical application of BWF samples in detecting epidermal growth factor receptor (EGFR) mutations in a large sample set of NSCLC patients. Methods: In diagnostic bronchoscopic examinations, before or after biopsy to target lesions, subsequent bronchial washing by saline was performed. Thereafter, EGFR mutation testing for both supernatant and sediment of BWF and histologic tissues was performed via amplification refractory mutation system real-time PCR (ARMS RTPCR) assay. Results: A total of 127 cases of histologic and corresponding BWF samples of patients underwent bronchoscopy for suspected lung malignant tumor lesions on chest radiography were successfully obtained. Of these, 72 cases were pathologically confirmed to be NSCLC based on forceps biopsy samples and EGFR mutations were identified in 26 cases. In 70 of 72 cases, the results of EGFR mutation status were concordant for BWF and histologic samples, and the concordance rate was 97%. In 13 cases that were not pathologically diagnosed with NSCLC with forceps biopsy samples but other samples, five cases (38.46%) with EGFR mutated-type were detected by BWF. The overall EGFR mutation concordance rate between supernatant and sediment specimens was 100%. The detection of EGFR mutations with supernatant/sediment of BWF samples showed a sensitivity of 86.5%, a specificity of 100%. Conclusions: This study demonstrates a clear comparison of supernatant/sediment of BWF samples and histologic tissues for EGFR mutation testing with largest clinical samples to date. Both supernatant and sediment of BWF samples showed high credibility and concordance via highly sensitive PCR analysis. BWF is considered a simple, rapid and effective alternative for histologic samples in EGFR mutation testing. Legal entity responsible for the study: Zhongshan Hospital Funding: None Disclosure: All authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx380 | 475

abstracts 1335P

Annals of Oncology

EGFR T790M detection in TKI-naı¨ve NSCLCs carrying sensitive EGFR mutations

G. Pinotti1, R. Cerutti2, N. Sahnane2, L. Lettig2, C. Albeni2, A. Tuzi1, F. Franzi2, A. Pastore1, F. Ogliari1, F. Sessa2, D. Furlan2 1 Oncology, Ospedale di Circolo Fondazione Macchi, Varese, Italy, 2Dept of Medicine and Surgery, University of Insubria-Ospedale del Circolo, Varese, Italy Background: To date, the frequency of EGFR T790M in TKI-naı¨ve patients remains unclear, ranging from 2% to 80% depending on the sensitivity and specificity of the methods. In this study we aimed to identify the frequency of EGFR T790M in NSCLCs before TKI treatment, comparing the detection rate of three highly sensitive molecular methods. Methods: Among 1100 NSCLCs (adenocarcinomas at stage IIIB or IV), we identified 130 NSCLCs with EGFR TKI-sensitive mutations, by MALDI-TOF mass spectrometry V (MALDI-TOF MS) and Myriapod Lung Status kit. The diagnostic performance in detecting de-novo EGFR T790M in these 130 tumors was evaluated comparing three V methods, namely MALDI-TOF MS, Real Time AS-PCR (Easy EGFR kit) and ddPCRTM (QX200 Droplet Digital PCR, PrimePCRTM Assays). Sensitivity and specificity of each method were defined by using a DNA reference standard set (Horizon). Limit of blank (LOB) of AS-PCR and ddPCR were determined by measuring replicates of 16 wild-type EGFR DNA samples obtained from peripheral blood lymphocytes and from formalin-fixed paraffin embedded (FFPE) normal lung tissues. Results: Comparison of the three methods was possible for 91 of the 130 NSCLCs. Overall, we identified a total of 16 de-novo T790M in the analyzable tumors (18%). In detail, 4 cases were identified by MALDI-TOF MS and confirmed by AS-PCR and ddPCR. Two T790M-mutated cases were additionally detected by AS-PCR. ddPCR confirmed EGFR T790M in all these tumors and additionally identified 10 mutated cases. Most of mutated cases showed a mutant-allele frequency between 5% and 0.1%. Titration experiments using a DNA reference standard set demonstrated higher sensitivity of ddPCR (0.1%) than AS-PCR (1%) and MALDI-TOF MS (5%). Analysis of wild-type EGFR DNA from FFPE samples was crucial for the determination of LOB of ddPCR in order to maximize sensitivity, avoiding loss of specificity. Conclusions: In this study, 18% of TKI-naı¨ve NSCLCs show EGFR T790M mutation together with an EGFR activating mutation. Most of mutated cases showed a mutant-allele frequency between 5% and 0.1%. ddPCR is a robust method enabling the detection of mutant-allele frequencies as low as 0.1%. However, a careful preliminary evaluation of the specificity of this test is mandatory, especially when FFPE tissues are investigated. Legal entity responsible for the study: Asst-Sette Laghi Ospedale di Circolo Varese Funding: AstraZeneca Disclosure: All authors have declared no conflicts of interest. R

ARISTOT (Association de Recherche, d’Information Scientifique et The´rapeutique en Oncologie Thoracique) grant in 2016. This work received funding from the GrenobleAlpes University Hospital (DRCI REALK project) and from the French Institut National du Cancer (INCa). Intergroupe Francophone de Cance´rologie Thoracique, Association de Recherche, d’Information Scientifique et The´rapeutique en Oncologie Thoracique, Grenoble-Alpes University Hospital (DRCI REALK project), French Institut National du Cancer (INCa). Disclosure: A. Mc Leer: Research funding from Novartis and Pfizer, reimbursed for travel, accommodation by ThermoFisher, ZytoVision/Clinisciences, Astra-Zeneca, Novartis, Pfizer and Roche and has served as a consultant for Pfizer. M. Duruisseaux: Research funding from Novartis and Pfizer. Reimbursed for travel, accommodation, and/or other expenses by Novartis, Pfizer and Roche. A.C. Toffart: Reimbursed for travel, accommodation, and/or other expenses by Novartis, Pfizer and Roche. D. Moro-Sibilot: Research funding from Pfizer. Served as a consultant (advisory board) for Novartis, Pfizer and Roche. S. Lantuejoul: Consultant in advisory boards for BMS, MSD, Pfizer, Novartis, Astra-Zeneca, Roche, Boehringer Ingelheim. All other authors have declared no conflicts of interest.

R

1336P

ALK fusion variants detection by targeted RNA-next generation sequencing and clinical responses to crizotinib in ALK-positive nonsmall cell lung cancer

A. Mc Leer1, M. Duruisseaux2, J. Pinsolle2, S. Dubourg1, J. Mondet1, M. PhillipsHoulbracq2, N. Magnat1, J. Faure´3, A. Chatagnon3, F. de Fraipont1, M. Giaj Levra2, A.C. Toffart2, G. Ferretti4, E. Brambilla5, P. Hainaut1, D. Moro-Sibilot2, S. Lantue´joul5 1 Cancer Molecular Genetics Platform, CHU Grenoble - Hopital Michallon, La Tronche, France, 2Thoracic Oncology, CHU Grenoble - Hopital Michallon, La Tronche, France, 3 Molecular Biology Platform, CHU Grenoble - Hopital Michallon, La Tronche, France, 4 Radiology and Imaging Dept, CHU Grenoble - Hopital Michallon, La Tronche, France, 5 Pathology Department, CHU Grenoble - Hopital Michallon, La Tronche, France Background: The aim of the present study was to assess the yields of an amplicon-based parallel sequencing (RNA-seq) assay for ALK fusion transcript variants detection in comparison with immunohistochemistry (IHC) and fluorescent in-situ hybridization (FISH) in a selected population of ALK-positive and ALK-negative non-small cell lung cancer (NSCLC) cases, and to evaluate the impact of the ALK variant on crizotinib efficacy. Methods: Our population comprised fifty-three NSCLC cases positive for ALK by IHC and/or FISH and twenty-three ALK-negative samples. For the ALK-positive samples, a distinction was made between ‘truly’ IHC/FISH positive or ‘truly’ IHC/FISH negative samples, and the samples for which the IHC and/or FISH were equivocal (IHC) or borderline-positive (FISH). Results: On the overall population, RNA-seq sensitivity and specificity were of 80% and 100%, respectively when IHC and FISH were combined and of 100% for both for ‘truly’ positive samples. Interestingly, this assay also appeared to be a promising rescue technique in equivocal and/or borderline-positive IHC/FISH cases. Moreover, when crizotinib efficacy was evaluated according to the type of ALK variant detected, better clinical outcomes were observed in crizotinib-treated patients with EML4-ALK v1/v2/ others variants compared to v3a/b variants. A lack of efficacy of crizotinib was noted in KLC1-ALK variants. Conclusions: RNA-seq detects ALK rearrangements with a high sensitivity and specificity and may be particularly useful in equivocal/borderline-positive IHC/FISH cases. In addition, it offers a unique opportunity to identify ALK fusion variants and to evaluate their predictive value for ALK inhibitors efficacy. Legal entity responsible for the study: Anne Mc Leer Funding: MD was the recipient of the IFCT (Intergroupe Francophone de Cance´rologie Thoracique) Alain Depierre Grant in 2014. JP was the recipient of the

476 | NSCLC, metastatic

1337P

A meta-analysis on epidemiology of ROS1 rearrangement in Asian and non-Asian population

V. Gupta1, N. Godre1, M. Alam2 Medical & Research Division, Pﬁzer Limited, Mumbai, India, 2Oncology, Pﬁzer Limited, West Ryde, Australia 1

Background: ROS1 is now recognised as a definite molecular target in NSCLC.1 Studies showed that ROS1 positive patients are significantly younger & more likely to be nonsmokers.2 Previous studies found epidermal growth factor receptor (EGFR) mutation in NSCLC was significantly higher in Asian population compared to non-Asian.3 Objective of the study is to evaluate difference between epidemiological parameters of ROS1 rearrangements in Asian & non-Asian patients by performing meta-analysis. Methods: We systematically searched databases like PubMed & e-journals since 2011. Statistical analysis for this study was done on 20 studies (11,665 patients) conducted globally & open source software R (version 3.3.2) was used for meta-analysis. Results: Global prevalence of ROS1 rearrangement was 2% (95% CI: 0.016- 0.026) & it was higher in Asian [2.2% (95% CI: 0.016- 0.029)] than non-Asian [1.9% (95% CI: 0.0120.027)] (p ¼ 0.92). Mean age of Asian & non-Asian was 54.5 yrs & 59 yrs respectively. The prevalence rate in non-Asian females, was significantly higher [3.8% (95% CI: 0.011-0.078)] than non-Asian males [0.7% (0.003- 0.012)] (p ¼ 0.003). Similarly, prevalence rate in Asian females [2.8% (95% CI: 0.019-0.038)] was higher than Asian males [2.1% (95% CI: 0.0120.033)] (p ¼ 0.88). Smokers in Asia were more likely to have ROS1 rearrangement [1.8% (95% CI: 0.005-0.037)] compared to non-smokers [0.6% (95% CI: 0.0004-0.0152)] (p ¼ 0.93). Whereas, prevalence rate amongst non-smokers in non-Asia [7.1% (95% CI: 0.039-0.110)] was significantly higher than smokers [0.7% (95% CI: 0.002-0.007)] (p ¼ 0.008). Clinical stage IV was more common in both population than other stages. In 5 out of 20 studies, ROS1 positivity was higher (3-12.5%) in enriched (EGFR-ve/ALK-ve) population [non-smokers 6.5% (95% CI: 0.046-0.083)] compared to overall population. Conclusions: Our meta-analysis showed that ROS1 gene rearrangement was more prevalent in NSCLC, females, non-smokers & patients with clinical stage IV while there is no significant difference in Asians vs non-Asians. References 1] Bubendorf et al. Virchows Arch. 2016; 469:489–503. 2] Bergethon et al. JCO. 2012; 30(8): 863-870. 3] Expert Opin Pharmacother. 2015 Jun;16(8):1167-76. Legal entity responsible for the study: Pfizer Limited Funding: None Disclosure: V. Gupta, N. Godre: Employee of Pfizer Ltd. Receipt of grants/research support: Pfizer Ltd. M. Alam: Employee and stock ownership: Pfizer Ltd. Receipt of grants/research support: Pfizer Ltd.

1338P

EGFR mutation detection in plasma cell-free DNA correlates with clinical outcomes in non-small cell lung cancer

J. Shi1, J. Mong1, T.M. Chin2, Y.H. Lim1, W.L. Tan3, C.K. Toh3, H.S. Tan3, S-W. Wong4, A. Tee5, D. Chan6, K. Wong6, S. Yeap7, L. Ngo8, Y-O. Tan6, M-H. Tan1 1 Biodevices and Diagnostics, Insitute of Bioengineering and Nanotechnology (IBN), Singapore, 2National University Cancer Institute, National University Cancer Institute, Singapore, 3Division of Medical Oncology, National Cancer Centre, Singapore, 4The Cancer Centre, The Cancer Centre, Singapore, 5Department of Respiratory & Critical Care Medicine, Changi General Hospital, Singapore, 6Singapore Oncology Consultants, Singapore Oncology Consultants, Singapore, 7Novena Cancer Centre, Novena Cancer Center-Mount Elizabeth Specialist Centre, Singapore, 8Rafﬂes Cancer Centre, Rafﬂes Hospital, Singapore Background: Cell-free DNA (cfDNA) testing of epidermal growth factor receptor mutations (EGFRmut) is being investigated as an adjunct for diagnosis and monitoring in nonsmall cell lung cancer (NSCLC) patients. The performance of various amplicon-based targeted next-generation sequencing (NGS) methods, both with and without error correction, is of high interest. Outcomes of error-corrected NGS in plasma EGFRmut testing have not been previously independently reported. We deployed an in-house amplification-refractory mutation system PCR (ARMS-PCR) assay in a prospective study, benchmarking its performance against two NGS platforms in a patient subset.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Methods: An ultrasensitive ARMS-PCR assay for hotspot EGFRmut was established, with detection limits between 0.02% and 0.1%. A total of 134 plasma samples were prospectively analysed from 68 patients with metastatic lung adenocarcinoma at diagnosis or progression, recruited between Jan 13-Apr 17 from 5 centres, with serial monitoring of plasma EGFRmut till radiologic progression in one centre. We further evaluated the performance of ARMS-qPCR assay, AmpliSeq Lung and Colon NGS assay and Oncomine Lung cfDNA NGS assay in 29 NSCLC and 20 healthy plasma controls. Results: Concordance rate between cfDNA and tumor was 83.8%, with sensitivity 80.0%, specificity 94.4%, positive predictive value 97.6%, and negative predictive value 63.0%. Dynamic monitoring of plasma EGFRmut levels demonstrated rising levels a median of 2.1 months [0.9-3.9] before radiological progression. This detection also held true for tissue EGFRmut positive patients negative for plasma EGFRmut at study entry. 20 of 49 patients at progression were plasma T790M-positive, and clinical benefit rates were 91.0% for osimertinib-treated patients. Evaluation of ARMS-PCR and NGS platforms yielded an average concordance rate, sensitivity and specificity was 85.9%, 63.4%, 92.3% (ARMSqPCR), 87.2%, 47.8%, 100% (Ampliseq) and 84.1%, 83.1%, 87.4% (Oncomine). Conclusions: ARMS-PCR provides a useful diagnostic and monitoring adjunct for NSCLC EGFRmut patients. Amplicon-based targeted next-generation sequencing approaches with error correction is a promising approach requiring additional validation. Legal entity responsible for the study: Institute of Bioengineering and Nanotechnology Funding: Agency for Science, Technology and Research (A*STAR) Disclosure: All authors have declared no conflicts of interest.

1339P

A large prospective cohort study of the clinical features of advanced lung cancer harboring HER2 aberrations (HER2-CS STUDY)

T. Tanaka1, H. Yoshioka1, A. Bessho2, N. Ishikawa3, M. Yamasaki4, T. Shibayama5, K. Aoe6, T. Kozuki7, N. Fujimoto8, K. Genba9, Y. Ueda10, M. Inoue11, T. Murakami12, S. Kuyama13, H. Kawai14, K. Fujitaka15, K. Inoue16, M. Takata17, H. Yanai18, K. Kiura19 1 Respiratory Medicine, Kurashiki Central Hospital, Kurashiki, Japan, 2Respiratory Medicine, Japanese Red Cross Okayama Hospital, Okayama, Japan, 3Respiratory Medicine, Hiroshima Prefectural Hospital, Hiroshima, Japan, 4Respiratory Medicine, Hiroshima Red Cross Hospital and Atomic-Bomb Survivors Hospital, Hiroshima, Japan, 5 Respiratory Medicine, NHO Okayama Medical Center, Okayama, Japan, 6Respiratory Medicine, Yamaguchi-Ube Medical Center, Ube, Japan, 7Respiratory Medicine, NHO Shikoku Cancer Center, Matsuyama, Japan, 8Medical Oncology, Okayama Rosai Hospital, Okayama, Japan, 9Medical Oncology, Chugoku Central Hospital, Fukuyama, Japan, 10Respiratory Medicine, Kagawa Prefectural Central Hospital, Takamatsu, Japan, 11 Thoracic Surgery, Shimonoseki City Hospital, Shimonoseki, Japan, 12Respiratory Medicine, Japanese Red Cross Society Himeji Hospital, Himeji, Japan, 13Respiratory Medicine, Iwakuni Medical Center, Iwakuni, Japan, 14Respiratory Medicine, Okayama Saiseikai General Hospital, Okayama, Japan, 15Respiratory Internal Medicine, Hiroshima University, Hiroshima, Japan, 16Respiratory Medicine, Ehime Prefectural Central Hospital, Matsuyama, Japan, 17Department of Late Effects Studies, Radiation Biology Center, Kyoto University, Kyoto, Japan, 18Diagnostic Pathology, Okayama University Hospital, Okayama, Japan, 19Respiratory Medicine, Okayama University Hospital, Okayama, Japan Background: HER2 is a potential driver oncogene. HER2-targeted precision therapy has been tested in NSCLC. However, the demographics of HER2-positive NSCLC have not been defined systematically.

Methods: Pts with advanced NSCLC were registered. HER2-IHC and FISH assays were performed with commercial kits. HER2 mutations were identified by the direct sequencing. The aim of this study was to clarify the frequency, characteristics and outcome of HER2-positive NSCLC. Results: Of 1,126 tumors screened (Table A), 34 (3.0%) were IHC3þ, and 34 (3.0%) were IHC2þ/FISHþ. Among the 724 EGFR wild-type tumors, 21 (2.9%) were HER2mutant tumors, including A775_G776insYVMA (n ¼ 15). Interestingly, the IHC3þ tumors and mutant tumors were entirely exclusive. Female pts had HER2 mutant tumors more frequently, while IHC/FISHþ tumors were detected more often in males (Table B). HER2-positive tumors had similar survival outcome to triple negative tumors, but significantly worse prognoses than EGFR-mutant and ALK-positive tumors (p < 0.05 each). Conclusions: This is the first prospective study showing a small fraction of NSCLC possessed HER2 aberrations. HER2-positive tumors had relatively poor prognosis. NSCLCs with HER2 IHC3þ and mutation seemed to be distinct subsets. Clinical trial identification: UMIN registration number 000017003 Legal entity responsible for the study: HER2-CS Network Funding: Japan Agency for Medical Research and Development Disclosure: H. Yoshioka: Honoraria: Eli Lilly, Chugai Pharma., Boehringer Ingelheim, Taiho Pharmaceutical, Pfizer, AstraZeneca, Bristol-Myers Squibb, Ono Pharmaceutical, Takeda Pharma. T. Shibayama: Honoraria: Meiji Seika Pharma Co., Ltd. Astra Zeneca K.K. Boehringer Ingelheim Pharmaceuticals, Inc. Eli Lilly Japan K.K. K. Aoe: Eli Lilly, Astrazeneca, BMS, Ono Pharmaceutical Co., Ltd. T. Kozuki: Honoraria: Chugai Pharma., AstraZeneca, Eli lilly Japan, Pfizer, Ono Pharm., BristolMyers Squibb, Kyowa-Hakko Kirin, Boehringer Ingelheim. N. Fujimoto: Honoraria: Kissei Co. Ltd. S. Kuyama: Honoraria: Chugai Pharma., AstraZeneca, Pfizer, Ono Pharm., Boehringer Ingelheim, Meiji Seika Pharma. H. Yanai: Honoraria: Bayer Yakuhin, Ltd, Janssen Pharmaceutical K.K., Kyowa Hakko Kirin Co., Ltd., Becton, Dickinson and Company, Fujiyakuhin Co., Ltd. K. Kiura: Honoraria: Chugai Pharmaceutical Co., Ltd. Pfizer Japan Inc. Novartis Pharma K.K. Taiho Pharmceutical Co., Ltd. Eli Lilly Japan K.K. All other authors have declared no conflicts of interest.

1340P

Transcriptomic analysis of bronchoalveolar lavage cells from advanced non-small cell lung cancer identifies overexpressed immunoglobulin genes of immunosuppressive implication

C-H.S. Kuo1, C-Y. Liu1, Y-L. Lo2, Y-W. Wang1, T-H. Wang3, C-T. Yang1 Thoracic Medicine, Chang Gung Memorial Hospital-Taipei, Taipei, Taiwan, 2Airway Disease, Chang Gung Memorial Hospital-Taipei, Taipei, Taiwan, 3Genomic Medicine Research Core Laboratory, Chang Gung Memorial Hospital-Taipei, Taipei, Taiwan

1

Background: Diverse pattern of inflammatory cells infiltration in the microenvironment of non-small cell lung cancers (NSCLCs) has key implication for successful immunotherapeutic approaches (Gajewski et al. Nat Immunol 2013). However, study of these cells remains challenging particularly in advanced disease where tumor resection is never possible. We hypothesized that transcriptomic study of bronchoalveolar lavage (BAL) cells is useful in this setting to identify characteristic gene expression of immunological significance. Methods: BAL cells were obtained from 13 patients of advanced NSCLC and 6 normal controls. In NSCLC group, lavage was performed from the lung segment where tumor was located. RNA was extracted and hybridized to Affymetrix HG-U133 plus2

Table: 1339P A. The Genotype-Specific Subsets* HER2 (n ¼ 88)

Age, median Sex (male) Smoking habit Non-Sq Stage III/IV

69 61 (69%) 58 (66%) 78 (89%) 51 (58%)

MST (mo) 1-yr OS rate 17.5 59% B. The Subsets of HER2 aberrations** IHC3 þ (n ¼ 34) Age, median Sex (male) Smoking 71 habit Non-Sq Stage III/IV 27 (79%) 24 (71%) 30 (88%) 17 (50%) MST (mo) 1-yr OS rate 10.5 46%

EGFR (n ¼ 358)

ALK (n ¼ 44)

Triple negative/ unknown (n ¼ 664)

69 142 (40%) 142 (40%) 351 (98%) 220 (61%)

62 21 (48%) 19 (43%) 44 (100%) 35 (80%)

69 516 (78%) 544 (82%) 503 (76%) 423 (64%)

NR 85%

NR 79%

15.1 59%

IHC2þ/FISH þ (n ¼ 34) 71 27 (79%) 26 (76%) 28 (82%) 21 (62%) 16.0 70%

Total (n ¼ 1,126)

69 726 (64%) 754 (67%) 951 (84%) 714 (63%) 19.8 67%

Mutant (n ¼ 21) 65 8 (38%) 9 (43%) 21 (100%) 14 (67%) NR 59%

*including 22 pts with HER2-positive tumors with EGFR mutations, 2 with both HER2- and ALK-positive tumors, and 2 had ALK-positive tumors with EGFRmutations. **1 had an IHC2þ/FISHþ tumor with mutation.

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx380 | 477

abstracts transcriptomic microarray. Raw intensity data was normalized by Robust Multi-Array Average and analyzed for differential expressed genes (DEGs) using Bioconductor R limma package. Results: A total 129 DEGs were identified whose gene ontology enrichment analysis revealed the top over-represented pathways as circulating immune complex (GO: 0042571, p¼2.2x1010), FCGR activation (REAC: 2029481, p¼5.4x1011), and regulation of B cell activation (GO: 0050864, p¼2.4x105), in which a number of upregulated genes encoding immunoglobulins, Src family kinases and interleukin (IL)-10 were noted. We integrated the immunoglobulin genes into a signature and calculated the enrichment score for each subject using Gene Set Variation Analysis (Sonja et al. BMC Bioinformatics 2013). Medium to high correlation of immunoglobulin signature with IL-10 (Pearson’s r: 0.64, p¼0.003), with FCGR2B (Pearson’s r: 0.43, p¼0.066) and FCGR2B with IL-10 (Pearson’s r: 0.50, p¼0.029) were determined. In addition, mild to medium correlation were identified between IL-10 and Src family kinases BLK (Pearson’s r: 0.48, p¼0.038), LCK (Pearson’s r: 0.40, p¼0.090) and YES1 (Pearson’s r: 0.25, p¼0.300). Conclusions: Transcriptome of BAL cells around advanced NSCLCs showed characteristic expression of immunoglobulin signature that may implicate the immunosuppressive property in tumor milieu. Legal entity responsible for the study: Chih-Hsi Scott Kuo MD Funding: Chang Gung Medical Foundation Disclosure: All authors have declared no conflicts of interest.

1341P

Assessing response to immunotherapy in patients with non-small cell lung cancer using circulating tumor DNA

S.B. Goldberg1, A. Narayan2, A.J. Kole2, R.H. Decker2, J. Teysir2, N.J. Carriero3, A. Lee2, R. Nemati2, S.K. Nath2, S.M. Mane4, Y. Deng5, N. Sukumar5, D. Zelterman5, D.J. Boffa6, K. Politi7, S.N. Gettinger1, L.D. Wilson2, R.S. Herbst1, A.A. Patel2 1 Medicine (Medical Oncology), Yale University School of Medicine, New Haven, CT, USA, 2 Therapeutic Radiology, Yale University School of Medicine, New Haven, CT, USA, 3 Scientiﬁc Computing Core, Simon’s Foundation, New York, NY, USA, 4Genetics, Yale University School of Medicine, New Haven, CT, USA, 5Biostatistics, Yale University School of Medicine, New Haven, CT, USA, 6Thoracic Surgery, Yale University School of Medicine, New Haven, CT, USA, 7Pathology, Yale University School of Medicine, New Haven, CT, USA Background: Evaluation of response to immune checkpoint inhibitors by serial imaging can be complicated by the possibility of pseudo-progression or delayed response, sometimes resulting in discontinuation of an effective therapy or delay of alternate treatment. Monitoring tumor cell death by measuring changes in circulating tumor DNA (ctDNA) levels in blood may permit early assessment of immunotherapy efficacy. Methods: We examined ctDNA levels in plasma samples from patients with metastatic non-small cell lung cancer (NSCLC) undergoing treatment with a PD-1 or PD-L1 inhibitor. CtDNA was quantified in plasma by determining the allele fraction of cancerassociated somatic mutations using a multi-gene next-generation sequencing assay. A ctDNA response was defined as more than 50% decrease in mutant allele fraction from baseline, with a second confirmatory measurement. Radiographic response assessment was performed using RECIST 1.1. Changes in ctDNA levels over time were correlated with imaging findings and with clinical outcomes. Results: Twenty-eight patients with metastatic NSCLC had ctDNA quantified in serial blood samples collected before and during treatment with a PD-1 axis inhibitor. Strong agreement was observed between ctDNA response and radiographic response (Cohen’s Kappa ¼ 0.753, P < 0.001). The median time to response was 24.5 days by ctDNA versus 72.5 days by imaging. Patients who had a ctDNA response remained on immunotherapy for a median of 205.5 days compared to a median of 69 days for those who did not have a ctDNA response (p < 0.001). Progression-free survival (PFS) and overall survival (OS) were significantly better for patients with a ctDNA response versus those without (hazard ratio [HR] for PFS, 0.29; 95% confidence interval [CI], 0.09-0.89; P ¼ 0.03 and HR for OS 0.17; 95% CI, 0.05-0.62; P ¼ 0.007). Conclusions: An early drop in ctDNA level enables assessment of response to immune checkpoint inhibitor therapies at a time when radiographic response may be uncertain for patients with metastatic NSCLC. Achievement of such a ctDNA response is predictive of a longer duration of therapeutic benefit as well as superior PFS and OS. Legal entity responsible for the study: Yale University Funding: The National Cancer Institute (RO1-CA197486-01A1), the Yale SPORE in Lung Cancer (P50-CA196530), the LUNGevity Foundation, the Honorable Tina Brozman Foundation, the Kalimeris Fund, and the Yale Cancer Center. Disclosure: S.B. Goldberg: Research funding from AstraZeneca. Membership on an AstraZeneca advisory board. A.J. Kole: Immediate family member is employed by and owns stock in Sarepta Therapeutics. R.H. Decker: Research support from Merck. Advisory board for Regeneron. K. Politi: Research support from Astra Zeneca, Roche, and Kolltan. Consultant for AstraZeneca, Novartis, and Merck. MSKCC for IP licensed to Molecular MD. S.N. Gettinger: Consultant for Bristol Myers Squibb and Alexion. L.D. Wilson: Owns stock in Medtronic, Alexion, Biogen, Bristol Myers Squibb, Celgene, Johnson and Johnson, Merck, United Health, and Vertex. R.S. Herbst: Research support from Genetech and Merck. Advisory board for AstraZeneca, Eli Lilly, Genentech/Roche, Merck and Pfizer. A.A. Patel: Research support from AstraZeneca. Advisory board for Novartis. Patent application covering circulating tumor DNA assay. All other authors have declared no conflicts of interest.

478 | NSCLC, metastatic

Annals of Oncology

1342P

Detection of epidermal growth factor receptor mutations in circulating cell-free DNA versus tumor biopsy

I-J. Oh1, H-W. Seo1, H-J. Cho1, C-K. Park1, J-H. Lim1, Y-C. Kim1, Y-D. Choi2 Department of Internal Medicine, Chonnam National University Medical School & Chonnam National University Hwasun Hospital, Hwasun, Republic of Korea, 2 Department of Pathology, Chonnam National University Medical School, Hwasun, Republic of Korea

1

Background: Epidermal growth factor receptor (EGFR) mutations are predictive marker of EGFR-tyrosine kinase inhibitor (TKI) therapy. We compared the sensitivity of EGFR mutation detection techniques between matched tumor tissue and peripheral blood sample in patients with lung adenocarcinoma. Methods: We collected the paired samples from plasma and paraffin-embedded tumor tissue in 295 patients before EGFR-TKIs therapy. DNA extraction was performed using the QIAamp MinElute virus spin kit and EGFR mutation analysis was done by two detection methods. One is the PNAClampTM (Clamp) which is the PNA-based PCR clamping that selectively amplifies only the mutated target DNA sequence as minor portion in mixture with the major wild type DNA sequences. The other is the TM PANAMutyper EGFR kit (Mutyper), which use PNA clamping-assisted fluorescence melting curve analysis to perform mutation detection and genotyping. Results: In tissue samples, the positive rates of EGFR sensitive mutation were not different between two methods (26.8% in Mutyper vs. 24.7% in Clamp). Plasma sensitivity was significantly higher in Mutyper than Clamp (70.9% vs. 30.1%, p < 0.001) with tissue as reference. The overall concordance and degree of agreement between two samples were better in Mutyper (91.2%, k ¼ 0.756, p < 0.001) than Clamp (81.7%, k ¼ 0.369, p < 0.001). The median progression-free survival (PFS) was significantly higher in EGFR sensitive group in tissue sample regardless of the two methods. In plasma sample, the median PFS of EGFR sensitive group was significantly longer than negative group only by Mutyper (11.8 vs. 3.7 months, p ¼ 0.020), not by Clamp (9.8 vs. 11.0 months, p ¼ 0.968). Conclusions: The plasma sensitivity and concordance of Mutyper were better than Clamp test. And Mutyper could better predict the PFS of EGFR mutation in plasma. So this technique is expected to be useful to detect EGFR mutation in circulating cell-free DNA sample. Legal entity responsible for the study: N/A Funding: This research was financially supported by grants from the Panagene Inc. Disclosure: All authors have declared no conflicts of interest.

1343P

Efficacy and safety of lorlatinib in patients (pts) with ALK1 non-small cell lung cancer (NSCLC) previously treated with 2nd-generation ALK TKIs

E. Felip Font1, A.T. Shaw2, B.J. Solomon3, T.M. Bauer4, S-H.I. Ou5, S. Gadgeel6, R.A. Soo7, T. Seto8, J.S. Clancy9, L.P. James10, A. Abbattista11, B. Besse12 1 Oncology, Vall d’Hebron Institute of Oncology, Barcelona, Spain, 2Cancer Center, Massachusetts General Hospital, Boston, MA, USA, 3Medical Oncology, Peter MacCallum Cancer Center, Melbourne, Australia, 4Medical Oncology, Sarah Cannon Research Institute, Nashville, TN, USA, 5Chao Family Comprehensive Cancer Center, University of California Irvine School of Medicine, Orange, CA, USA, 6Oncology, Karmanos Cancer Center, Detroit, MI, USA, 7Haematology-Oncology, National University Cancer Institute, Singapore, 8Department of Thoracic Oncology, National Kyushu Cancer Center, Fukuoka, Japan, 9Clinical, InVentiv Clinical, Princeton, NJ, USA, 10 Global Product Development, Pﬁzer Oncology, New York, NY, USA, 11Global Biometrics and Data Management, Pﬁzer Oncology, Milan, Italy, 12Department of Medical Oncology, Gustave Roussy Cancer Campus, Villejuif, France Background: Standard therapy for advanced anaplastic lymphoma kinase (ALK)þ NSCLC consists of first-line crizotinib, followed by 2nd-generation ALK tyrosine kinase inhibitors (TKIs) such as ceritinib, alectinib, or brigatinib. While pts can achieve some clinical benefit from these agents, most will subsequently develop resistance. Lorlatinib, a brain-penetrant next-generation ALK TKI, is active against most known resistance mutations that can develop during treatment with 1st- and 2nd-generation TKIs. Here, we explore the antitumor activity and safety of lorlatinib in pts with ALKþ NSCLC treated with 1 prior 2nd-generation TKI. Methods: In this ongoing phase 2 study (NCT01970865), pts with ALKþ or ROS1þ NSCLC, 6 asymptomatic untreated or treated central nervous system (CNS) metastases, were enrolled into 1 of 5 ALKþ cohorts dependent on the number of prior TKIs received, and 1 ROS1þ cohort with no restriction on the extent of previous therapy. Pts received lorlatinib 100 mg QD. The primary objective was overall and intracranial (IC) antitumor activity, measured as confirmed overall and IC response by independent central review (ICR). Results: At the data cutoff (15 Aug 2016), 65 pts with ALKþ NSCLC were treated with 1 prior 2nd-generation TKI (alectinib, ceritinib, brigatinib, or other). Of these pts, 41 had CNS metastases at baseline. The table shows the confirmed overall and IC responses (complete response þ partial response) by ICR. The most common treatmentrelated adverse events (TRAEs) of any grade in pts who received prior ALK TKIs were hypercholesterolemia, hypertriglyceridemia, and peripheral neuropathy. Hyperlipidemia was successfully managed by appropriate medical treatment.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology 1344P

Table: 1343P Type of 2nd-Generation TKI

Overall N

Alectinib Ceritinib Brigatinib Other

a

33 29 6 5b

Intracranial a

Confirmed N Responses, n (%)

Confirmed Responses, n (%)

9 (27) 9 (31) 1 (17) 2 (40)

8 (50) 11 (48) 0 (0) 2 (50)

16 23 3 4

a

Patients in each group of prior TKI; a patient could have received more than one type of prior TKI. b Other TKIs included entrectinib (n ¼ 3) and ensartinib (n ¼ 2).

Conclusions: Lorlatinib has shown clinical activity in pts with ALKþ NSCLC who had received 1 prior 2nd-generation TKI. Clinical trial identification: NCT01970865 Legal entity responsible for the study: Pfizer Funding: Pfizer Disclosure: E. Felip Font: Advisory boards: Eli Lilly, Pfizer, Roche, MSD, Boehringer Ingelheim. Speaker’s bureau/lectures fees: AstraZeneca, BMS, Novartis. A.T. Shaw: Advisory board or board of directors: Blueprint medicines, KSQ therapeutics (scientific advisory board). Consulting/Honoria Pfizer, Novartis, Ariad, Genentech/Roche, Ignyta, Daiichi-sankyo, Taiho, Loxo, Blueprint medicines, EMD Serono, Foundation Medicine. B.J. Solomon: Advisory Boards: Pfizer, Novartis, Roche-Genentech, AstraZeneca, Merck, Bristol Myers Squibb. S-H.I. Ou: Membership of an advisory board or board of directors: Genentech/Roche, Ariad, Pfizer, Novartis, Astra Zeneca. S. Gadgeel: Membership of an advisory board or board of directors: Genentech/Roche, Ariad, Pfizer, Novartis. R.A. Soo: Membership of an advisory board or board of directors: AstraZeneca, Boehringer Ingelheim, BMS, Lilly, Merck, Novartis, Pfizer, Roche, Taiho Corporate sponsored research: AstraZeneca. T. Seto: Research funding from Eisai Co., Ltd., MSD, K.K., Nippon Boehringer Ingelheim Co., Ltd, Pfizer Japan Inc., AstraZeneca K.K., Astellas Pharma Inc., Chugai Pharmaceutical Co., Ltd., Daiichi Sankyo Co. Ltd., Eli Lilly Japan K.K., Merck Serono Co., Ltd., Novartis Pharma K.K., Pfizer Japan Inc. Honoraria from AstraZeneca K.K., Bristol-Myers Squibb, Chugai Pharmaceutical Co., Ltd., Eli Lilly Japan K.K., Kissei Pharmaceutical Co. Ltd., Kyowa Hakko Kirin Co., Ltd., MSD K.K., Nippon Boehringer Ingelheim Co., Ltd, Nippon Kayaku Co., Ltd., Ono Pharmaceutical Co. Ltd., Pfizer Japan Inc. Roche Singapore Pte Ltd, Taiho Pharmaceutical Co., Ltd., YakultHonsha Co., Ltd. J.S. Clancy: Employed by inVentiv Clinical but under contract position with Pfizer. L.P. James, A. Abbattista: Employee and stock owner of Pfizer. B. Besse: Corporate sponsored research: Pfizer. All other authors have declared no conflicts of interest.

Brigatinib (BRG) in anaplastic lymphoma kinase (ALK)-positive nonsmall cell lung cancer (NSCLC): Long-term efficacy and safety results from a phase 1/2 trial

L.A. Bazhenova1, S.N. Gettinger2, C.J. Langer3, R. Salgia4, K.A. Gold5, R. Rosell6, A.T. Shaw7, G.J. Weiss8, J. Haney9, V.M. Rivera10, D. Kerstein11, R. Camidge12 1 Department of Medicine, Moores Cancer Center, University of California San Diego, La Jolla, CA, USA, 2Yale Cancer Center, Yale School of Medicine, New Haven, CT, USA, 3 Abramson Cancer Center, University of Pennsylvania, Philadelphia, PA, USA, 4 Department of Medical Oncology and Therapeutics Research, City of Hope, Duarte, CA, USA, 5Moores Cancer Center, University of California San Diego, La Jolla, CA, USA, 6 Germans Trias i Pujol Health Sciences Institute and Hospital, Catalan Institute of Oncology (ICO Badalona), Badalona, Spain, 7Massachusetts General Hospital Cancer Center, Massachusetts General Hospital, Boston, MA, USA, 8Clinical Research, Cancer Treatment Centers of America, Goodyear, AZ, USA, 9Biomedical Data Sciences and Information, ARIAD Pharmaceuticals Inc., Cambridge, MA, USA, 10Preclinical and Translational Research, ARIAD Pharmaceuticals Inc., Cambridge, MA, USA, 11Clinical Research, ARIAD Pharmaceuticals Inc., Cambridge, MA, USA, 12Cancer Center, University of Colorado, Aurora, CO, USA Background: The next-generation ALK inhibitor BRG has shown activity in ALKþ NSCLC patients (pts) in clinical trials. Methods: Pts with advanced malignancies (N ¼ 137; including 79 pts with ALKþ NSCLC) received oral BRG (30–300 mg/d) in a phase 1/2, open-label, multicenter trial (NCT01449461). We report activity by RECIST v1.1 in ALKþ NSCLC pts and safety in all pts, with long-term follow-up (>31 months since last pt was enrolled). Results: Among 79 ALKþ NSCLC pts, median age was 54 years; 90% (71/79) had received prior crizotinib (CRZ). As of 21 Feb 2017, 32% of ALKþ NSCLC pts (25/79) and 25% (7/ 28) of those receiving 180 mg qd with a 7-day lead-in at 90 mg (a regimen evaluated in the phase 2 portion of the trial) continued to receive BRG. Median treatment duration was 20.0 months (1 day to 56.1 months). Confirmed objective response rate (ORR) was 63% (45/71) in pts with prior CRZ and 100% (8/8) in CRZ-naive pts. Additional efficacy data are shown in the table. At 180 mg qd (with lead-in), confirmed ORR was 76% (95% CI, 55%–91%) and median progression-free survival (PFS) was 16.3 months (95% CI, 9.2– 28.1) in pts with prior CRZ. Treatment-emergent adverse events (AEs) in 30% of all 137 pts, mostly grade 1/2, were nausea (55%), fatigue (45%), diarrhea (42%), headache (36%), and cough (34%). Grade 3 treatment-emergent AEs in 5% of pts were increased lipase (12%), pneumonia (7%), dyspnea (6%), and hypertension (6%). Eleven percent of pts (15/137) discontinued BRG due to an AE. Conclusions: BRG shows major antitumor activity in ALKþ NSCLC pts with an acceptable safety profile in this long-term follow-up. PFS of > 16 months in pts receiving 180 mg qd with a 7-day lead-in at 90 mg is among the longest reported in CRZ-resistant ALKþ NSCLC. This dosing regimen is being investigated in a randomized phase 3 trial of BRG vs CRZ in ALK inhibitor–naive pts with advanced ALKþ NSCLC (ALTA-1L; currently recruiting pts). Clinical trial identification: NCT01449461. First received by ClinicalTrials.gov: September 30, 2011 Legal entity responsible for the study: ARIAD Pharmaceuticals, a wholly owned subsidiary of Takeda Pharmaceutical Company Limited. Funding: ARIAD Pharmaceuticals, a wholly owned subsidiary of Takeda Pharmaceutical Company Limited.

Table: 1344P Efficacy in ALKþ NSCLC Pts With Prior CRZ

CRZ-Naive a,b

90 mg qd n 5 13

90 mg ﬁ 180 mg qd

45 (63) 51–75 14.5 9.0–26.1

7 (54) 25–81 11.1 3.8–16.7

19 (76) 55–91 14.9 7.9–33.3

8 (100) 63–100 32.4 5.6–NR

13.2 9.2–16.7 53 41–65 30.1 21.4–NR 77 65–85 61 48–71

11.9 3.5–21.2 50 21–74 21.2 9.9–47.6 69 37–87 46 19–70

16.3 9.2–28.1 62 40–78 29.5 21.4–NR 84 63–94 64 42–79

34.2 7.4–NR 75 32–93 NR NR–NR 100 100–100 100 100–100

All n 5 71 cORR, n (%) 95% CI Median duration of response in conﬁrmed responders, months 95% CI Median PFS, months 95% CI Probability of PFS at 1 year, % 95% CI Median OS, months 95% CI Probability of OS at 1 year, % 95% CI Probability of OS at 2 years, % 95% CI

a

n 5 25

All n 5 8

Time-to-event data reﬂect Kaplan-Meier estimates ALKþ NSCLC, anaplastic lymphoma kinase–positive non–small cell lung cancer; BRG, brigatinib; CI, conﬁdence interval; cORR, conﬁrmed objective response rate; CRZ, crizotinib; NR, not reached; OS, overall survival; PFS, progression-free survival; pts, patients a BRG regimens used in the pivotal phase 2 trial b 180 mg qd with 7-day lead-in at 90 mg.

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx380 | 479

abstracts Disclosure: L.A. Bazhenova: Stock and other ownership interests (Epic Sciences), consulting or advisory role (AbbVie, ARIAD, AstraZeneca, Genoptix, Heat Biologics, Novartis, Pfizer, Trovagene), speakers bureau (AstraZeneca, Novartis, Roche/ Genentech), research funding (BeyondSpring). S.N. Gettinger: Consulting or advisory role (ARIAD, BMS, Janssen), research funding (ARIAD, AstraZeneca/MedImmune, BMS, Boehringer Ingelheim, Incyte, Pfizer, Roche/Genentech). C.J. Langer: Honoraria (BMS, Eli Lilly/ImClone, Roche/Genentech), consulting or advisory role (Abbott, ARIAD, Bayer/Onyx, BMS, Cancer Support Community, Celgene, Clarient, Clovis Oncology, Eli Lilly/ImClone, Merck, Roche/Genentech), research funding (Advantagene, ARIAD, Celgene, Clovis Oncology, GSK, Inovio, Merck, Roche/ Genentech, Stem CentRx). K.A. Gold: Honoraria (Roche/Genentech), consulting or advisory role (ARIAD), travel, accommodations, expenses (AstraZeneca), research funding (ARIAD, Astellas, AstraZeneca, BMS, Pharmacyclics, Roche/Genentech). A.T. Shaw: Honoraria (Novartis, Pfizer, Roche/Genentech), consulting or advisory role (ARIAD, Blueprint Medicines, Daiichi Sankyo, EMD Serono, Ignyta, Novartis, Pfizer, Roche/Genentech, Taiho), research funding (Novartis, Pfizer, Roche/Genentech). G.J. Weiss: Employment (Cancer Treatment Centers of America), consulting or advisory role (Blend Therapeutics, Paradigm, Pharmatech), speakers bureau (Amgen, Celgene, Medscape, Pfizer, Quintiles), travel, accommodations, expenses (Cambridge Healthtech Institute, Pharmatech), patents, royalties, other intellectual property (US Patent #8,911,940 issued 2014). J. Haney: Employment, stock and other ownership interests (ARIAD). V.M. Rivera, D. Kerstein: Employment, stock and other ownership interests (ARIAD). R. Camidge: Honoraria (ARIAD), research funding (ARIAD). All other authors have declared no conflicts of interest.

1345P

Intracranial efficacy of brigatinib (BRG) in patients (Pts) With crizotinib (CRZ)-refractory anaplastic lymphoma kinase (ALK)-positive non-small cell lung cancer (NSCLC) and baseline CNS metastases

S-H.I. Ou1, M. Tiseo2, R. Camidge3, M-J. Ahn4, R.M. Huber5, M.J. Hochmair6, S-W. Kim7, H.L. West8, K.L. Reckamp9, J.R. Molina10, G. Liu11, A. Delmonte12, S. Viteri13, A. Bearz14, Y. Summers15, W. Reichmann16, D. Kerstein17, S.N. Gettinger18, D-W. Kim19 1 Chao Family Comprehensive Cancer Center, University of California Irvine School of Medicine, Orange, CA, USA, 2Medical Oncology Unit, University Hospital of Parma, Parma, Italy, 3Cancer Center, University of Colorado, Aurora, CO, USA, 4Department of Hematology and Oncology, Samsung Medical Center, Seoul, Republic of Korea, 5 Division of Respiratory Medicine and Thoracic Oncology, University Hospital of Munich, Munich, Germany, 6Respiratory Oncology Unit, Otto Wagner Hospital, Vienna, Austria, 7Department of Oncology, Asan Medical Center, Seoul, Republic of Korea, 8 Thoracic Oncology Program, Swedish Cancer Institute, Seattle, WA, USA, 9Clinical Research Operations, City of Hope, Duarte, CA, USA, 10Department of Oncology, Mayo Clinic, Rochester, MN, USA, 11Department of Medicine, Princess Margaret Cancer Centre, Toronto, ON, Canada, 12Department of Medical Oncology, Istituto Scientiﬁco Romagnolo per lo Studio e la Cura dei Tumori (IRST) IRCCS, Meldola, Italy, 13Medical Oncology, Quir onSalud Dexeus University Hospital, Instituto Oncol ogico Dr Rosell (IOR), Barcelona, Spain, 14Division of Medical Oncology, National Cancer Institute, Aviano, Italy, 15Department of Medical Oncology, The Christie NHS Foundation Trust, Manchester, UK, 16Biomedical Data Sciences and Information, ARIAD Pharmaceuticals Inc., Cambridge, MA, USA, 17Clinical Research, ARIAD Pharmaceuticals Inc., Cambridge, MA, USA, 18Yale Cancer Center, Yale School of Medicine, New Haven, CT, USA, 19 Department of Internal Medicine, Seoul National University Hospital, Seoul, Republic of Korea Background: The CNS is often a site of first disease progression in CRZ-treated ALKþ NSCLC pts; we report BRG efficacy and safety in pts with CRZ-refractory advanced ALKþ NSCLC in the phase 2 ALTA trial who had baseline CNS metastases. Methods: ALTA (NCT02094573) permitted baseline CNS disease (including pts with prior whole brain radiotherapy/stereotactic radiosurgery and asymptomatic untreated pts). 222 pts stratified by presence of brain metastases and best response to prior CRZ were randomized 1:1 to receive BRG 90 mg qd (arm A, n ¼ 112) or 180 mg qd with a 7day lead-in at 90 mg (arm B, n ¼ 110). This analysis included an exploratory competing risks analysis to estimate cumulative incidence of CNS progression vs non-CNS progression vs death in pts with baseline brain metastases. Results: 80 (71%)/73 (66%) pts in A/B had baseline brain metastases per independent review committee (median age, 49/55 years; 76%/74% had received chemotherapy). As of 21 Feb 2017, median follow-up for pts with brain metastases was 17.7 months; 34%/ 40% continued to receive BRG in A/B. Table shows intracranial efficacy. 5 pts with measurable baseline brain metastases in A had progression in the brain (20% growth in target lesions or new lesions) while receiving BRG 90 mg qd and escalated to 180 mg qd with 1 additional scan; all 5 had a reduction in measurable lesions after escalation (14% to 100%). While pts without baseline brain metastases did not have routine brain MRI scans, 3/32 and 1/36 pts without baseline brain metastases per investigators in A and B, respectively, had a new brain lesion identified by MRI. Conclusions: In this update of ALTA, BRG continued to show robust intracranial efficacy in ALKþ NSCLC pts with baseline brain metastases, particularly at 180 mg (with lead-in), with a higher intracranial response rate and a numerically lower incidence of disease progression in the CNS and outside the CNS, compared to 90 mg.

480 | NSCLC, metastatic

Annals of Oncology

Table: 1345P Intracranial Efficacy of BRG in Pts With CRZ-Refractory ALKþ NSCLC and Baseline Brain Metastases (per IRC)

Arm A 90 mg qd

Arm B 90 mg ﬁ 180 mg qda

Conﬁrmed iORR (pts with measur13/26 (50) 12/18 (67) able brain metastases), n/N (%) Conﬁrmed iORR (pts with measur9/19 (47) 11/15 (73) able, activeb brain metastases), n/ N (%) Median duration of intracranial reNR (3.7–NR) 16.6 (3.7–16.6) sponsec (pts with measurable n ¼ 13 n ¼ 12 brain metastases), months (95% CI) Median iPFSc (pts with any baseline 12.8 (9.0–18.3) 18.4 (12.6–NR) brain metastases), months (95% n ¼ 80 n ¼ 73 CI) Competing risks analysis n ¼ 80 n ¼ 73 CIR of ﬁrst disease progression in CNS, % (95% CI) By 6 months 20 (12–29) 15 (8–25) By 12 months 33 (23–44) 27 (17–37) By 18 months 41 (30–52) 34 (23–45) CIR of ﬁrst disease progression at non-CNS site, % (95% CI) By 6 months 14 (7–23) 11 (5–20) By 12 months 21 (13–31) 20 (11–30) By 18 months 23 (15–34) 21 (13–32) CIR of death prior to all disease progression (in CNS or at non-CNS site), % (95% CI) By 6 months 5 (2–12) 3 (1–9) By 12 months 9 (4–17) 6 (2–13) By 18 months 9 (4–17) 7 (3–15) Pts with baseline brain metastases are shown; in these pts, CNS disease was tracked by MRI every 8 weeks Last scan date: 28 Feb 2017 ALKþ NSCLC, anaplastic lymphoma kinase–positive non–small cell lung cancer; BRG, brigatinib; CI, conﬁdence interval; CIR, cumulative incidence rate; CNS, central nervous system; CRZ, crizotinib; iORR, intracranial objective response rate; iPFS, intracranial progression-free survival; IRC, independent review committee; MRI, magnetic resonance imaging; NR, not reached; pts, patients a 180 mg qd with 7-day lead-in at 90 mg b Untreated, or treated and progressed c From Kaplan-Meier analysis.

Clinical trial identification: NCT02094573 First received by ClinicalTrials.gov: March 18, 2014 Legal entity responsible for the study: ARIAD Pharmaceuticals, Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company Limited Funding: ARIAD Pharmaceuticals, Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company Limited. Disclosure: S-H.I. Ou: Honoraria (ARIAD, AstraZeneca, Novartis, Pfizer, Roche), consulting or advisory role (ARIAD, AstraZeneca, Novartis, Pfizer, Roche), speakers bureau (AstraZeneca, Roche), research funding (ARIAD, AstraZeneca, Clovis Oncology, Daiichi Sankyo, Ignyta, Novartis, Pfizer, Roche). M. Tiseo: Consulting or advisory role (AstraZeneca, BMS, Boehringer Ingelheim, Eli Lilly, Novartis, Otsuka, Pierre Fabre), research funding (ARIAD). R. Camidge: Honoraria (ARIAD), research funding (ARIAD). M-J. Ahn: Honoraria (AstraZeneca, BMS, Boehringer Ingelheim, MSD, Novartis), consulting or advisory role (AstraZeneca, BMS, Boehringer Ingelheim, MSD, Novartis). R.M. Huber: Honoraria (ARIAD, AstraZeneca, BMS, Boehringer Ingelheim, Pfizer, Roche), consulting or advisory role (BMS, Boehringer Ingelheim, Celgene, Clovis Oncology, Eli Lilly, Novartis, Roche), research funding (Pierre Fabre). H.L. West: Consulting or advisory role (ARIAD, AstraZeneca, BMS, Boehringer Ingelheim, Celgene, Guardant Health, Merck, Novartis, Roche/Genentech), speakers bureau (ARIAD, Roche/Genentech). K.L. Reckamp: Consulting or advisory role (Amgen, ARIAD, Astellas, Boehringer Ingelheim, Euclises, Nektar), research funding (Adaptimmune, ARIAD, BMS, Clovis Oncology, Eisai, Gilead Sciences, GSK, Novartis, Pfizer, Roche/Genentech, Xcovery). G. Liu: Honoraria (Novartis, Pfizer), consulting or

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology advisory role (AstraZeneca/MedImmune, Novartis, Pfizer, Roche Canada, Takeda), research funding (Roche). A. Delmonte: Consulting or advisory role (BMS, Boehringer Ingelheim, Novartis). S. Viteri: Consulting or advisory role (Boehringer Ingelheim, Clovis Oncology, Idea Pharma, Novartis, Promega Biotech Ibe´rica, Roche, Targovax), research funding (AbbVie, ARIAD, Astex, AstraZeneca/MedImmune, Boehringer Ingelheim, Clovis Oncology, CytRx, Daiichi Sankyo, GSK, Hanmi, Incyte, Merck KGaA, Novartis, Pfizer, Puma, Roche, Servier, Vaxon). A. Bearz: Consulting or advisory role (Boehringer Ingelheim, Eli Lilly, Novartis, Roche). Y. Summers: Consulting or advisory role (AstraZeneca, BMS, Boehringer Ingelheim, Clovis Oncology, Eli Lilly, MSD, Pfizer, Tesaro). W. Reichmann, D. Kerstein: Employment, stock and other ownership interests (ARIAD). S.N. Gettinger: Consulting or advisory role (ARIAD, BMS, Janssen), research funding (ARIAD, AstraZeneca/MedImmune, BMS, Boehringer Ingelheim, Incyte, Pfizer, Roche/Genentech). All other authors have declared no conflicts of interest.

1346P

CNS efficacy results from the phase III ALUR study of alectinib vs chemotherapy in previously treated ALK1 NSCLC

J. de Castro1, S. Novello2, J. Mazieres3, I-J. Oh4, M.R. Migliorino5, A. Helland6, R. Dziadziuszko7, F. Griesinger8, F. de Marinis9, A. Zeaiter10, A. Cardona10, B. Balas10, H. Johannsdottir10, M. Chlistalla10, V. Smoljanovic10, J. Wolf11 1 Department of Medical Oncology, University Hospital, La Paz, Spain, 2Department of Oncology, University of Turin, Turin, Italy, 3Thoracic Oncology Department, Toulouse University Hospital, Toulouse, France, 4Department of Internal Medicine, Chonnam National University Hwasun Hospital, Jeonnam, Republic of Korea, 5Oncology, A.O. San Camillo Forlanini, Rome, Italy, 6Department of Cancer Genetics and Department of Oncology, Institute for Cancer Research and University Hospital, Radiumhospitalet, Oslo, Norway and Oslo University Hospital, Radiumhospitalet, Oslo, Norway, 7 Department of Oncology and Radiotherapy, Medical University of Gdansk, Gdansk, Poland, 8Department of Hematology and Oncology, Pius Hospital, OIdenburg, Germany, 9Division of Thoracic Oncology, European Institute of Oncology, Milan, Italy, 10 Oncology, F. Hoffmann-La Roche Ltd., Basel, Switzerland, 11Center for Intergrated Oncology, University Hospital Cologne, Cologne, Germany Background: Alectinib has shown central nervous system (CNS) activity in phase II trials of previously treated ALKþ NSCLC. We report CNS efficacy data from the phase III ALUR study (NCT02604342) of alectinib vs standard relapse chemotherapy (CT) in pts with ALKþ NSCLC who previously failed platinum-based doublet CT and crizotinib. Methods: Pts aged 18 years with ALKþ NSCLC, previously treated with CT and crizotinib were randomised 2:1 to alectinib (600mg twice daily) or CT (pemetrexed 500mg/ m2 every three weeks [q3w] or docetaxel 75mg/m2 q3w) until disease progression (PD), death or withdrawal. After PD, pts could crossover from CT to alectinib. Primary endpoint was progression-free survival (PFS) by investigator assessment in the intent-totreat (ITT) population. A key secondary endpoint was CNS overall response rate (CORR) by Independent Review Committee (IRC) in pts with measurable CNS mets at baseline (BL) (mC-ITT). Other CNS outcomes were CORR in pts with measurable and non-measurable CNS mets (C-ITT); 6-month cumulative incidence rate in the ITT, and C-ITT; CNS duration of response (CDOR) and disease control rate (CDCR); and safety. Results: In total, 107 pts were randomised (alectinib n ¼ 72, CT n ¼ 35; ITT) of whom 76 had BL CNS disease (alectinib n ¼ 50, CT n ¼ 26; C-ITT); 40 had measurable CNS mets (alectinib n ¼ 24; CT n ¼ 16; mC-ITT) and 36 had non-measurable CNS mets (alectinib n ¼ 26, CT n ¼ 10; nC-ITT). CNS efficacy endpoints are shown in the Table. The 6-month cumulative incidence rate of CNS PD was 11% (alectinib) vs 48% (CT) in the ITT, 15% vs 52% in the C-ITT and 0% vs 39% in pts without BL CNS disease. Safety and tolerability profile compared favourably for alectinib vs CT. Conclusions: CNS-related outcomes were significantly improved with alectinib vs chemotherapy in previously treated ALKþ NSCLC. Alectinib reduces CNS PD and prevents the development of new CNS mets. Clinical trial identification: NCT02604342 Legal entity responsible for the study: F. Hoffmann-La Roche Ltd. Funding: F. Hoffmann-La Roche Disclosure: J. de Castro: Membership on an advisory board: Astra-Zeneca, Boehringer, MSD, Novartis, Pfizer, Roche. S. Novello: Speaker Bureau: Eli Lilly, BMS, MSD, Astra

Zeneca, Roche. J. Mazieres: Honoraria: Novartis, Roche, Pfizer, BMS, MSD. M.R. Migliorino: Honoraria and Advisory board: Bristol Mayer Squibb, Boehringer Ingelheim, MSD, Astra Zeneca. R. Dziadziuszko: Travel accommodation: Roche. Honoraria: Novartis, Tesaro, Clovis. Honoraria and speaker’s bureau: Pfizer, Honoraria and consulting role: Boehringer-Ingelheim, Astra-Zeneca, Ignyta. F. Griesinger: Advisory Board and corporate sponsored research: Roche, Astra Zeneca, Lilly, BMS, MSD, Celgene, Boehringer, Pfizer, Novartis. F. de Marinis: Advisor for BMS, Pfizer, Roche, Astra Zeneca, BI. A. Zeaiter, A. Cardona, B. Balas: Employment with F. Hoffmann-La Roche. H. Johannsdottir: Employee of Hoffmann-La Roche. M. Chlistalla: Employee of F. Hoffmann-La Roche Ltd. V. Smoljanovic: Employee of F. Hoffmann-La Roche Ltd. with stock ownership. J. Wolf: Advisory boards and lecture fees from AstraZeneca, BMS, Boehringer-Ingelheim, Clovis, Lilly MSD, Novartis, Pfizer, Roche. Research support from BMS, MSD, Novartis, Pfizer, Roche. All other authors have declared no conflicts of interest.

1347P

Treatment beyond disease progression: ALK inhibitors in ALKrearranged advanced NSCLC

N.B. Leighl1, J. Kuo1, A. Pavel1, M. Prescilla1, F.A. Shepherd2, G. Liu3, P. Bradbury4, M. Moskovits1 1 Department of Medical Oncology and Hematology, Princess Margaret Cancer Centre, University of Toronto, Toronto, ON, Canada, 2Department of Medical Oncology and Hematology, Princess Margaret Cancer Centre, Toronto, ON, Canada, 3Department of Medical Oncology and Hematology, Princess Margaret Hospital, Toronto, ON, Canada, 4 Medical Oncology, Princess Margaret Cancer Centre, Toronto, ON, Canada Background: Treatment of patients with ALK-EML4 gene rearrangement with ALK inhibitors (ALK-I) yields high response rate (RR) and prolonged progression free survival (PFS). Defining progressive disease (PD) using RECIST has been challenging in the era of targeted and immunotherapy, as some patients are perceived to be deriving ongoing benefit despite PD by RECIST. We explore the impact of treatment beyond progression with ALK-I on patient symptoms and treatment duration. Methods: Patients with advanced ALK-rearranged lung cancer treated at the Princess Margaret Cancer Centre between 2009 and 2017 were identified. Treatment duration was obtained from medical records, assessment of PFS by RECIST 1.1 from patient imaging, and patient self-reported symptoms and performance status (Edmonton Symptom Assessment Scale (ESAS), ECOG PS) were collected prospectively. Results: 66 patients were identified with advanced ALK-rearranged lung cancer. The median age at diagnosis was 61 years, 49% were male, 78% presented with stage 4 disease, 47 received ALK-I therapy (median 2 lines, range 1-4). Over half (26/47, 55%) continued ALK-I treatment beyond RECIST PD. PD occurred most commonly in brain (15/47), lung and/or pleura (11/47); 17/47 received local therapy (predominantly radiation) and continued ALK-I. Data on time to RECIST PD and treatment failure are shown below. Only 34/47 patients had symptom data available at baseline, 22 with severe symptoms; 70% improved with initial ALK-I treatment. At the time of RECIST PD, most of those continuing ALK-I beyond progression had not experienced deterioration of symptoms.

Table: 1347P ALK-I Treatment Patterns Initial ALK-I (N ¼ 47)

Second ALK-I (N ¼ 26)

Median Time to RECIST PD

10.1 months 4.8 months (range 0.3-– NR) (range 0.5 – NR) Pts continuing ALK-I beyond PD 17/47 (36%) 8/26 (30%) Median duration of treatment 5.0 (range 0.6 – NR) 3.9 (range 1.3-21.5) beyond progression NR- not reached.

Table: 1346P CNS efficacy endpoints C-ITT N ¼ 76 Alectinib N ¼ 50 CORR, % (95% CI) Difference (95% CI) P value CDOR, months (95% CI) CDCR, %

36 0 36% (13%–57%) p < 0.001 NE (6.2–NE) 80

mC-ITT N ¼ 40

Chemotherapy N ¼ 26

0 26.9

Alectinib N ¼ 24

Chemotherapy N ¼ 16

54.2 0 54% (23%–78%) p < 0.001 NE (3.6–NE) 0 79.2 31.3

NE, not evaluable

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx380 | 481

abstracts

Annals of Oncology

Conclusions: Treatment beyond disease progression for patients with advanced lung cancer harboring ALK rearrangement is common and is often associated with maintenance of symptom burden. Legal entity responsible for the study: UHN, University of Toronto Funding: None Disclosure: All authors have declared no conflicts of interest.

1348P

Overall survival (OS) in patients (pts) with EGFR T790M-positive advanced non-small cell lung cancer (NSCLC) treated with osimertinib: Results from two phase II studies

T. Mitsudomi1, M-J. Ahn2, L. Bazhenova3, F. Blackhall4, T. Hida5, M. Majem Tarruella6, S.L. Vowler7, G. Laus8, P.A. J€anne9, J.C-H. Yang10 1 Thoracic Surgery, Kindai University School of Medicine, Osaka, Japan, 2Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea, 3UC San Diego Health, Moores Cancer Center, San Diego, CA, USA, 4The Christie NHS Foundation Trust and Division of Molecular and Clinical Cancer Services, University of Manchester, Manchester, UK, 5Department of Thoracic Oncology, Aichi Cancer Center, Nagoya, Japan, 6 Medical Oncology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, 7Biostatistics and Informatics, AstraZeneca, Cambridge, UK, 8Global Medicines Development, AstraZeneca, Cambridge, UK, 9Lowe Center for Thoracic Oncology, Dana Farber Cancer Institute, Boston, MA, USA, 10Cancer Research Center, National Taiwan University, Taipei, Taiwan Background: Osimertinib is an oral, potent, CNS active, irreversible EGFR-TKI approved to treat pts with T790M-positive advanced NSCLC. Here we present updated data from a pre-planned pooled analysis of Phase II studies: AURA extension (NCT01802632), AURA2 (NCT02094261). V Methods: Pts with centrally confirmed T790M-positive (by cobas EGFR Mutation Test) advanced NSCLC, who had progressed following EGFR-TKI treatment, received osimertinib 80 mg once daily. Other inclusion criteria were measurable disease, WHO performance status 0/1 and acceptable organ function. Pts with stable CNS metastases were eligible. The primary efficacy endpoint was objective response rate (ORR) by RECIST 1.1 per blinded independent central review (BICR). Secondary endpoints included duration of response (DoR), progression-free survival (PFS), OS and safety. Results: As of the 1 Nov 2016 data cut-off (DCO) 411 pts had received osimertinib (129 pts as second-line and 282 pts as third-line); median treatment exposure was 16.4 months (mo; range 0–29.7 mo). Pooled results (BICR) were: ORR (evaluable for response set; EFR), 66% (95% confidence interval [CI] 61, 70); median DoR (EFR), 12.3 mo (95% CI 11.1, 13.8); median PFS (full analysis set), 9.9 mo (95% CI 9.5, 12.3). At DCO, 188 pts (46%) had died, 191 pts (47%) remained in survival follow up, and 32 pts (8%) had discontinued before death. Pooled median OS was 26.8 mo (95% CI 24.2, not calculable [NC]); median OS in the second-line and third-line cohorts (95% CI) was 25.8 mo (24.0, NC) and NC (22.1, NC), respectively. The 12 and 24 mo survival rates were 80% and 56%, respectively. The most common (investigator assessed) possibly causally-related adverse events (AEs) were rash (grouped term 42%, [grade 3, 1%]) and diarrhoea (39% [1 yr. Response assessments were reported for 16 pts; disease control rate was 69% (11/ 16 pts) and objective response rate was 19% (3/16 pts). No new or unexpected safety findings were observed. Conclusions: This analysis of a heavily pre-treated pt population with HER2mþ NSCLC from the afatinib NPU program showed a promising 32% of pts with TTF >1 yr, durable disease control and a manageable safety profile. With the limitation of retrospective analysis, and also considering that different HER2 mutations can display different treatment sensitivities, these findings suggest that the evaluation of afatinib in earlier treatment lines in HER2mþ NSCLC pts may be warranted. Legal entity responsible for the study: Boehringer Ingelheim Funding: Boehringer Ingelheim Disclosure: J-Y. Shih: Advisory board: AstraZeneca, Roche, Boehringer Ingelheim, MSD Oncology, Chugai Pharma. Honoraria: AstraZeneca, Roche, Eli Lilly, Boehringer Ingelheim, Pfizer, MSD Oncology, Bristol-Myers Squibb, Novartis. Other substantive relationship: Travel, Accommodations, Expenses: Roche, Boehringer Ingelheim, Bristol-Myers Squibb. V.W-Y. Liao: Honoraria: AstraZeneca, Roche,Pfizer, Boehringer Ingelheim, Norvatis, Eli Lilly, Merck Sharp & Dohme, Sanofi. V. Spataro: Advisory board: Roche, Novartis, Bristol-Myers, MSD. R. Lorence: Employee and consultant for Boehringer Ingelheim Pharmaceuticals, Inc. A. Cseh: Employee of Boehringer Ingelheim. All other authors have declared no conflict of interest.

1356P

Octogenarians with EGFR-mutated non-small cell lung Cancer (NSCLC) treated by Tyrosine Kinase Inhibitor (TKI): A multicentric real world study assessing tolerance and efficacy. OCTOMUT study GFPC 07-15

R. Corre1, R. Gervais2, L. Tassy3, F. Guisier4, R. Lamy5, G. Fraboulet6, L. Greillier7, H. Doubre8, C. chouaid9, J.B. Auliac10 1 Pneumology, CHU Pontchaillou, Rennes, France, 2Pneumo-oncology, Centre Francois Baclesse, Caen, France, 3Oncologie, Institut paoli calmet, Marseille, France, 4Pneumology, CHU Hoˆpitaux de Rouen-Charles Nicolle, Rouen, France, 5Oncology, CH Sud Bretagne, Lorient, France, 6Pneumology, Hopital Rene´ Dubos, Pontoise, France, 7Pneumo-oncology, Hopital St. Marguerite Assistance Publique Hopitaux de Marseille, Marseille, France, 8 Pneumology, Hopital Foch, Suresnes, France, 9Pneumology, CH Intercommunal de Cre´teil, Cre´teil, France, 10Pneumology, CH Franc¸ois Quesnay, Mantes La Jolie, France Background: Tyrosine kinase inhibitors (TKI) are the standard of treatment in first line for advanced EGFR-mutated NSCLC. Few data exist about their tolerance and efficacy in octogenarians particularly in Caucasian population. The purpose of this multicentric real world study was to assess tolerance and efficacy of EGFR-TKI in this population. Methods: We retrospectively identified patients aged 80 years or older with EGFRmutated NSCLC treated by EGFR-TKI between 01/01/2011 and 31/03/2015 whatever the line of treatment. Patients were described according to their clinical characteristics, management and outcomes (progression-free survival (PFS) and overall survival (OS). Results: The 20 french participating centers included 114 patients: 77% were women, the median age was 83,9 63,9 years; 98,2% were caucasians and 84,6% were home life (45% had home help), 71% took more than 5 drugs/d. Respectively 64%,17.5% and 8,5% of patients had a performance status of 0-1/2/3 at diagnosis. 76% of them were non-smokers, 95,6% had adenocarcinomas; 80%/13%/7% had respectively stage IV/ III/II–I at treatment initiation. EGFR mutations were identified on exon 19 (46,5%), exon 21 (40,3%), exon 20 (5,2%). A geriatric assessment was assessed in 35% of cases. Median time between first symptoms and diagnosis was 55 days. 97.3% of the patients were treated by TKI as first or second line. Median PFS was 11,9 months, 95% CI: 8,614,7. Response and disease control rates were 67% and 79% respectively. In 40% of the cases EGFR-TKI treatment was maintain beyond progression. After progression, 44,7% of the patients received another line of treatment (chemotherapy: 44,7%). Median OS was 20,9 months, 95% CI: 14,3-27,1. Main toxicities were cutaneous: 66% (grade 3/ 4:10%), diarrhea 56% (grade 3/4:15%, grade 5: 2%), others 25,6% (grade 3/4: 41%). Conclusions: In this real-world analysis, compared to younger, octogenarians patients with EGFR-mutated NSCLC treated by EGFR TKI present comparable outcomes and toxicity profile. Geriatric assessment is still under-used in this population. Clinical trial identification: IRBN 112016/CHUSTE CCTIRS N 15.779 Legal entity responsible for the study: Groupe Franc¸ais de Pneumo Cance´rologie (GFPC) Funding: AstraZeneca Disclosure: R. Corre: In the last five years, has received honoraria for attending scientific meetings, speaking, organizing research or consulting from Roche, Astra-Zeneca, Boehringer-Ingelheim, Lilly, Bristol-Myers-Squibb, Novartis. H. Doubre: In the last

Volume 28 | Supplement 5 | September 2017

five years, has received honoraria for attending scientific meetings, speaking, organizing research or consulting, from Astra-Zeneca, Novartis, Lilly, Boehringer-Ingelheim, Roche, Leo Pharma. C. Chouaid: In the past 5 years, has received fees for attending scientific meetings, speaking, organizing research or consulting from AZ, Boehringer Ingelheim, and Roche. J.B. Auliac: In the last five years, has received honoraria for attending scientific meetings, speaking, organizing research or consulting, from Boehringer Ingelheim, Hoffman-Roche, BMS, Lilly and Pfizer. All other authors have declared no conflicts of interest.

1357P

Prospective evaluation of the relationship between erlotinib concentration and efficacy in patients with non-small cell lung cancer harboring EGFR-activating mutations

H. Kenmotsu1, C.K. Imamura2, T. Kawamura1, T. Oyakawa1, S. Omori1, K. Nakashima1, K. Wakuda1, A. Ono1, T. Taira1, T. Naito1, H. Murakami1, N. Yamamoto3, T. Takahashi1, Y. Tanigawara2 1 Division of Thoracic Oncology, Shizuoka Cancer Center, Shizuoka, Japan, 2Department of Clinical Pharmacokinetics and Pharmacodynamics, Keio University School of Medicine, Tokyo, Japan, 3Internal Medhicine, Wakayama Medical University, Wakayama, Japan Background: Erlotinib is an epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI), which has shown improved clinical outcomes in the treatment of patients with non-small cell lung cancer (NSCLC) harboring EGFR exon 19 deletions and exon 21 L858R mutations. Previous reports regarding the relationship between erlotinib pharmacokinetics and pharmacodynamics evaluated toxicities but not efficacy. Therefore, to evaluate the relationship between erlotinib exposure and efficacy, we conducted this prospective study. Methods: Erlotinib was orally administered at a dose of 150 mg/body once daily to patients with NSCLC, who were not previously treated with EGFR-TKIs. A series of blood samples were taken at predetermined times on day 1 to calculate the area under the concentration-time curve (AUC). Erlotinib trough concentrartions (Ctrough) at each visit and the level of alpha1-acid glycoprotein (AAG), which is a binding protein of erlotinib in serum, were measured. Results: Of 70 patients enrolled, 61 had activating EGFR mutations (30 patients with exon 19 deletions and 31 with exon 21 L858R mutations). The AUC was 37.0 lg h/mL in median (range; 9.7-63.3). Objective response rate and median progression-free survival (PFS) were 72% and 12.4 months in the patients with EGFR-activating mutations. Response was not associated with AUC. There was also no significant difference in PFS between patients with AUC > 37.0 lg h/mL and 37.0 lg h/mL. Ctrough was significantly correlated with the grade of skin rash (p < 0.01), but not with objective response. In multivariate analyses, pretreatment AAG level, which was 0.97 g/L in median (0.533.83), was found to be a significant factor in PFS for patients with EGFR-activating mutations (median PFS; AAG > 0.97 g/L, 7.9 months; AAG 0.97 g/L, 16.8 months, p < 0.01). Conclusions: The lack of a relationship between erlotinib exposure and efficacy shows that the approved dose of erlotinib is sufficient to reach the therapeutic range in EGFRactivating mutant NSCLC, even with dose reduction due to toxicities. AAG level can be a prognostic factor for patients with NSCLC harboring EGFR-activating mutations treated with erlotinib. Clinical trial identification: UMIN000012862 (16-Jan-2014) Legal entity responsible for the study: Shizuoka Cancer Center Funding: JSPS KAKENHI Grant Disclosure: H. Kenmotsu, N. Yamamoto, T. Takahashi: Grants and personal fees from Chugai Pharmaceutical Co, Ltd. C.K. Imamura, Y. Tanigawara: Honoraria from Chugai Pharmaceutical Co, Ltd. T. Kawamura, S. Omori, K. Wakuda, A. Ono, H. Murakami: Personal fees from Chugai Pharmaceutical Co, Ltd. All other authors have declared no conflicts of interest.

1358P

First-in-human phase I study of PF-06747775, a third-generation mutant selective EGFR tyrosine kinase inhibitor (TKI) in metastatic EGFR mutant NSCLC after progression on a first-line EGFR TKI

H. Husain1, R.G. Martins2, S.B. Goldberg3, P. Senico4, W. Ma5, J. Masters6, N. Pathan6, D-W. Kim7, M.A. Socinski8, Z. Goldberg5, B.C. Cho9 1 Medicine, UC San Diego Moores Cancer Center, San Diego, CA, USA, 2Oncology/ Hematology, Seattle Cancer Care Alliance, Seattle, WA, USA, 3Medical Oncology, Yale Cancer Center, New Haven, CT, USA, 4Oncology, Pﬁzer, Collegeville, USA, 5Oncology, Pﬁzer, New York, NY, USA, 6Oncology, Pﬁzer, La Jolla, USA, 7Oncology, Seoul National University Hospital, Seoul, Republic of Korea, 8Oncology, University of Pittsburgh, Pittsburgh, USA, 9Oncology, Yonsei Cancer Center, Yonsei, Republic of Korea Background: PF-06747775 (“PF-7775”) is a selective, 3rd generation Epidermal Growth Factor Receptor (EGFR) TKI effective against EGFR sensitizing and T790M mutations in preclinical models. Enrollment was completed in a phase I study of PF7775 in EGFR mutant (EGFRmþ) NSCLC patients (pts) who progressed after a prior EGFR TKI. Methods: EGFRmþ NSCLC pts with resistance to first-line EGFR-TKIs enrolled in a multicenter trial in 6 dose escalation (25 mg–600 mg) and 2 dose expansion cohorts

doi:10.1093/annonc/mdx380 | 485

abstracts (200 mg and 300 mg). Biopsy for EGFR T790M at clinical progression was not required for study entry. PF-7775 was given orally once daily. All pts were assessed for response, adverse events (AEs), and pharmacokinetics. Plasma T790M testing was conducted retrospectively. Results: As of a 3 Feb17 data cutoff, 44 pts were enrolled (59% female, median age 63.5, Asian/Caucasian 73/25%).The most common all grade AEs in all enrolled patients ( 25%) were: diarrhea (57%), rash (59%), paronychia (52%), dermatitis acneiform (34%), stomatitis (32%), pruritus (27%), dry skin (25%), and rhinorrhea (25%). Grade 3 events were consistent with known EGFR TKI toxicities (diarrhea and skin toxicity) and easily managed. No grade 4 treatment related AEs were noted in dose escalation or expansion. Nineteen pts (43%) had dose reductions due to treatment related AEs at the higher dose cohorts and the RP2D of 200 mg was selected based on the AE profile and tolerability of drug at this dose for longer term administration. The prevalence rate of T790M, L858R, and del 19 in plasma samples, along with data correlating ORR (CR and PR rate) and CBR (CR þ PR þ SD) is ongoing. Conclusions: PF-7775 is well tolerated at 200 mg dose in EGFRmþ NSCLC pts with acquired resistance to first-line EGFR-TKIs. Data relating mutational status to response rate in plasma is ongoing. Clinical trial identification: NCT02349633 Legal entity responsible for the study: Pfizer Inc. Funding: Pfizer Inc. Disclosure: P. Senico, W. Ma, J. Masters, N. Pathan, Z. Goldberg: Employee of Pfizer and declare Pfizer stock ownership. B.C. Cho: Conducting Pfizer corporate-sponsored research. All other authors have declared no conflicts of interest.

1359P

Impact on OS and PFS of 2nd and 3rd generation TKI in EGFR mt1 and ALK1 pts: Results of the NOWEL network

J. Roeper1, A. Lueers1, M. Netchaeva1, M. Falk2, C. Hallas2, M. Tiemann2, N. Neemann3, L. Heukamp3, C. Wesseler4, G.H. Wiest4, D. Ukena5, S. Sackmann5, F. Griesinger1 1 Department of Internal Medicine-Oncology, Pius Hospital Oldenburg, University of Oldenburg, Oldenburg, Germany, 2Department of Hematopathology, Hematopathology Hamburg, Hamburg, Germany, 3Department of Pathology, NEO New Oncology, Cologne, Germany, 4Department of Oncology, Asklepios Klinik Harburg, Hamburg, Germany, 5Department of Oncology, Klinikum Bremen Ost, Bremen, Germany Background: Clinical research data shows that early mutation testing for pts with NSCLC stage IV could lead to an effective choice of therapy for pts with proven mutations. Targeted therapies achieve a higher ORR, OS, PFS and a better quality of life than chemotherapy in mtþ pts. With the advent of 2nd and 3rd generation TKÍs effective in 1st generation TKI resistant tumors, we wanted to study the impact of these drugs on the outcome of pts in a real life setting. Methods: 1381 pts from three cancer centers diagnosed with non-squamous cell NSCLC stage IV (UICC 7) were examined. Methods for the mutation testing was performed according to the German Oncopedia guidelines using either Sanger Sequencing V or COBAS or Next Generation Sequencing (hybrid capture NGS, New Oncology Cologne). Results: 879/1381 (64%) consecutive pts with non-squamous cell NSCLC from three cancer centers were studied for the presence of tumor mutations, especially for EGFR and ALK mtþ. The EGFR mtþ rate was 16.6% (141/847), and the ALK-translocation rate 3.8% (24/635). Median OS in EGFR mtþ pts was 28 (n ¼ 79) vs. 28 (n ¼ 38) vs. 16 (n ¼ 14) months respectively (center 1 vs. center 2 vs. center 3). Median OS in ALK mtþ pts was 24 months (n ¼ 17) in center 1 and 11 months (n ¼ 5) in center 2 (p < 0.033). The ORR in the CR/PR group was 54.2% for pts treated with chemotherapy and 77% for pts treated with TKI on 1st line therapy. The chance to reach a CR/PR on 1st line therapy is 2.83 higher for pts on TKI than for pts on chemotherapy (p < 0.02). The use of 3rd generation TKI Osimertinib (n ¼ 19) lead to a significantly higher OS (n ¼ 19, median OS 67 months) than the use of only 1st and 2nd generation TKI (n ¼ 111, median OS 23 months, p < 0.000). Pts treated with 3rd gen TKI had significantly longer PFS (11 months, n ¼ 7) than patients treated without 3rd generation TKI (5 months, n ¼ 45) (p < 0.037). Similarly, use of 2nd and 3rd generation ALKi impacted significantly on median OS: Crizotinib alone (n ¼ 8), 17 months, Crizotinib followed by Ceritinib and/or Brigatinib (n ¼ 10) median OS not reached and 3 months for other therapies (n ¼ 6) (p < 0.001). Conclusions: Small differences in OS were observed, depending on the treatment centers, but the use of multiple EGFR and ALK-I impacted highly significantly on the outcome of pts with EGFR and ALK mtþ in a real life setting. Legal entity responsible for the study: Faculty 6 University of Oldenburg Funding: None Disclosure: M. Falk: Member of Advisory board: Boehringer Ingelheim, Pfizer Professional fee: Roche, Astra Zeneca, Boehringer Ingelheim. M. Tiemann: Member of Advisory board: Novartis, Boehringer, Roche, Astra Zeneca Professional fee: Novartis, Boehringer, Roche, Astra Zeneca Corporate-sponsored Research: Novartis. L. Heukamp: Member of advisory board: BMS, Boehringer, Roche Diaganostics, Novartis F. Griesinger: Advisory board/Scientific Support/Professional fee: Ariad, Astra-Zeneca, Boehringer-Ingelheim, Bristol-Myer-Squibb, Celgene, Clovis, Lilly, Merck-SharpDome, Novartis, Pfizer, Roche. All other authors have declared no conflicts of interest. R

486 | NSCLC, metastatic

Annals of Oncology

1360P

Phase II trial of AZD9291 in second line treatment after acquired resistance with T790M mutation detected from circulating tumor DNA (LiquidLung-O-Cohort 2)

Y-C. Kim, C.K. Park, I-J. Oh, J-H. Lim, Y-D. Choi, H. Cho, S-J. Ahn, S-Y. Song, J.S. Yun, K-J. Na Lung Cancer Clinic, Chonnam National University Medical School & Chonnam National University Hwasun Hospital, Hwasun, Republic of Korea Background: Administering the best treatment after acquiring resistance to epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) requires the knowledge of the resistance status. In this trial, the treatment efficacy of osimertinib (AZD9291) was assessed in patients with non-small-cell lung carcinoma (NSCLC) harboring T790M resistance mutation, which was detected in the circulating tumor DNA (CtDNA) without re-biopsy of the tumor tissue. Methods: To prove 60% response rate of osimertinib compared to 30% as null hypothesis, and considering 10% drop out rate, 19 subjects was recruited. To extract CtDNA, 15 mL of peripheral blood was withdrawn and centrifuged immediately before storage. V V Cobas v2 RUO (Roche diagnostics) and PANA mutyper (Pangene, Korea) were used to detect the EGFR mutations from CtDNA. Osimertinib was prescribed as an 80mg tablet once in a day irrespective of the food intake. Results: Eighty patients with acquired resistance to prior EGFR TKIs were screened for T790M resistance mutation, and the CtDNA of 21 subjects (26.3%) showed T790M V V mutation. T790M mutation was detected by both PANA mutyper and Cobas in 13 V V cases, T790M was detected only by PANA mutyper in 4 cases, and only by Cobas in 4 cases. Nineteen subjects (age: 64.4 6 11.6 years old, 14 women, 5 men) were enrolled in this prospective single arm trial from September 2016 to April 2017. Prior EGFR TKIs were afatinib (n ¼ 3), erlotinib (n ¼ 4), gefitinib (n ¼ 10), erlotinib and afatinib (n ¼ 1), and gefitinib and afatinib (n ¼ 1). Twelve subjects had exon 19 deletion of EGFR gene, 4 had L858R point mutation, one showed exon 19 deletion and L858R, 1 had G719X, and 1 case showed no activating mutation. Conclusions: By April 2017, the response to osimertinib was evaluated in 13 subjects; 4 subjects dropped out from this trial before response evaluation, and the responses in 2 subjects are still pending for evaluation. Among the 13 subjects whose responses were evaluated (efficacy analysis set), partial remission was observed in 8 cases (61.5%). In the final efficacy analysis, toxicity and survival analyses will be performed. Clinical trial identification: NCT02769286 Legal entity responsible for the study: Young-Chul Kim Funding: AstraZeneca Disclosure: Y-C. Kim: This study was funded by AstraZeneca. All other authors have declared no conflicts of interest. R

R

R

R

1362P

R

R

Safety and clinical activity of DS-6051b, a ROS1/NTRK inhibitor, in Japanese patients with NSCLC harboring ROS1 fusion gene

M. Takeda1, T. Seto2, Y. Fujiwara3, N. Yamamoto3, K. Nosaki2, R. Toyozawa2, C. Abe4, R. Shiga4, K. Nakamaru5, K. Nakagawa1 1 Department of Medical Oncology, Kindai University Faculty of Medicine, Osaka, Japan, 2 Department of Thoracic Oncology, National Kyushu Cancer Center, Fukuoka, Japan, 3 Department of Experimental Therapeutics, National Cancer Center Hospital, Tokyo, Japan, 4Oncology Clinical Development Department, Daiichi Sankyo Co., LTD, Tokyo, Japan, 5Biomarker Department, Daiichi Sankyo Co., LTD, Tokyo, Japan Background: ROS1 or NTRK genetic abnormality is a key oncogenic driver, especially in non-small cell lung cancer (NSCLC). DS-6051b is an orally available and potent selective small molecule inhibitor of ROS1 and NTRK. Preclinical pharmacology studies exhibited antitumor activity of DS-6051b against several types of tumor with ROS1 or NTRK fusion gene. Methods: This is an ongoing phase 1 study in Japanese patients with advanced solid tumors harboring either a ROS1 or NTRK fusion gene. Patients receive continuous once daily (QD) dosing of DS-6051b. Pharmacokinetics (PK) samples are collected from Day1 to Day22. If patients consented, liquid biopsy samples for circulating tumor DNA (ctDNA) analysis are collected at screening and study discontinuation. The primary objective is to evaluate the safety profile and secondary objectives are to determine the maximum tolerated dose (MTD), the recommended phase 2 dose (RP2D) and to assess the PK profile. Exploratory objectives are to assess the tumor responses to DS6051b by means of Investigator evaluation per RECIST v.1.1 and to explore biomarkers. ClinicalTrials.gov identifier: NCT02675491. Results: As of Apr 10, 2017, 15 patients are enrolled. The median age is 51 (34-69) years, 46.7% are female, all 15 patients are ROS1 fusion positive NSCLC, and 4 had prior crizotinib (CRZ) treatment. Patients received DS-6051b at doses of 400mg QD (n ¼ 6), 800mg QD (n ¼ 3) and 600mg QD (n ¼ 6). Common adverse events are AST increased, ALT increased, diarrhea and nausea. There are no DLTs in the 400mg and 600mg QD cohorts, and 2 out of 3 patients in the 800mg QD cohort experienced DLTs with grade 3 ALT increased. In 12 patients who had measurable target lesions, 7 demonstrated partial response (PR) and 5 demonstrated stable disease. In 9 CRZ-naı¨ve patients, the overall response rate is 66.7% (6 PRs) and the disease control rate is 100%. Early clinical response was also shown in a subject with brain metastasis. The plasma drug concentration increased in a dose-dependent manner. Currently, ctDNA analyses obtained from pre/post DS-6051b treatment are ongoing.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology Conclusions: DS-6051b is well tolerated in Japanese patients with NSCLC and effective especially in CRZ-naı¨ve. The MTD/RP2D is identified as 600mg QD. Legal entity responsible for the study: Daiichi Sankyo Co., Ltd. Funding: None Disclosure: T. Seto: Consulting and advisory services, speaking or writing engagements, public presentations; 14 companies. Direct research support to the responsible project lead; 12 companies (including Daiichi Sankyo Co. Ltd.). Y. Fujiwara: Consulting and advisory services, speaking or writing engagements, public presentations; Bristol-Myers Squibb, Ono Pharmaceutical Co. Ltd., MSD, K.K. (e500-e20’000 p.a). N. Yamamoto: Consulting and advisory services, speaking or writing engagements, public presentations; Eisai Co., Ltd, Takeda Pharmaceutical Company, Ltd., OncoTherapy Science (e500 - e20’000 p.a). K. Nosaki: Consulting and advisory services, speaking or writing engagements, public presentations; 7 companies. Direct research support to the responsible project lead; 2 companies. C. Abe, R. Shiga, K. Nakamaru: Full-time employee of Daiichi Sankyo. Co., Ltd. K. Nakagawa: Consulting and advisory services, speaking or writing engagements, public presentations; Daiichi Sankyo Co,. Ltd., Pfizer Japan Inc. Direct research support to the responsible project lead; Daiichi Sankyo Co,. Ltd., Pfizer Japan Inc. All other authors have declared no conflicts of interest.

1363P

Efficacy and safety of abemaciclib combined with either LY3023414 or pembrolizumab in stage IV NSCLC

P. Garrido Lopez1, J. Goldman2, K. Kelly3, E.S. Kim4, T.J. Beck5, M. Alonso Garcia6, Z. Yang7, K-J. Ingram8, L. Amstutz8, W.J. John8, M. Provencio Pulla9 1 Medical Oncology, Hospital Universitario Ram on y Cajal, Madrid, Spain, 2Hematology & Oncology, University of California-Los Angeles, Santa Monica, USA, 3Medical Oncology, UC Davis Comprehensive Cancer Center, Sacramento, CA, USA, 4Medical Oncology, Levine Cancer Institute, Charlotte, USA, 5Medical Oncology, Highlands Oncology Group, Fayetteville, USA, 6Medical Oncology, Hospital Universitario Virgen del Rocıo, Seville, Spain, 7Biometrics, Eli Lilly and Company, Indianapolis, USA, 8Medical Affairs, Eli Lilly and Company, Indianapolis, USA, 9Medical Oncology, Servicio de Oncologıa Me´dica Hospital Universitario Puerta de Hierro, Madrid, Spain Background: Abemaciclib (abema), a cyclin D kinase 4 & 6 inhibitor, has single-agent activity and an acceptable safety profile when dosed continuously in patients with previously treated metastatic NSCLC (NCT01394016). In tumor models, CDK inhibition induces an escape pathway involving PI3kinase (PI3K) and abema induces synergistic immune activation with checkpoint inhibitors. We report on activity and safety of abema plus LY3023414 (LY), a PI3K/mTOR dual inhibitor, and abema plus pembrolizumab (pembro), an anti-PD-1 antibody, in an ongoing Phase 1b open-label, 3 þ 3 multicenter trial of previously treated advanced NSCLC (NCT02079636). Methods: For escalation, Abema [100, 150 mg, or 200 mg (cohort D only)] was given orally on a continuous schedule every 12 hours (q12h) with LY at 100, 150, or 200 mg q12h (cohort D) or with pembro at 200 mg I.V. infusion q3 weeks (cohort E). Confirmatory cohorts were given 150 mg abema with 150 mg LY or 200 mg pembro. Pts were treated until progression or other discontinuation criteria were met. Responses were evaluated with RECIST v1.1. Safety assessments followed the NCICTCAE v4.0. Results: As of 01-Mar-2017, cohort D (n ¼ 29) had 62.1% males, 37.9% 65 years of age, median # prior systemic therapies¼3; 86.2% stage IV; 72.4% adenocarcinoma; 62.1% ECOG PS ¼ 1. 9 pts (31%) had stable disease (SD), 3 pts had progressive disease (PD), and the status for the remaining 17 pts was unknown or under evaluation. There were 3 deaths unrelated to study drug (2 disease related and 1 stroke). 24/29 pts had a treatment emergent, related AE (TRAE), 10/24 had a Grade 3/4 TRAE. Any grade TRAEs (>30% pts) were nausea (51.7%), diarrhea (51.7%), vomiting (41.4%), and decreased appetite (31%). Cohort E had 19 pts entered (42.1% male, 42.1% 65 years of age, median # prior systemic therapies¼2; 52.6% stage IV; 89.5% adenocarcinoma; 57.9% ECOG PS ¼ 1). 8 pts (42.1%) had SD, 1 had PD, and the status for the remaining 10 pts was unknown or under evaluation. There were 3 disease related deaths. 15/19 pts had a TRAE, 5/15 had a G3/4 TRAE. Any grade TRAEs (>30% pts) were fatigue (n ¼ 47.4%) and diarrhea (36.8%). Conclusions: To date, stable disease as best response and acceptable safety have been observed using combinations of abema and either LY or pembro in advanced NSCLC. Clinical trial identification: NCT02079636 Legal entity responsible for the study: Eli Lilly and Company Funding: Eli Lilly and Company Disclosure: P. Garrido Lopez: Advisor/Board member: MSD, Pfizer, BMS, Novartis, Roche, BI, Guardant Speakers Bureau: MSD, Pfizer, BMS, Novartis, Roche Honorarium recipient: BI J. Goldman: Research grant from Eli Lilly and company Eli Lilly and Company’s Scientific Advisory Board member. K. Kelly: Attended a Lilly advisory board in Feb 2017 that discussed Abemaciclib. E.S. Kim: Celgene - Consultant Boehringer Ingelheim - Consultant Eli Lilly - Consultant AstraZeneca – Consultant. Z. Yang, L. Amstutz, W.J. John: Full time employee of Eli Lilly and Company. K-J. Ingram: Stock ownership in Eli Lilly and Company. Employee of Eli Lilly and company. M. Provencio Pulla: Corporate-sponsored research (investigator): GI Therapeutics, Eli Lilly and Company. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

1364P

Final clinical results from SUNRISE: A phase III, randomized, doubleblind, placebo-controlled multicenter trial of bavituximab plus docetaxel in patients with previously treated stage IIIb/IV nonsquamous non-small cell lung cancer

R. Palmero1, P. Bidoli2, I.M. Bondarenko3, M. Boyer4, P. Germonpre5, D. Ghizdavescu6, A. Kotsakis7, H. Lena8, G. Losonczy9, K. Park10, M. Reck11, W-C. Su12, N. Kallinteris13, M. Tang13, J. Lai13, J. Shan13, D.R. Spigel14 1 Medical Oncology, Institut Catal a d’Oncologia Hospital Duran i Reynals, Barcelona, Spain, 2Medical Oncology, Azienda Ospedaliera S. Gerardo U.O. Oncologia Medica, 3 Monza, Italy, Oncology, Dnipropetrovsk Medical Academy, City Multiple-Discipline Clinical Hospital, Dnepropetrovsk, Ukraine, 4Medical Oncology, Chris O’Brien Lifehouse, Camperdown, Australia, 5Medical Oncology, AZ Maria Middelares, Gent, Belgium, 6 Medical Oncology, Municipal Hospital, Ploiesti, Romania, 7Medical Oncology, University Hospital of Heraklion, Heraklion, Greece, 8Medical Oncology, CHU de Pontchaillou, Rennes, France, 9Medical Oncology, Semmelweis University I. Faculty of Medicine, Budapest, Hungary, 10Division of Heamatology/Oncology, Department of Medicine, Samsung Medical Center, Seoul, Republic of Korea, 11Department of Thoracic Oncology, Lung Clinic Grosshansdorf, Airway Research Center North (ARCN), Member of the German Center for Lung Research (DZL), Grosshansdorf, Germany, 12Medical Oncology, National Cheng Kung University, Tainan, Taiwan, 13Clinical and Regulatory, Peregrine Pharmaceuticals, Inc., Tustin, CA, USA, 14Medical Oncology, Sarah Cannon Research Institute/Tennessee Oncology, PLLC, Nashville, TN, USA Background: Exposed phosphatidylserine (PS) in the tumor microenvironment is highly immunosuppressive. Bavituximab targets PS and repolarizes M2 macrophages to M1 resulting in production of pro-inflammatory cytokines such as IFN-c and IL-12, maturation of dendritic cells, and tumor specific cytotoxic T lymphocyte immunity. In a prior blinded Phase II trial in 2nd-line nonsquamous NSCLC, bavituximab þ docetaxel was well-tolerated and demonstrated 60% improvement (11.7 vs 7.3 months) in median overall survival (mOS) (HR, 0.66; P ¼ 0.11) compared to control. Methods: 597 patients with Stage IIIb/IV nonsquamous NSCLC that progressed on platinum-doublet chemotherapy were randomized 1:1 to receive up to six 21-day cycles of docetaxel in combination with weekly 3 mg/kg bavituximab (BþD) or placebo (D) until progression or toxicity. The primary endpoint was OS. Secondary endpoints included progression-free survival (PFS), objective response rate (ORR) and safety. Results: With 12 months follow-up from the last patient randomized and 85% of the targeted OS events reached, mOS was 10.5 months (95% confidence interval [CI], 8.411.9) among 297 patients in BþD and 10.9 months (95% CI, 9.2-12.1) among 300 patients in D (HR, 1.06; P ¼ 0.533). PFS was 4.2 months (95% CI, 3.9-4.6) in BþD and 4.1 months (95% CI, 3.2-4.8) in D (HR, 1.02; P ¼ 0.876). The ORR was 15% in BþD vs. 11% in D (odds ratio, 0.7; P ¼ 0.15). The safety profile was similar between groups. Grade 3 or higher adverse events occurred in 68% of patients in BþD and 60% in D. In an exploratory analysis of OS for patients who received subsequent immune checkpoint inhibitors (ICI), the mOS was not reached (95% CI, 15.2-NA) in BþD (n ¼ 46) and 12.6 months (95% CI, 10.4-17.8) in D (n ¼ 47) (HR, 0.46; P ¼ 0.006). Conclusions: The combination of BþD was well-tolerated though no OS difference was observed compared to D alone in the ITT population of previously treated nonsquamous NSCLC. An exploratory analysis of patients who received subsequent ICI found significantly longer OS in patients who received prior BþD than those who received D and support further clinical investigation of BþICI in NSCLC. Clinical trial identification: NIH ¼ NCT01999673 EudraCT ¼ 2013-003953-13 Legal entity responsible for the study: Peregrine Pharmaceuticals Inc. Funding: Peregrine Pharmaceuticals Inc. Disclosure: P. Bidoli: Eli Lilly personal fees and advisory board BMS personal fees and advisory board Boehringer personal fees and advisory board. M. Reck: Honoraria for lectures and consultancy with Hoffmann-La Roche, Lilly, MSD, Merck, BMS, AstraZeneca, Celgene, Boehringer-Ingelheim, Pfizer, Novartis. N. Kallinteris, J. Lai: Peregrine Pharmaceuticals Inc. ¼ employee, stock ownership M. Tang: Peregrine ¼ Employee, stock owner. J. Shan: Employee, officer and stock owner for Peregrine Pharmaceuticals Inc. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx380 | 487

abstracts 1365P

Effect on quality of life (QOL) of adding cisplatin to single-agent firstline chemotherapy in elderly patients with advanced non-small cell lung cancer (NSCLC): A joint analysis of the multicentre, randomized, phase 3 MILES-3 and MILES-4 studies

M.C. Piccirillo1, A. Morabito2, P. Maione3, A. Luciani4, L. Cavanna5, L. Bonanno6, E. Piazza7, S. Leo8, S. Cinieri9, F. Morgillo10, M.A. Burgio11, D. Bilancia12, F. Rosetti13, A. Montanino2, A. Manzo2, L. Arenare1, S. Signoriello14, C. Gallo14, F. Perrone1, C. Gridelli3 1 Clinical Trials Unit, Istituto Nazionale Tumori – I.R.C.C.S - Fondazione Pascale, Naples, Italy, 2Thoraco-Pulmonary Oncology, Istituto Nazionale Tumori – I.R.C.C.S - Fondazione Pascale, Napoli, Italy, 3Medical Oncology, Azienda Ospedaliera S. Giuseppe Moscati, Avellino, Italy, 4Medical Oncology, Azienda Ospedaliera S. Paolo, Milan, Italy, 5Medical Oncology, Azienda Ospedaliera Piacenza, Piacenza, Italy, 6Medical and Experimental Oncology Department, Medical Oncology 2, Istituto Oncologico Veneto IRCCS, Padua, Italy, 7Medical Oncology Unit, AOU Sacco Hospital, Milan, Italy, 8Geriatric Oncology, Ospedale Vito Fazzi, Lecce, Italy, 9Medical Oncology Unit A, Perrino Hospital, Brindisi, Italy, 10Dipartimento Medico-Chirurgico di Internistica Clinica e Sperimentale “F. Magrassi E A. Lanzara”, Universit a degli Studi della Campania “Luigi Vanvitelli”, Naples, Italy, 11Oncology, IRCCS Istituto Scientiﬁco Romagnolo per lo Studio e la Cura dei 12 Tumori, Meldola, Italy, Medical Oncology, AZ. Ospedaliera San Carlo, Potenza, Italy, 13 Oncologia ed Ematologia Oncologica, Mirano ULSS 3, Serenissima Regione Veneto, a degli Studi della Campania “Luigi Mirano, Italy, 14Medical Statistics, Universit Vanvitelli”, Naples, Italy Background: The effect on QOL of adding platinum to first line treatment of elderly patients (pts) with advanced NSCLC is unknown. In this setting, MILES-3 and MILES4 trials prospectively showed that adding cisplatin (Cis) to single-agent gemcitabine (Gem) or pemetrexed (Pem) does not significantly prolong overall survival. Methods: Advanced NSCLC pts, >70 years old, ECOG performance status 0-1, were eligible. In MILES-3, pts with any tumor histology were randomly assigned to CisGem or Gem. In MILES-4, pts with non-squamous histology were randomly assigned to CisGem, Gem, CisPem or Pem. The trials were joined together because of slow accrual. Overall survival was the primary endpoint. QOL (EORTC QLQ C30 and LC13) was a secondary endpoint. Five questionnaires were planned in MILES-3 and 7 in MILES-4; QOL was measured in both the trials at 3 time points (baseline, end cycle 1, end cycle 2) used for joint analysis. Intention-to-treat strategy was applied; analyses were adjusted for baseline QOL, stage, PS, gender, age, size of centre, trial, histotype and companion drug. Results: Overall, 458/531 pts (86.3%) answered baseline questionnaire. Rate of missing questionnaires at end cycle 1 and 2 was slightly higher among pts receiving Cis. Mean change in fatigue after cycle 1 (P ¼ 0.01) and in sore mouth after cycle 2 (P ¼ 0.02) were worse with Cis. Using a 10% change from baseline as clinically relevant threshold to categorize response, alopecia was significantly worse with Cis (P ¼ 0.05). In time to deterioration analysis with progression/death as competitive risk, sore mouth and alopecia deteriorated more with Cis (HR 1.72 95%CI 1.02-2.89 P ¼ 0.04 and HR 1.84 95%CI 1.09-3.10, P ¼ 0.02, respectively). Response analysis in MILES-3 confirmed findings of the joint analysis while time to deterioration analysis in MILES-4 did not find any significant difference. Cis did not significantly improve any QOL item, in any type of analysis. Conclusions: The addition of Cisplatin did not improve QOL of elderly patients with advanced NSCLC. Partially supported by AIFA (grant FARM8KAJZK) and Eli Lilly. Clinical trial identification: MILES 3 EudraCT number: 2009 01354036 MILES 4 EudraCT number: 2012-000164-25 Legal entity responsible for the study: Istituto Nazionale Tumori, IRCCS, Fondazione G. Pascale Funding: Eli Lilly Disclosure: M.C. Piccirillo: Received honoraria by Bayer for consulting and travel payment by Roche and Bayer for congresses. A. Morabito: Consulting or Advisory Role for Bristol-Myers Squibb and that received Honoraria by Roche, Boehringer Ingelheim, AstraZeneca, Lilly, Bayer, Italfarmaco and Pfizer. P. Maione: Honoraria by Lilly, Boehringer Ingelheim, AsaraZeneca and Roche, and that had Consulting or Advisory Role for Lilly, Boehringer Ingelheim, AsaraZeneca, Roche. F. Perrone: Received Travel, Accommodations, Expenses by Roche, Lilly, Bayer, Daiichi Sankyo, and received Honoraria by Amgen, Novartis, Lilly, Roche, Bayer, Daiichi Sankyo, and that his Institution had Research Funding from Roche, Bayer. C. Gridelli: Consulting Role for MSD, BMS, Celgene, Novartis, Roche, had Speakers’ Bureau from BMS, MSD, Celgene, Novartis, Roche, had Research Funding from Lilly, received Travel Expenses by MSD, BMS, Celgene, Novartis All other authors have declared no conflicts of interest.

Annals of Oncology 1366P

Effect of nab-paclitaxel/carboplatin (nab-P/C) induction therapy on quality of life (QoL) of patients with squamous (SCC) non-small cell lung cancer (NSCLC) (ABOUND.sqm)

S. Ponce Aix1, V. Villaflor2, J. Knoble3, M. Thomas4, J. von Pawel5, S. Bailey6, M.A. Hussein7, M.A. Razaq8, K. Sabbath9, P. Staib10, T. Chen11, N. Trunova11, D.R. Spigel12 1 Oncology/Hematology, University Hospital 12 de Octubre, Madrid, Spain, 2Oncology/ Hematology, Robert H Lurie Comprehensive Cancer Center of Northwestern University, Chicago, USA, 3Oncology/Hematology, Mark H. Zangmeister Center, Columbus, USA, 4 Oncology/Hematology, Thoraxklinik Heidelberg, Heidelberg, Germany, 5Oncology/ Hematology, Asklepios-Fachklinikum, Gauting, Germany, 6Oncology/Hematology, Hazard ARH Regional Medical Center Cancer Center, Hazard, USA, 7Oncology, Florida Cancer Specialists, Leesburg, FL, USA, 8Oncology, Stephenson Cancer Center OU, Oklahoma City, USA, 9Oncology/Hematology, Harold Leever Regional Cancer Center, amatologie/ Waterbury, USA, 10Oncology/Hematology, St. Antonius Hospital Klinik fu¨r H€ Onkologie, Eschweiler, Germany, 11Oncology, Celgene Corporation, Summit, USA, 12 Medical Oncology, Sarah Cannon Research Institute/Tennessee Oncology, PLLC, Nashville, TN, USA Background: Patients with advanced NSCLC experience a high symptom burden; therefore, identifying a treatment that maintains or improves QoL is important. QoL outcomes in patients with SCC NSCLC receiving nab-P/C in the induction part of the ABOUND.sqm study are reported. Methods: Patients with stage IIIB/IV SCC NSCLC and no prior chemotherapy for metastatic disease received 4 cycles of induction therapy with nab-P 100 mg/m2 days 1, 8, and 15 þ C area under the curve 6 on day 1 (21-day cycles). Patients not progressing after induction received (2:1) maintenance nab-P 100 mg/m2 days 1 and 8 (21-day cycles) þ best supportive care (BSC) or BSC alone until progression/unacceptable toxicity. The primary endpoint is progression-free survival (randomization to maintenance). Patient-reported QoL (exploratory endpoint) was assessed on day 1 of each cycle using the Lung Cancer Symptom Scale (LCSS) and EuroQoL 5 Dimensions-5 Levels (EQ-5D-5L). Results: In 343 patients receiving treatment in the induction phase were evaluated. Median age was 68 years; 90% were white, 68% male, and 67% had ECOG PS 1. Of 332 patients treated for 2 cycles, 298 (90%) completed baseline þ 1 postbaseline QoL assessment. During induction, the mean change from baseline in LCSS symptom burden index and total score ranged from 5.5%-7.8% and 5.5%-7.7%, respectively. Clinically meaningful improvements ( 10 mm [visual analog scale]) from baseline were observed in composite LCSS pulmonary symptom items of cough, shortness of breath, and hemoptysis in 44% of patients. Each individual dimension of the EQ-5D5L was maintained/improved from baseline in the majority of patients (82%-91%), and 32% reported complete resolution at least once during treatment. Conclusions: QoL was improved/maintained in patients with advanced SCC NSCLC treated with nab-P/C induction therapy. These results continue to support nab-P/C as a treatment option in patients with SCC NSCLC, as was initially demonstrated in a subset analysis of the phase III registration trial. NCT02027428. Clinical trial identification: NCT02027428 Legal entity responsible for the study: Celgene Corporation Funding: Celgene Corporation Disclosure: V. Villaflor: Research funding from Celgene and Novartis paid directly to University of Chicago. J. Knoble: Consulting or advisory role for Cardinal Health, Speakers’ bureau for Novartis, Alexion and Celgene. M. Thomas: Received honoraria for an advisory/speaker role from: Celgene, Astrazeneca, Roche, BMS, Lilly, Novartis, Boehringer. P. Staib: Honoraria, consulting or advisory role, speaker’s bureau and research funding: Celgene. T. Chen, N. Trunova: Employment and Stock Ownership: Celgene. D.R. Spigel: Research funding, consulting or advisory role, and travel, accommodations, expenses: Celgene. All other authors have declared no conflicts of interest.

1367P

Quality of life (QoL) in elderly NSCLC patients (pts) treated with nabpaclitaxel/carboplatin (nab-P/C) in the ABOUND.701 trial

C.J. Langer1, E. Anderson2, R. Jotte3, J. Goldman4, D. Haggstrom5, M. Modiano6, M. Socoteanu7, D. Smith8, C. Dahkil9, K. Konduri10, E. Kim10, A. Sanford11, K. Amiri11, J. Weiss12 1 Oncology/Hematology, Abramson Cancer Center, Philadelphia, PA, USA, 2Oncology/ Hematology, Knight Cancer Institute, Portland, OR, USA, 3Oncology/Hematology, Rocky Mountain Cancer Center, Denver, CO, USA, 4Medical Oncology, University of CaliforniaLos Angeles, Santa Monica, CA, USA, 5Oncology/Hematology, Levine Cancer Institute, Charlotte, NC, USA, 6Oncology/Hematology, ACRC/Arizona Clinical Research Center, Tucson, AZ, USA, 7Oncology/Hematology, Texas Oncology, Longview, TX, USA, 8 Oncology/Hematology, Compass Oncology, Vancouver, BC, USA, 9Oncology/ Hematology, Cancer Center of Kansas, Wichita, KS, USA, 10Oncology/Hematology, Baylor Charles A. Sammons Cancer Center, Dallas, TX, USA, 11Oncology/Hematology, Celgene Corporation, Summit, NJ, USA, 12Oncology/Hematology, UNC Lineberger Comprehensive Cancer Center, Chapel Hill, NC, USA Background: Treatment of elderly pts with advanced NSCLC remains challenging, and the impact of therapies on QoL can be an important factor in clinical decisions. nab-P/ C demonstrated efficacy in a subset of pts 70 yrs with NSCLC in a phase 3 trial.

488 | NSCLC, metastatic

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology ABOUND.70þ was designed to determine whether a 1-wk break can further improve tolerability of nab-P/C in these patients. QoL outcomes are reported here. Methods: Pts 70 yrs with locally advanced/metastatic NSCLC were randomized 1:1 to first-line nab-P 100 mg/m2 on d 1, 8, and 15 þ C AUC 6 on d 1 of a 21-day cycle (Arm A) or the same regimen with a 1-week break between cycles (Arm B). Primary endpoint: percentage of pts with grade 2 peripheral neuropathy or grade 3 myelosuppression. Key secondary endpoints: PFS, ORR, OS for which statistical analyses do not control for type I error (P values unadjusted). QoL (exploratory endpoint) was assessed using Lung Cancer Symptom Scale (LCSS) and EuroQoL-5 Dimensions-5 Levels (EQ-5D-5L) at d 1 of each cycle. Results: At interim evaluation, primary endpoint was similar across arms, resulting in early closure of enrollment. In Arms A and B, 78% and 79% completed a baseline and 1 postbaseline QoL assessment. LCSS item of cough improved with each cycle; at the end of cycle 6, mean change from baseline in Arms A and B was 25.4 and 13.8 mm (visual analog scale). For cough, median time to deterioration (TTD) was 4.4 and 4.7 mos (P ¼ 0.7003). For the composite LCSS pulmonary symptom items of cough, shortness of breath, and hemoptysis, the median TTD was 4.4 and 6.0 mos (P ¼ 0.3347). Mean maximum improvement (at any point during treatment) in EQ-5D-5L visual analog scale was 10.1 and 12.8 points. Table lists key safety, efficacy and QoL data.

Table: 1367P Arm A n ¼ 71 Safety Primary endpoint, n (%) 52/68 (76) P value 0.9258 Grade 2 peripheral neuropathy 25/68 (37) Grade 3 myelosuppression 48/68 (71) Neutropenia 39/68 (57) Anemia 14/68 (21) Thrombocytopenia 17/68 (25) Efficacy Conﬁrmed ORR, % 24 P value 0.0376 PFS, median, months 3.6 P value 0.0019 HR (95% CI) 0.48 (0.30 - 0.76) OS, median, months 15.2 P value 0.1966 HR (95% CI) 0.72 (0.44 - 1.19) QoL Mean maximum improvement from baseline, mm LCSS Total score 5.8 LCSS Pulmonary symptom 9.2

Arm B n ¼ 72

54/70 (77) 25/70 (36) 45/70 (64) 39/70 (56) 17/70 (24) 12/70 (17) 40 7.0

16.2

11.7 14.9

1368P

T. Ninomiya1, K. Hotta2, K. Ohashi1, T. Kubo1, D. Harada3, T. Kozuki3, N. Nogami3, I. Oze4, S. Hosokawa5, A. Bessho5, H. Yoshioka6, K. Kudo7, S. Kuyama7, S. Harita8, I. Takata9, N. Fujimoto10, T. Moritaka11, H. Ichikawa12, N. Takigawa13, K. Kiura1 1 Respiratory Medicine, Okayama University Hospital, Okayama, Japan, 2Center for Innovative Clinical Medicine, Okayama University Hospital, Okayama, Japan, 3 Respiratory Medicine, National Hospital Organization Shikoku Cancer Center, Matsuyama, Japan, 4Division of Molecular and Clinical Epidemiology, Aichi Cancer Center, Nagoya, Japan, 5Department of Respiratory Medicine, Japanese Red Cross Okayama Hospital, Okayama, Japan, 6Department of Respiratory Medicine, Kurashiki Central Hospital, Kurashiki, Japan, 7Respiratory Medicine, Iwakuni Medical Center, Iwakuni, Japan, 8Respiratory Medicine, Chugoku Central Hospital, Fukuyama, Japan, 9 Respiratory Medicine, Fukuyama City Hospital, Fukuyama, Japan, 10Department of Medical Oncology, Okayama Rosai Hospital, Okayama, Japan, 11Respiratory Medicine, Ehime Prefectural Central Hospital, Matsuyama, Japan, 12Respiratory Medicine, KKR Takamatsu Hospital, Takamatsu, Japan, 13Internal Medicine 4, Kawasaki Medical School General Medical Center, Okayama, Japan Background: Although nab-PTX plus carboplatin is one of the standard treatment for chemo-naive advanced non-small cell lung cancer (NSCLC), the efficacy, safety and optimal schedule of nab-PTX monotherapy as 2nd or 3rd line for NSCLC pts without any driver mutations remains unknown. Methods: This was a single arm phase I/II study. Eligible pts are advanced NSCLC without EGFR mutation and ALK rearrangement that progressed after platinum-doublet chemotherapy. The pts were received 100 mg/m2 of nab-PTX on day 1, 8, 15 and 22 (level 0) or on day 1, 8, and 15 (level -1) every 4-week in phase I. Dose limiting toxicities (DLT) were assessed and the recommended schedule was determined in the phase I. The primary endpoint was objective response rate (ORR), assuming that estimated ORR was 15% and threshold ORR was 5% with a error of 0.05 and b error of 0.2 in the phase II part. Total 55 pts were planned to be enrolled. Results: The recommended schedule of nab-PTX was determined as the level -1, because the DLTs were found in 4 of 5 pts in level 0. Total 55 pts were enrolled in the phase II and the characteristics were as followings; median age, 66 years (range, 41–90 years), male/female ¼ 40/15, PS 0/1/2 ¼ 12/39/4, 2nd/3rd line ¼ 34/21, adeno/squamous/ large/others ¼ 34/17/2/2. The median number of treatment cycles was three (range, 1– 10). The ORR was 7.3% (95% [CI], 2.0 – 17.6%) (PR (n ¼ 4), SD (n ¼ 26), PD (n ¼ 24), and NE (n ¼ 1)). At the median follow-up time of 5.3 months (range, 1.9 – 26.0 months) for all pts, the median PFS was 3.4 months (95%[CI], 1.9 – 4.0 months). Treatment related grade 3 or 4 toxicities were neutropenia (36%), pulmonary infection (3.6%), and pneumonitis (5.4%). One patient (2%) was died due to treatment-related ARDS. Conclusions: In phase I part, we confirmed that schedule level -1 was tolerable and the schedule had been recommended. In phase II part, this study failed to meet predefined primary endpoint although PFS was comparable and toxicity was acceptable for pts with advanced NSCLC without any driver mutations as 2nd or 3rd line treatment. Clinical trial identification: UMIN000012404. Legal entity responsible for the study: Okayama Lung Cancer Study Group Funding: None Disclosure: All authors have declared no conflicts of interest.

1369P Conclusions: These results support nab-P/C as a treatment option in elderly pts with NSCLC. Safety (primary endpoint) and OS were similar across the two arms, while there was a signal of improvement in ORR, PFS, and QoL with a 1-wk break. NCT02151149. Clinical trial identification: NCT02151149 Legal entity responsible for the study: Celgene Corporation Funding: Celgene Corporation Disclosure: C.J. Langer: Consultant/Advisor Role: Celgene. J. Goldman, E. Kim: Research funding: Celgene. K. Konduri: Consultant/Advisory Role: Celgene, Boehringer Ingelheim, DAVA – Pharmaceuticals. A. Sanford, K. Amiri: Employment and stock ownership: Celgene. J. Weiss: Astellas: CME & company sponsored trials (CST); AZ, Biodesix, Clovis, Oncoples: consulting; Celgene: IIT support, CST support, SC; EMD Serono: DSMB member; GSK: IIT support; Medimmune, Pfizer: CST support, IIT support; Merck: CS Tsupport, IIT support. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

Phase I/II trial of weekly nab-paclitaxel as 2nd or 3rd line treatment in NSCLC without driver mutations. (OLCSG1303)

nab-paclitaxel/carboplatin (nab-P/C) induction therapy in squamous (SCC) non-small cell lung cancer (NSCLC): Interim safety results from ABOUND.sqm

C. Gridelli1, M. McCleod2, D. Morgensztern3, D.B. Daniel4, R. Page5, T. Wehler6, O. Juan7, B. Levy8, A. Ardizzoni9, T. Berry10, T. Chen10, N. Trunova10, R. Jotte11 1 Oncology, S.G. Moscati Hospital, Avellino, Italy, 2Oncology, Florida Cancer Specialists, Cape Coral, USA, 3Oncology, Washington University Schoo of Medicine, St Louis, MO, USA, 4Oncology, Tennessee Oncology, Chattanooga, TN, USA, 5Oncology, The Center for Cancer and Blood Disorders, Fort Worth, TX, USA, 6Oncology, University Hospital Mainz, Mainz, Germany, 7Medical Oncology, Hospital Universitari i Polite`cnic La Fe, Valencia, Spain, 8Oncology, Beth Israel Medical Center, New York, NY, USA, 9Oncology, Universitari iPolite´cnicLa Fe, Valencia, Italy, 10Oncology, Celgene Corporation, Summit, USA, 11Oncology/Hematology, Rocky Mountain Cancer Center, Denver, CO, USA Background: The ABOUND.sqm trial is currently investigating treatment outcomes of 4 cycles of nab-P/C followed by nab-P maintenance therapy in patients (pts) with SCC NSCLC. Given that tolerability is an important consideration for treatment decisions, this analysis evaluated the safety of nab-P/C during induction. Methods: Chemotherapy-naive pts with stage IIIB/IV SCC NSCLC received 4 cycles of induction therapy with nab-P 100 mg/m2 d1, 8, and 15 þ C area under the curve 6 d 1 (21-d cycles). Pts not progressing after induction received (2:1) maintenance nab-P 100 mg/m2 d1 and 8 (21-d cycles) þ best supportive care (BSC) or BSC alone until progression/unacceptable toxicity. The primary endpoint is progression-free survival (randomization to maintenance). Secondary endpoints include safety analyzed as treatment-emergent adverse events (TEAEs), overall survival, overall response rate, and disease control rate.

doi:10.1093/annonc/mdx380 | 489

abstracts Results: A total of 334 pts were included in this analysis. Median age was 68 yrs; 90% were white, 68% male, and 68% had ECOG PS 1. During induction, 145/334 pts (43%) discontinued treatment. Of these, 51/145 (35%) discontinued due to progressive disease, 39/145 (27%) due to adverse events, 16/145 (11%) each due to death and pt withdrawal, 13/145 (9%) due to symptomatic deterioration, 9/145 (6%) due to other, and 1/145 (< 1%) due to protocol violation. The median percentage of per-protocol dose of nab-P was 74%; median nab-P dose intensity and cumulative dose were 74.08 mg/ m2/week and 900 mg/m2, respectively. nab-P dose modifications included 1 reduction, missed dose, or dose delay in 55%, 60%, and 60% of pts, respectively. Grade 3 TEAEs were mainly hematologic and included neutropenia (148/334 [44%]), anemia (92/334 [28%]), and thrombocytopenia (48/334 [14%]). Grade 3 peripheral neuropathy was observed in 14/334 pts (4%). Conclusions: This interim analysis demonstrates the safety and tolerability of nab-P/C induction therapy in pts with SCC histology. The findings are consistent with those reported in subset analysis of the phase III study and provide additional support for the use of this regimen in this pt population. NCT02027428. Clinical trial identification: NCT02027428 Legal entity responsible for the study: Celgene Corporation Funding: Celgene Corporation Disclosure: C. Gridelli: Honoraria: Celgene. D. Morgensztern: Advisory role: Celgene; Speaker bureau: Boehringer. O. Juan: Advisory or Speaker role: Roche, Astra, MSD, Boehringer, BMS, Lilly, Pfizer, Pierre-Fabre. B. Levy: Honoraria: Eli Lilly, Genetech, Astra-Zeneca, Celgene; Consulting or Advisory Role: Eli Lilly, Genetech, Astra-Zeneca, Celgene; Speakers’ Bureau: Eli Lilly, Genetech. A. Ardizzoni: Honoraria: Eli Lilly, BMS, MSD, Boehringer; Consulting Role: Eli Lilly, BMS, MSD, Boehringer. T. Berry: Employment and Stock Ownership: Celgene. T. Chen, N. Trunova: Employment and stock ownership: Celgene. All other authors have declared no conflicts of interest.

1370P

Comparison of platinum agents cisplatin and carboplatin in routine treatment of advanced NSCLC: Results from prospective German TLK cohort study

N. Marschner1, U. von Verschuer2, R. Schnell3, M-O. Zahn4, J. Eggert5, A. Binninger6, L. Spring6, M. J€anicke6 1 Oncology and Hematology, Praxis fu¨r interdisziplin€ are Onkologie & H€ amatologie, amatologie und Onkologie Freiburg, Germany, 2Oncology and Hematology, MVZ fu¨r H€ 3 Essen gGmbH, Essen, Germany, Oncology and Hematology, Praxis Internistischer Onkologie und H€ amatologie (PIOH), Frechen, Germany, 4Oncology and hematology, € ¨ rtliche Berufsausu¨bungsgemeinschaft MVZ Onkologische Kooperation Harz, Ubero Goslar, Germany, 5Oncology and hematology, Gemeinschaftspraxis fu¨r Innere Medizin, Moers, Germany, 6Clinical Epidemiology and Health Economics, IOMEDICO AG, Freiburg, Germany Background: Lung cancer is the leading cause of cancer-related mortality and the majority of patients (pts) are diagnosed with advanced or metastatic disease. Despite advances in targeted therapies for selected patient subgroups, the majority of pts (80%) are treated with platinum-based doublet chemotherapies (CT). The choice between the platinum agents cisplatin (CIS) or carboplatin (CAR) has been subject of a long debate. Here we present data on the treatment of advanced non-small cell lung cancer (aNSCLC) in routine practice. Such real-world data are of central importance to improve the standard of care. Methods: 107 sites in Germany recruited 1,239 pts with aNSCLC at start of 1st-line therapy into the prospective clinical cohort study TLK (Tumour Registry Lung Cancer) between Feb 2010 and Dec 2013. Details on systemic treatment and outcome were collected until Jan 2016. A longitudinal health-related quality of life (HRQOL) analysis using the questionnaires EORTC QLQ-C30 and –LC13 was conducted every 2 months (mts) for a period of up to 10 months. Results: 46% of the pts received CAR- and 35% CIS-based doublet CT in 1st-line treatment. Pts receiving CIS- were younger than pts receiving CAR-combinations (median age at start of treatment 62 vs. 69 years), more often had a good performance status (33% vs. 17% ECOG ¼0) and less comorbidities (34% vs. 56% Charlson Comorbidity Index 1). The main combination partner was pemetrexed for CIS (33%) and paclitaxel for CAR (24%). Median overall survival was 11.9 mts (95% CI 10.2-13.8) for CISand 12.2 mts (95% CI 10.0-13.3) for CAR-combinations. The median time to deterioration of the global health status was 6.8 mts for CIS- and 6.4 mts for CARcombinations. Considerable deteriorations in the symptoms nausea, fatigue, dyspnoea and pain were reported after 4-6 mts, with no difference between CIS and CAR. Conclusions: Numerous meta analyses have been dedicated to finding the optimal platinum agent for the 1st-line treatment of aNSCLC. With our data from the prospective, population-based cohort study TLK, we complement the results from clinical trials. We show that there is no considerable difference in outcome or HRQOL between CISand CAR-combinations in the treatment of aNSCLC. Clinical trial identification: NCT01192919 Legal entity responsible for the study: iOMEDICO AG Funding: Roche Pharma AG, Boehringer Ingelheim Pharma GmbH & Co. KG, Onkovis GmbH, TEVA Deutschland GmbH Disclosure: N. Marschner: Roche Pharma AG: Travel Grant þ Advisory Board Fee. All other authors have declared no conflicts of interest.

490 | NSCLC, metastatic

Annals of Oncology

1371P

Diagnostic and therapeutic strategies for elderly patients with advanced non-small cell lung cancer (NSCLC): Results from an EORTC pan-European survey

M. Giaj Levra1, J. Menis2, A. Luciani3, E. De Maio4, B. Hasan4, T. Berghmans5, M.A. Massiani6, M. De Waele7, A-M. Dingemans8, J. Donckele4, C. Faivre-Finn9, N. Girard10, L. Greillier11, S. Lantue´joul12, M. O’Brien13, M. Reck14, K. Tryfonidis4, H. Wildiers15, B. Besse2, S. Novello16 1 Chest Department, CHU Grenoble - Hopital Michallon, La Tronche, France, 2Cancer Medicine, Institut de Cance´rologie Gustave Roussy, Villejuif, France, 3Medical Oncology, Azienda Ospedaliera S. Paolo, Milan, Italy, 4EORTC, European Organisation for Research and Treatment of Cancer, Brussels, Belgium, 5Oncology, Institute Jules Bordet, Brussels, Belgium, 6Oncology, Institut Curie, Paris, France, 7Oncology, U.Z.A. University Hospital Antwerp, Edegem, Belgium, 8Pulmonology, Maastricht University Medical Center, Maastricht, Netherlands, 9Radiotherapy, The Christie NHS Foundation Trust, Manchester, UK, 10Pneumologie, Hoˆpital Louis Pradel-Hospices Civils de Lyon, Bron, France, 11Pneumologie, Hopital St. Marguerite Assistance Publique Hopitaux de Marseille, Marseille, France, 12Pathology department, CHU Grenoble - Hopital Michallon, La Tronche, France, 13Oncology, Royal Marsden Hospital NHS Foundation Trust, London, UK, 14Thoracic Oncology, Krankenhaus Grosshansdorf, Grosshansdorf, Germany, 15General Medical Oncology, University Hospitals Leuven, Leuven, Belgium, 16 Oncology, University of Turin, Orbassano, Italy Background: The EORTC Lung Cancer Group (LCG) and the Elderly Task Force (ETF) developed a pan-european survey that aims to provide an overview of the management and treatment strategies for elderly patients (pts) diagnosed with advanced NSCLC, as well as to identify potential needs and scientific pending questions that could be addressed in new trials. Methods: An electronic 13-topic survey explaining the study purpose was developed and sent to all EORTC LCG and ETF members. The 25-items included multiple-choice and open-ended questions requesting the following information on general demographics (6 items), pts population (3 items) and diagnostic, treatment preferences and outcomes (4 items). Elderly pts were defined as those older than 70 years. Results: Sixty-two individual sites, from 19 countries, completed the online questionnaire. In 42 centers (67.7%) there is no dedicated team for the management and treatment of elderly pts; on the other hand, only in 2 centers (3.2%) pts with suspected NSCLC are not discussed by a multidisciplinary board. Notably, oncogeriatric assessment is routinely performed in 17 (27.4%) centers; G8, CGA or both scales are the preferred evaluation tools (35.3%, 23.6% and 11.8% respectively). In fit pts, the preferred first-line chemotherapy regimens are Carboplatin (CBDCA)-Pemetrexed (PEM) 19 (31.7%), CBDCA-Paclitaxel (PAC) 15 (24.2%), Cisplatin (CDDP)-PEM 14 (22.6%), CBDCA-Gemcitabine (GEM) 10 (16%), other 3 (4.8%). In the second line setting the preferred treatments are Nivolumab 30 (45.5%), PEM 11 (16.7%), Docetaxel 9 (13.6%), PAC 8 (12.1%), GEM 4 (6.1%), Erlotinib 4 (6.1%); while PEM 15 (24.2%), Nivolumab 13 (20.1%), PAC 9 (14.5%), Docetaxel 6 (9.7%), GEM 6 (9.7%), other 11 (17.7%) represent a second level option. Conclusions: The survey provides an overview of the clinical practice in the management of elderly patients with advanced NSCLC, summarizing relevant and updated background for the possible development of future collaborative trials. In this survey, different treatment regimens are used by different centers, and geriatric assessment is used heterogeneously, reflecting the lack of a “standardized” approach and the need for further research in this area. Legal entity responsible for the study: EORTC Funding: None Disclosure: All authors have declared no conflicts of interest. 1372P

Diversity of brain metastasis (BM) management in non-small cell lung cancer (NSCLC) in Europe (EU): Results of the Young Investigators European Organisation for Research and Treatment of Cancer Lung Cancer Group (YI EORTC LCG) survey

L. Hendriks1, C. Faivre-Finn2, B. Hasan3, E. de Maio4, A. Berghoff5, A-M. Dingemans1, S. Novello6, T. Berghmans7, B. Besse8, A. Levy9 1 Pulmonary Diseases, Maastricht University Medical Center (MUMC), Maastricht, Netherlands, 2Radiotherapy, The Christie NHS Foundation Trust, Manchester, UK, 3 Statistics, EORTC - European Organisation for Research and Treatment of Cancer, Brussels, Belgium, 4Medical oncology, EORTC - European Organisation for Research and Treatment of Cancer, Brussels, Belgium, 5Department of Medicine I, Medical University of Vienna, Vienna, Austria, 6Department of Oncology, University of Turin, Turin, Italy, 7 Oncology, Institute Jules Bordet, Brussels, Belgium, 8Cancer Medicine, Institut de Cance´rologie Gustave Roussy, Villejuif, France, 9Radiation oncology, Institut de Cance´rologie Gustave Roussy, Villejuif, France Background: BM are frequent in NSCLC patients (pts) but management can vary. Methods: An online survey containing questions on NSCLC BM screening and treatment (tx) was widely distributed between 16/02/17 and 16/04/17 to all EORTC LCG members, and through several EU societies involved in lung cancer tx. Results: 478 physicians (phys) (radiation oncologist: 51.9%, pulmonologist: 27%, medical oncologist: 15.3%, others: 5.8%; 73.2% with >5 years experience in NSCLC) responded. Italy [17.8%], Netherlands [14.2%], UK [13.0%], and France [11.5%] contributed most. 84.9% screened neurologically asymptomatic pts for BM at diagnosis (49% used MRI). Phys screened stage III (66.9%) and IV (40.8%) most often. 35.4%

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology used a prognostic (p) classification to guide initial tx decisions. In 48.1% lowest p-score threshold to actively treat pts did not differ between driver mutation (MUTþ) and non-driver (MUT-) pts. 38.1% used less WBRT in poor prognosis pts based on QUARTZ trial (NCT3826061) results. 88.9% had access to stereotactic radiosurgery (SRS). After single BM surgery, 50.8% systematically prescribe adjuvant SRS or WBRT, and 44.2% only in case of incomplete resection. Preferred tx in neurologically asymptomatic tx-naive pts diagnosed with >5 BM was systemic tx (78.4%). 46.9% stated that WBRT could increase systemic tx efficacy. 44.8/49.8% stated that all tyrosine kinase inhibitors (TKI) and immune checkpoint blockers (IO) were discontinued (timing varied) during SRS/WBRT, respectively. Drugs that were most often continued during SRS-WBRT were erlotinib (44.5-40.1%), gefitinib (38.0-34.0%), afatinib (29.9-25.1%), crizotinib (32.2-26.7%) and IO (PD-(L)-1: 28.4-22.8%), CTLA4: 10-9.8%), because of no perceived safety issues (44.6%) or risk of systemic flare (37.9%). MUTþ pts with > 3 BM were more likely to receive SRS than MUT-. 76% of phys preferred local tx & TKI continuation over a switch to next-line tx in pts with only intracranial progression. Conclusions: BM management differs: screening is not uniform, p-classifications are not often used, and MUTþ NSCLC pts generally receive more aggressive local tx. Legal entity responsible for the study: EORTC Lung Cancer Group Funding: None Disclosure: All authors have declared no conflicts of interest.

1373P

Clinical features of never smoker patients with lung squamous cell carcinoma: A retrospective multicenter study

S. Frega1, M. Macerelli2, A. Del Conte3, L. Bonanno1, M. Bartoletti2, V. Polo4, G. Zago1, A. Follador2, I. Attili1, A. Pavan1, L. Urso4, S.M.M. Basso5, G. Fasola2, P.F. Conte1, G. Pasello1 1 Medical and Experimental Oncology Department, Medical Oncology, Istituto Oncologico Veneto IRCCS, Padua, Italy, 2Medical Area Department, AOU Santa Maria della Misericordia, Udine, Italy, 3Medical Oncology Department, Centro di Riferimento Oncologico Aviano, Aviano, Italy, 4Surgical, Oncological and Gastroenterological Sciences Department, University of Padova, Padua, Italy, 5General Surgery and Perioperative Medicine Department, Azienda Ospedaliera Sta Maria degli Angeli, Pordenone, Italy Background: Squamous cell carcinoma of the lung (LSCC) is the second most common histological subtype of non-small cell lung cancer (NSCLC) having smoking habit as the major risk factor. LSCC in non-smokers is an exceptional finding possibly related to professional exposure and subsequent carcinogenesis even though clinical and biological landscape is largely unexplored. Methods: This is a retrospective multicenter study investigating clinical features of never-smoker LSCC patients (pts) referred to three Italian Centers between 2010 and 2016. Relapse (RFS) or progression free (PFS) and overall (OS) survival curves were calculated by Kaplan-Meier method. Cox regression proportional hazards model was used to estimate the impact of covariates on OS. Results: Among 791 LSCC pts, 37(4,6%) occurred in never-smokers; our case series included 19 males and 18 females with a median age of 63 years. ECOG PS was 0-1 in 30(81%) pts. Median Charlson Comorbidity Index (CCI) was 6. Two (5%) pts referred second-hand smoking history and 13(35%) occupational exposure. Additional tumor history was reported by 15(41%) patients: head and neck (N¼4), basocellular skin (N¼5), breast (N¼2), lung (N¼2), prostate (N¼1) cancer and leukemia (N¼1). Molecular characterization was performed in 12(32%)pts: EGFR and KRAS mutations were found in 2 and 1 pts respectively. Median time from symptoms appearance and diagnosis was 7 weeks. Twelve (32%) pts showed a limited stage, while the other 25(68%) showed advanced/metastatic disease at the diagnosis. Nineteen (52%) pts received a first-line palliative chemotherapy (pct), mostly platinum-based doublets plus gemcitabine (N¼11) or taxane (N¼3), achieving a response rate and disease control rate of 37% and 58% respectively. Median RFS in resected patients (N¼9) was 21 months. Median PFS and OS after first-line pct were 5 months and 8.5 months respectively. No covariate significantly impacted on OS. Conclusions: Never-smoker LSCC pts represent a rare subgroup characterized by more females, younger age and a not negligible CCI and second-tumor history compared with the known features of smoker LSCC. Treatment outcome of advanced disease is still dismal as for most LSCC pts. Legal entity responsible for the study: Istituto Oncologico Veneto IRCCS Funding: None Disclosure: All authors have declared no conflicts of interest. 1374P

outcomes of such ineligible patients. Therefore, we investigated the characteristics, outcomes, and survival of advanced NSCLC patients who were ineligible for clinical trials. Methods: We analyzed a retrospective cohort of 786 consecutive patients diagnosed with advanced NSCLC between January 2006 and December 2014. We reviewed the criteria in phase 3 clinical trials, and classified patients using the common first-line eligibility criteria for lung cancer. Results: Of the 786 patients, 469 (60%) were ineligible for clinical trials. The main reasons for ineligibility were brain metastasis (41%), a poor performance status (PS) (25%), and respiratory disease (24%). In all patients, ineligibility was identified as an independent predictor of overall survival (OS) (adjusted hazard ratio [HR] 0.78, 95% confidence interval [CI], 0.65–0.93, P ¼ 0.008), even in patients with a good PS who received chemotherapy (HR 0.80, 95% CI, 0.65–0.99, P ¼ 0.037). In subgroup analyses of ineligible patients, the survival varied depending on the reasons for their ineligibility. In particular, prior cancer history was not associated with a poor outcome, though this was a common reason for ineligibility (14.5%). Conclusions: Most patients were ineligible for clinical trials and had shorter survival. The survival of ineligible patients varied depending on the reasons for their ineligibility. We should consider these results when applying clinical trial outcomes to real-world patients. More studies for ineligible patients are needed to improve real-world treatment. Legal entity responsible for the study: Daichi Fujimoto Funding: None Disclosure: All authors have declared no conflicts of interest.

Characteristics and prognostic impact of advanced non-small-cell lung cancer patients who were ineligible for clinical trials

1375P

Treatment paradigm shift in NSCLC: Patient data analysis from 2005 to 2016

L. Mitrofan1, M. Sahni2, D. Dhingra2, H. Mistry3, C. Anger3 Global Oncology, RWI, QuintilesIMS, Paris, France, 2Global Delivery Center, QuintilesIMS, Gurgaon, India, 3Global Oncoloy, RWI, QuintilesIMS, London, UK

1

Background: In the last decade, chemotherapies were the SoC in advanced NSCLC treatment with limited benefit to long-term survival. Discovery of EGFR/ALK and PD1/PDL1 followed by of approval targeted therapies (TTs) and immunotherapies (IOs), respectively, marked two major treatments shifts. In addition, science advancement on drug resistance issued from TTs and new drug gable mutations continue to transform the treatment landscape and options for patients with advanced NSCLC. Methods: This study used IMS Oncology AnalyzerTM a syndicated, retrospective, longitudinal cancer treatment database collecting anonymized patient-level oncology data in EU5, projected to national level. Data collected between 2005 and 2016 was used to identify changes in the treatment paradigm in advanced NSCLC. Three time period groups have been compared: period 1: from 2005 to 2008; period 2: from 2009 to 2014 and period 3: 2015 and 2016. Results: Of the currently 1,602,026 (projected number) treated populations, there is an increase in protein kinase Inhibitors (TKIs) usage, mainly represented by anti EGFR and anti ALK, from 8% to 23% to 30% in period 1, 2 and period 3 respectively. Monoclonal antibodies (MAb) follows a similar trend increasing from 1% to 15% in the last years, respectively, while the platinum agents slightly decreases. IOs captures 52% in the last couple of years from the overall MAb group. Till recently, bevacizumab (BEVA) was leading this therapeutic class. Increased granularity in patient stratification, will allow identification of more spectacular treatment changes or identification of those who would have passed unnoticed. In 1L, mutant segment, paradigm switch occurred end 2008 when TKIs reached directly 84%. In 2L, IOs jump is much less noticeable, entering directly in the last analyzed period with 25% from MAb group. In WT segment, we can notice 2 switches: one in 1L, end 2014 when BEVA reached directly 13% and a second one in 2L, end 2016, when MAbs captured 28%, with IOs representing 90% from this therapeutic group. Conclusions: Currently, the advent of IOs has completely overshadowed existing TTs. Emerging genetic markers (ROS-1, KRAS, RET), specific EGFR/ALK mutations due to resistance along with combinations of IOs and TTs will continue to add new treatment options. Legal entity responsible for the study: QuintilesIMS Funding: QuintilesIMS Disclosure: All authors have declared no conflicts of interest.

1376TiP

Phase III study of atezolizumab (atezo) vs chemotherapy (chemo) in patients (pts) with treatment-naive advanced, recurrent or metastatic NSCLC unsuitable for platinum (plat)-based chemo

H. Kawachi1, D. Fujimoto1, T. Morimoto2, M. Ito1, S. Teraoka1, Y. Sato1, K. Nagata1, A. Nakagawa1, K. Otsuka1, Y. Imai3, K. Tomii1 1 Respiratory Medicine, Kobe City Medical Center General Hospital, Kobe, Japan, 2Clinical Epidemiology, Hyogo College of Medicine, Nishinomiya, Japan, 3Clinical Pathology, Kobe City Medical Center General Hospital, Kobe, Japan

S.M. Lee1, C. Schulz2, A. Cardona3, P. Bartakova4, S. Peters5 Department of Oncology, UCLH/UCL Cancer Institute/Cancer Research UK Lung Cancer Centre of Excellence, London, UK, 2Department of Internal Medicine II, University Hospital Regensburg, Regensburg, Germany, 3PDMA Biometrics, F. Hoffmann-La Roche Ltd., Basel, Switzerland, 4Global Product Development Medical Affairs (Oncology), F. Hoffmann-La Roche Ltd., Basel, Switzerland, 5Oncology, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland

Background: The majority of non-small cell lung cancer (NSCLC) patients are ineligible for clinical trials. Nevertheless, very few studies report the profiles and treatment

Background: Most pts with newly diagnosed NSCLC have locally advanced or metastatic disease, with 30%-40% having poor performance status (ECOG PS 2) due to

1

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx380 | 491

abstracts disease burden and comorbidities. These pts have poor prognosis vs pts with PS < 2 and higher toxicity with standard doublet chemo. Despite limited efficacy, single-agent chemo (vinorelbine [vin], gemcitabine [gem], docetaxel [doc]) is often used for pts who do not tolerate plat-based regimens, highlighting the need for new treatments. Atezo (antiPD-L1) prevents PD-L1 from binding its receptors PD-1 and B7.1, restoring cancer-specific T-cell immunity. In OAK, a Phase III 2Lþ NSCLC study, atezo monotherapy was well tolerated and showed significant OS improvement vs doc regardless of PD-L1 status. Thus, atezo could provide superior clinical benefit vs singleagent chemo and demonstrate a more favorable safety profile in these pts. The openlabel, randomized, multicenter Phase III PS2 study will evaluate the efficacy and safety of atezo vs single-agent chemo (vin, gem) in pts with untreated advanced, recurrent or metastatic NSCLC unsuitable for plat-based chemo. Trial design: Eligible pts have Stage IIIB/IV NSCLC; are considered unsuitable for platbased chemo due to poor PS (ECOG PS 2-3); have substantial comorbidities or contraindications for plat-based doublet chemo; have measurable disease; have no EGFR/ALK mutations and have received no prior systemic therapy. Pts will be randomized 2:1 to receive atezo 1200 mg IV q3w or single-agent chemo (vin [PO/IV], gem [IV]) per local practice until PD. Pts on atezo may continue therapy until loss of clinical benefit. Stratification factors include histology subtype (squamous vs nonsquamous), PD-L1 status (by VENTANA SP142 IHC assay) and presence of brain metastases. Primary endpoint is OS. Secondary endpoints include OS at 6, 12, 18 and 24 mo; ORR; PFS; DOR and safety. Tumor biopsies and blood samples (pre-treatment all mandatory; at PD only blood mandatory) will be assessed for biomarkers associated with atezo responses and immune escape. Planned enrollment is 441 pts. Clinical trial identification: NCT pending (available on poster) Legal entity responsible for the study: F. Hoffmann-La Roche Ltd. Funding: F. Hoffmann-La Roche Ltd. Disclosure: S.M. Lee: Advisory Board for Roche, Merck, Bristol-Myers, Astra Zeneca. C. Schulz: Scientific Advisor or Membership: AstraZeneca, BMS, Boehringer, Lilly, Novartis, Roche Honoraria: AstraZeneca, Celgene, BMS, Boehringer, Lilly, Novartis, Roche. A. Cardona: Roche employee. P. Bartakova: F. Hoffmann-La Roche employee and company stock ownership. S. Peters: Edu grants, consultation, AB, lectures: Amgen, AZ, BI, BMS, Clovis, Eli Lilly, F. Hoffmann-La Roche, Guardant health, Janssen, Merck Sharp and Dohme, and Merck Serono, Merrimack, Morphotek, Pfizer, Regeneron, Takeda.

1377TiP

JAVELIN Lung 100: updated design of a phase 3 trial of avelumab vs platinum doublet chemotherapy as first-line (1L) treatment for metastatic or recurrent PD-L11 non-small-cell lung cancer (NSCLC)

M. Reck1, C-H. Yang2, P.E. Postmus3, F. Barlesi4, E.F. Font5, M. Thomas6, T.M. Kim7, M. Cobo Dols8, H. Skuladottir9, K. Park10, R. Sullivan11, A.L. Veatch12, N. Pavlakis13, € ¨ ro glu15, F. Morgillo16, M. Schlichting17, F. Teofilovici18, L.M. Dreosti14, M. Ozgu V. Chand19, V. Westeel20 1 Medical Oncology, LungenClinic, Airway Research Center North (ARCN), German Center for Lung Research (DZL), Grosshansdorf, Germany, 2Medical Oncology, National Taiwan University Hospital, Taipei City, Taiwan, 3Medical Oncology, The Clatterbridge Cancer Centre, Liverpool, UK, 4Medical Oncology, Aix Marseille University; Assistance Publique Hoˆpitaux de Marseille, Marseille, France, 5Medical Oncology, Hospital Universitari Vall d’Hebron, Barcelona, Spain, 6Medical Oncology, Thoraxklinik Heidelberg, Heidelberg, Germany, 7Medical Oncology, Seoul National University alaga Hospital, Seoul, Republic of Korea, 8Medical Oncology, Hospital Universitario M Regional y Virgen de la Victoria. IBIMA., M alaga, Spain, 9Medical Oncology, Herning 10 Sygehus, Herning, Denmark, Division of Heamatology/Oncology, Department of Medicine, Samsung Medical Center, Seoul, Republic of Korea, 11Medical Oncology, Auckland Hospital, Auckland, New Zealand, 12Medical Oncology, Northwest Medical Specialties, Puyallup, WA, USA, 13Medical Oncology, Royal North Shore Hospital, Sydney, Australia, 14Medical Oncology, University of Pretoria, Pretoria, South Africa, 15 Medical Oncology, Istanbul University, Istanbul, Turkey, 16Medical Oncology, AOU Seconda Universita degli Studi di Napoli, Naples, Italy, 17Biostatistics, Merck KGaA, Darmstadt, Germany, 18Global Research & Early Development, EMD Serono, Inc., Billerica, USA, 19Medical Oncology, EMD Serono Inc., Billerica, MA, USA, 20Chest Disease, Jean Minjoz University Hospital, Besanc¸on, France Background: Avelumab is a human anti–PD-L1 IgG1 antibody that has shown promising antitumor activity and manageable safety in several tumor types, including NSCLC, and is approved in the US for the treatment of metastatic Merkel cell carcinoma. Reports with anti–PD-(L)1 therapies suggest a role in 1L treatment of NSCLC using tumor PD-L1 expression selection criteria; however, optimal treatment for patients (pts) with different PD-L1 expression status is currently undefined. Preliminary exposure-response analysis from pts with NSCLC treated with avelumab 10 mg/kg Q2W in the JAVELIN Solid Tumor trial (NCT01772004) suggest enhanced antitumor activity with higher avelumab exposure. This phase 3 trial (NCT02576574) compares avelumab vs platinum doublet chemotherapy as 1L treatment for PD-L1þ NSCLC. Here, we describe the updated design, including an additional treatment arm to investigate the role of higher avelumab exposure in this setting. Trial design: JAVELIN Lung 100 is an ongoing global, multicenter, randomized, openlabel trial. The study protocol has been amended to demonstrate superiority of avelumab in prolonging PFS or OS as co-primary endpoints vs platinum doublet therapy, per RECIST v1.1 by blinded independent review. A hierarchical testing strategy will be applied to compare avelumab arms vs chemotherapy in terms of PFS and OS by PD-

492 | NSCLC, metastatic

Annals of Oncology L1þ enriched populations (expression cutoff levels: high, moderate, and any, determined by the Dako 73-10 assay) at an overall significance level of 2.5% (one-sided). Eligibility criteria include: stage IV NSCLC, ECOG PS 1, no prior systemic treatment for advanced disease, and no EGFR mutation/ALK translocation. Approximately 1,095 pts will be randomized to 1 of 3 arms: arm A (avelumab 10 mg/kg 1-hour IV Q2W), arm B (investigator’s choice of specified platinum-based chemotherapy), or arm C (avelumab 10 mg/kg every week for 12 weeks, then 10 mg/kg Q2W), stratified by NSCLC histology and baseline tumor PD-L1 expression level. Secondary endpoints include objective response, duration of response, safety, pt-reported outcomes, PK, and biomarker assessments. Clinical trial identification: NCT02576574 Protocol number: EMR 100070-005 Legal entity responsible for the study: Pfizer Inc., New York, NY, USA and Merck KGaA, Darmstadt, Germany. Funding: Pfizer Inc., New York, NY, USA and Merck KGaA, Darmstadt, Germany. Disclosure: M. Reck: Advisory Role: Roche, Lilly, BMS, MSD, Merck, AstraZeneca, Boehringer Ingelheim, Pfizer, Novartis, Celgene. Speaker’s Bureau: Roche, Lilly, BMS, MSD, Merck, AstraZeneca, Boehringer Ingelheim, Pfizer, Novartis, Celgene. C-H. Yang: Advisory Role: BI, Eli Lilly, Bayer, Roche/Genentech/Chugai, Astellas, MSD, Merck Serono, Pfizer, Novartis, Clovis Oncology, Celgene, Merrimack, Yuhan Pharmaceuticals, BMS, Ono pharmaceutical Daiichi Sankyo and Astrazeneca, Hansoh Pharmaceuticals. P.E. Postmus: Consulting: Celgene, BMS, Roche, Boehringer Ingelheim. Travel, Accomodations, expenses: Boehringer Ingelheim. Speaker’s Bureau: Eli Lilly. F. Barlesi: Honoraria: Astra-Zeneca, Bristol-Myers Squibb, Boehringer– Ingelheim, Clovis Oncology, Eli Lilly Oncology, F. Hoffmann–La Roche Ltd, Novartis, Merck, MSD, Pierre Fabre and Pfizer. E.F. Font: Advisory Role: Eli Lilly, Pfizer, Roche, Boehringer Ingelheim, MSD. Speaker’s Bureau: AstraZeneca, BMS, Novartis. M. Thomas: Grants/Research Support Recipient: Lilly, Pierre Fabre, BMS, AstraZeneca. Advisory Role: Lilly, Pierre Fabre, BMS, AstraZeneca, MSD, Novartis, Celgene, Roche, Pfizer. Honoraria: Lilly, MSD, Celgene, Roche. R. Sullivan: Travel, accommodations, expenses: BMS and Roche. N. Pavlakis: Grants/Research Support: Bayer. Advisory Role: Bayer, BI, AZ, MSD, BMS, Roche, Pfizer, Amgen, Merck Serono. L.M. Dreosti: Grants/Research Support: Novartis. Advisory Role: Novartis, Roche. Honoraria: MSD, € ¨ roglu: Novartis, Janssens. Congress sponsorship: Roche, Merck, Janssens M. Ozgu Advisory Role: Sanofi. M. Schlichting: M.S. is an employee of Merck KGaA, Darmstadt, Germany. F. Teofilovici: F.T. is an employee of EMD Serono Inc. V. Chand: V.C. is an employee of EMD Serono Inc. V.C holds Bristol Myers Squibb stock. V. Westeel: Research funding: Roche, Genentech, BI, Merck Serono. Expert Testimony: Teva. Consulting: Pierre-Fabre Onc, MSD, Eli Lilly, AZ, BI, Novartis. Travel: Pierre-Fabre Onc, Roche, BMS, Eli Lilly, Merck Serono, Pfizer, BI. Speaker’s Bureau: AZ, MSD, BMS, BI. All other authors have declared no conflicts of interest.

1378TiP

A phase 3 study of first-line durvalumab vs platinum-based chemotherapy in patients with advanced NSCLC and high PD-L1 expression: PEARL

Y-L. Wu1, S. Lu2, S. Clarke3, K. Lactionov4, P. Li5, M. Kirkby5, Y. Xie6, P. Stockman7 Guangdong Lung Cancer Institute, Guangdong General Hospital, Guangzhou, China, 2 Shanghai Lung Cancer Center, Shanghai Chest Hospital, Shanghai, China, 3Northern Sydney Cancer Centre, Royal North Shore Hospital, St Leonards, Australia, 4Medical Oncology, N. N. Blokhin Russian Cancer Research Center, Moscow, Russian Federation, 5 Medical Oncology, AstraZeneca, Beijing, China, 6CDU (China Development Unit) B&I, AstraZeneca, Shanghai, China, 7Global Medicines Development, AstraZeneca, Alderley Park, UK

1

Background: NSCLC is highly prevalent in Asia. Differences in outcomes to therapy, including longer survival and higher response rates, have been noted in Asian vs Caucasian NSCLC patients due to the higher prevalence of targetable oncogenes (i.e. EGFR sensitizing mutation or ALK translocation) in the Asian population. Until recently, standard of care (SoC) first-line treatment of EGFR/ALK wild-type advanced NSCLC comprised systemic platinum-based doublets. However, over recent years, immune checkpoint blockade has also become an important option, including in the firstline setting where pembrolizumab is approved in the US, EU and Japan in patients whose tumours express high levels of PD-L1. Durvalumab is a selective, high-affinity, anti-PD-L1 antibody that has shown antitumour activity and manageable tolerability across multiple tumour types, including NSCLC. The PEARL study aims to assess firstline durvalumab vs SoC in predominantly Asian populations with EGFR/ALK wildtype advanced NSCLC and high tumour PD-L1 expression. Trial design: PEARL (NCT03003962) is a randomised, open-label, multicentre Phase 3 study. Eligible patients are immunotherapy- and chemotherapy-naı¨ve with Stage IV, EGFR/ALK wild-type NSCLC, high PD-L1 expression (25% tumour cells with membrane staining using the Ventana PD-L1 [SP263] Assay) and ECOG performance status of 0 or 1. 440 patients will be stratified by level of PD-L1 expression, histology and smoking status, and randomised 1:1 to receive either durvalumab (20 mg/kg i.v. every 4 weeks [q4w]) or SoC platinum-based chemotherapy until disease progression. The coprimary endpoints are PFS using blinded independent central review assessments according to RECIST v1.1 and OS. Secondary endpoints include ORR; duration of response; proportion of patients alive and progression free at 12 months; PFS after subsequent anticancer therapy; disease-related symptoms and HRQoL; immunogenicity; safety (CTCAE v4.03) and tolerability. Tumour assessments will be performed q6w for the first 48 weeks and then q8w as defined by RECIST v1.1 until confirmed

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology disease progression. Recruitment is ongoing in Australia, China, Republic of Korea, Russia, Thailand and Vietnam. Clinical trial identification: NCT03003962 (release date: December 15, 2016) Legal entity responsible for the study: AstraZeneca PLC Funding: AstraZeneca Disclosure: Y-L. Wu: Speaker fees: Roche, AstraZeneca, Eli Lilly, Sanofi, Pfizer. S. Lu: Speakers’ Bureau: AstraZeneca, Eli Lilly, Roche, Sanofi Corporate sponsored research: AstraZeneca, Boehringer-Ingelheim, Hutchison, Roche Consultant: AstraZeneca, Boehringer-Ingelheim, Hutchison, MediPharma, Roche. S. Clarke: Corporate sponsored research: Ipsen, Merck Honoraria: AstraZeneca, Ipsen, Merck. K. Lactionov: Consultant BMS, AstraZeneca, MSD, Pfizer and travel grants BMS, AstraZeneca, MSD, Pfizer P. Li, M. Kirkby, P. Stockman: AstraZeneca: full-time employment and stock ownership. Y. Xie: AstraZeneca: full-time employment.

1379TiP

Phase I/II study of S 49076, a MET/AXL/FGFR inhibitor, in combination with gefitinib in EGFR-mutated NSCLC patients who progress on EGFR tyrosine kinase inhibitor

K. Park1, F. Ciardiello2, T. Hida3, W. Lim4, C-C. Lin5, H. Murakami6, M. Nishio7, F. Cantero8, V. Cattan8, C. Gabarroca9, E. Gandossi8, L. Paz-Ares10 1 Division of Heamatology/Oncology, Department of Medicine, Samsung Medical Center, Seoul, Republic of Korea, 2U.O.C. di Oncologia Medica e di Ematologia, Universita degli Studi della Campania “Luigi Vanvitelli”, Naples, Italy, 3Department of Thoracic Oncology, Aichi Cancer Center, Nagoya, Japan, 4Clinical Trial and Epidemiological Sciences, National Cancer Centre Singapore, Singapore, 5Department of oncology, National Taiwan University Hospital, Taipei, Taiwan, 6Division of Thoracic Oncology, Shizuoka Cancer Center, Shizuoka, Japan, 7Department of Thoracic Medical Oncology, The Cancer Institute Hospital of JFCR, Tokyo, Japan, 8Department of Oncology R&D, Servier, Suresnes, France, 9Department of Oncology Biostatistics, Servier, Suresnes, France, 10Medical Oncology, Hospital Universitario 12 de Octubre, Madrid, Spain Background: EGFR-mutated NSCLC patients treated with EGFR TKI ultimately develop resistance to therapy. Mechanisms of resistance include aberrant signalling of MET, AXL and FGFR. Inhibition of these receptors should thus abrogate their downstream signalling and restore sensitivity to EGFR TKI. S 49076 is a potent ATPcompetitive TKI, which targets MET, AXL and FGFR1/2/3 at clinically relevant doses. It showed strong activity in preclinical assays against EGFR mutated MET amplified cell lines to overcome the resistance to EGFR TKI. This study evaluates the relevance of targeting MET and/or AXL dysregulation to overcome acquired non EGFR-T790Mmediated resistance to EGFR-TKI. Trial design: The phase I/II is an international, open-label study to evaluate the safety and activity of S 49076 in combination with gefitinib in stage IIIb-IV NSCLC patients who progress on EGFR TKI (1st and 2nd generation), harbouring MET and/or AXL dysregulation without EGFR-T790M mutation. The phase I is a single arm dose-finding part of S 49076 in combination with standard dose of gefitinib. The primary objective is to determine the recommended phase 2 dose (RP2D) based on the dose-limiting toxicities and safety assessments. The secondary objectives are to evaluate the pharmacokinetic profile of both drugs and the anti-tumour activity of the combination. S 49076 and gefitinib are administered orally once daily over a continuous 28-day cycle with doses ranging from 500mg to 600mg. Dose levels are allocated using a modified Bayesian Continual Reassessment Method. This study part is active and recruiting over 5 countries across Asia and Europe. The end of phase I is expected in Q3 2017. A nonrandomised and non-comparative phase II part will then evaluate the anti-tumour activity of S 49076 at the RP2D in combination with gefitinib in cohorts of patients with MET and/or AXL dysregulation. The primary objective will be to determine the objective response rate according to RECIST and the secondary objectives will be to evaluate survival rate, progression free survival, clinical benefit rate and response duration as well as safety. An interim analysis will be performed in each cohort for futility. Clinical trial identification: EudraCT 2015-002646-31 Legal entity responsible for the study: Institut de Recheches Internationales Servier Funding: Institut de Recherches Internationales Servier Disclosure: K. Park: Servier study international coordinator and investigator. F. Ciardiello, W. Lim, C-C. Lin: Servier study national coordinator and investigator. T. Hida, H. Murakami, M. Nishio: Servier study investigator. F. Cantero, V. Cattan, C. Gabarroca, E. Gandossi: Servier employee. L. Paz-Ares: Study Servier national coordinator and investigator.

Volume 28 | Supplement 5 | September 2017

1380TiP

A randomized, open-label comparison of lorlatinib versus crizotinib as first-line treatment for advanced anaplastic lymphoma kinase (ALK)-positive non-small cell lung cancer

A.T. Shaw1, T.M. Bauer2, T. Takahashi3, C.S. Baik4, A. Polli5, M. Carpentieri6, J-F. Martini7, B.J. Solomon8 1 Cancer Center, Massachusetts General Hospital, Boston, MA, USA, 2Medical Oncology, Sarah Cannon Research Institute, Nashville, TN, USA, 3Division of Thoracic Oncology, Shizuoka Cancer Center, Shizuoka, Japan, 4Medical Oncology, University of Washington/Seattle Cancer Care Alliance, Seattle, WA, USA, 5Statistics, Pﬁzer Oncology, Milan, Italy, 6Global Oncology Research and Development, Pﬁzer Oncology, Milan, Italy, 7 Translational Oncology, Pﬁzer Oncology, La Jolla, CA, USA, 8Medical Oncology, Peter MacCallum Cancer Center, Melbourne, Australia Background: Lorlatinib and crizotinib are oral tyrosine kinase inhibitors with activity against ALK and ROS1 fusion proteins. Crizotinib is well tolerated and has superior efficacy compared to chemotherapy for treatment of patients (pts) with advanced ALKþ NSCLC. However, resistance to crizotinib can develop, and the central nervous system (CNS) is often a site of disease relapse. Lorlatinib is a CNS penetrant and has potent activity against de novo fusions and kinase domain resistance mutations. Lorlatinib has shown clinical activity in pts previously treated with crizotinib and other ALK inhibitors, including pts with progressive CNS metastases. This study aimed to determine if lorlatinib is superior to crizotinib in prolonging progression-free survival (PFS) in treatment-naı¨ve pts with advanced ALKþ NSCLC and to identify candidate biomarkers predictive of clinical efficacy or treatment resistance. Trial design: This global, multicenter, open-label phase 3 study will enroll 280 treatment-naı¨ve pts. Eligible pts must be aged 18 years, have Eastern Cooperative Oncology Group performance status of 0–2 and 1 measurable extracranial target lesion not previously treated with radiotherapy. Pts with asymptomatic brain metastases are eligible. Pts will be randomized (1:1) to lorlatinib 100 mg once daily or crizotinib 250 mg twice daily and stratified by presence of brain metastases (yes/no) and ethnicity (Asian/non-Asian). Treatment will continue until disease progression, pt refusal, or unacceptable toxicity. Crossover between treatment arms will not be permitted. The primary endpoint is PFS based on blinded independent central review (BICR) using RECIST v1.1. Secondary endpoints include PFS based on investigator assessment (IA), overall survival, objective response (OR) by BICR and IA, intracranial (IC) OR, IC time to progression, duration of response, time to response by BICR, tumor tissue and peripheral blood circulating free DNA biomarker assessment, safety, and pt-reported health-related outcomes. The first pt was screened on April 14, 2017. This study is registered with ClinicalTrials.gov as NCT03052608. Clinical trial identification: NCT03052608 Legal entity responsible for the study: Pfizer Funding: Pfizer Disclosure: A.T. Shaw: Membership of an advisory board or board of directors Blueprint medicines, KSQ therapeutics. Honoraria or Consulting - Pfizer, Novartis, Ariad, Genentech/Roche, Ignyta, Daiichi-sankyo, Taiho, LOXO, Blueprint medicines, EMD Serono, Foundation Medicine. T. Takahashi: Corporate sponsored research AstraZeneca, Pfizer, Eli Lilly, Chugai Pharmaceutical Co, Ono Pharmaceutical Other, please specify; Honoraria - AstraZeneca, Pfizer, Eli Lilly, Chugai Pharmaceutical, Ono Pharmaceutical. C.S. Baik: University of Washington: Pfizer; Novartis, Loxo Oncology, Genentech, MedImmune, Mirati Therapeutics, Clovis Oncology, GlaxoSmithKline, Eisai, Celgene, Bristol-Myers Squibb, Merck Sharp & Dohme Corp. Clovis Oncology and Novartis. A. Polli: Stock ownership - Pfizer. M. Carpentieri: Stock ownership Pfizer Other relationships (such as employment) with a pharmaceutical company Pfizer. J-F. Martini: Stock ownership - Pfizer Other relationships (such as employment) with a pharmaceutical company - Employee (Pfizer). B.J. Solomon: Membership of an advisory board or board of directors - Advisory Boards: Pfizer, Novartis, RocheGenentech, AstraZeneca, Merck, Bristol Myers Squibb. All other authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx380 | 493

abstracts 1381TiP

Annals of Oncology

A randomized double-blind phase II trial evaluating maintenance PARP inhibitor Olaparib versus placebo in patients with platinumsensitive advanced non-small cell lung cancer: PIPSeN trial

S. Postel-Vinay1, D. Planchard2, A.L. Ortega Granados3, A. Gazzah1, M.A. Sala Gonzalez4, M. Majem5, C.J. Camps6, A. Abou-Lauvergne7, J-P. Pignon8, J. Cadranel9, J. Bennouna10, F. Barlesi11, M.R. Garcia Campelo12, S. Viteri13, B. Besse14, J. Coves Sarto15, B. Massuti Sureda16, J-C. Soria2, R. Rosell17 1 De´partemement d’Innovation The´rapeutique et des Essais Pre´coces (DITEP), Institut Gustave Roussy, Villejuif, France, 2Medical Oncology Department, Institut Gustave Roussy, Villejuif, France, 3Oncologıa Me´dica, Complejo Hospitalario de Jae´n, Jae´n, Spain, 4Medical Oncology Department, Hospital de Basurto, Bilbao, Spain, 5Medical oncology Department, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, 6Medical Oncology Department, Hospital General Universitario Valencia, Valencia, Spain, 7 Recherche clinique, Institut Gustave Roussy, Villejuif, France, 8De´partement de Biostatistiques, Institut Gustave Roussy, Villejuif, France, 9Medical Oncology Department, APHP, CancerEst, Tenon University Hospital, Paris, France, 10Oncology, CHU de Nantes, Nantes, France, 11Multidisciplinary Oncology & Therapeutic Innovations, Aix Marseille University - Assistance Publique des Hoˆpitaux de Marseille, Marseille, France, 12Medical Oncology, Complejo Hospitalario Universitario de A Coru~ na, A Coru~ na, Spain, 13Oncology department, Instituto Universitario USP Dexeus, Barcelona, Spain, 14Cancer Medicine, Institut de Cance´rologie Gustave Roussy, Villejuif, France, 15Medical Oncology, Hospital Son Llatzer, Palma de Mallorca, Spain, 16Medical Oncology, Hospital General Universitario de Alicante, Alicante, Spain, 17Medical Oncology Department, Catalan Institute of Oncology (ICO Badalona), Hospital Germans Trias i Pujol, Badalona, Spain Background: Poly (ADP-ribose) polymerase inhibitors (PARPi) have demonstrated impressive efficacy in BRCA-mutated gynaecological malignancies. Several lines of evidence now support that the DNA repair (DDR)-deficient populations that benefit from PARPi go far beyond BRCA-deficiency. Non-small cell lung cancer (NSCLC), the first cause of cancer death worldwide, displays frequent DDR defects, the most frequent being ERCC1. This defect leads to platinum and PARPi sensitivity. Beyond ERCC1, DDR defects leading to platinum sensitivity widely overlap with those underlying PARPi sensitivity. Maintenance PARPi could therefore benefit to patients (pts) with V platinum-sensitive NSCLC. Olaparib (Lynparza , Astra Zeneca), a potent and selective PARPi, was the first-in-class approved PARPi in BRCA-mutated ovarian cancer. Trial design: PIPSeN is a randomized double-blind phase II investigator-initiated study evaluating maintenance Olaparib versus placebo in pts with platinum-sensitive advanced NSCLC. Chemonaı¨ve ECOG PS 0-1 pts with stage III-IV NSCLC with no EGFR mutation or ALK translocation are eligible. Treatment consists of an “induction phase” of 4-6 cycles platinum-based therapy (any doublet), followed by a “randomized phase” where pts presenting with partial or complete response are randomized between Olaparib maintenance (tablets; 300mg bd) and placebo until progression or unacceptable toxicity. Primary objective is to assess the efficacy of maintenance Olaparib as measured by Progression-Free Survival from randomisation (RECIST v1.1). Secondary objectives include comparison of overall survival, disease control rate and safety. Randomization is stratified according to age, histology and country. With an anticipated HR ¼ 0.65 (bilateral a ¼ 0.2; b ¼ 0.2), approximately 500 enrolled pts will be required to randomize 144 pts and observe 97 events. Recruitment is ongoing since the 5th of Feb. 2016 across 21 centres in France and Spain; 95 (19) pts have been enrolled (randomised) to date. Translational studies looking notably for biomarkers of platinum and PARPi sensitivity (using WES, RNAseq, proteomics and ctDNA) are associated. Clinical trial identification: EudraCT: 2014-005586-75 NCT02679963 Legal entity responsible for the study: Gustave Roussy Cancer Campus (sponsor), in collaboration with the Spanish Lung Cancer Group Funding: Astra Zeneca Disclosure: D. Planchard: Consultancy fees for AstraZeneca, Boehringer Ingelheim, BMS, Lilly, MSD, Pfizer, Roche, Novartis, Chugai. M. Majem: Merck Sharp and Dhome, Boerhinger Ingelheim, Bristol-Myers Squibb, Roche, Novartis honoraria. F. Barlesi: AstraZeneca Honoraria. S. Viteri: Consulting/advisory (BI, Clovis, Idea Pharma, Novartis, Roche, Targovax), Research (AbbVie, ARIAD, Astex, AstraZeneca/MedImmune, BI, Clovis, CytRx, Daiichi Sankyo, GSK, Hanmi, Incyte, Merck, Novartis, Pfizer, Puma, Roche, Servier, Vaxon) B. Besse: Research grants from Astra Zeneca. J-C. Soria: Consultancy fees from AZ. All other authors have declared no conflicts of interest. R

1382TiP

Investigation of biomarkers in patients with adenocarcinoma of the lung receiving nintedanib according to approved label: Noninterventional LUME-BioNIS study

M. Reck1, A.J. Staal-van den Brekel2, A. Mellemgaard3, N. Morsli4, K. Pietzko5, T. Kitzing5, J. Braunger5, K.M. Kerr6 1 Department of Thoracic Oncology, Lung Clinic Grosshansdorf, Airway Research Center North (ARCN), Member of the German Center for Lung Research (DZL), Grosshansdorf, Germany, 2Department of Pulmonary Diseases, Twente Hospitals, Almelo, Netherlands, 3 Department of Oncology, Herlev University Hospital, Herlev, Denmark, 4-, Boehringer Ingelheim France S.A.S., Paris, France, 5Boehringer Ingelheim GmbH & Co. KG, Biberach an der Riss, Biberach, Germany, 6Department of Pathology, Aberdeen Royal Inﬁrmary, Aberdeen, UK Background: In the Phase 3 LUME-Lung 1 trial, the addition of nintedanib to docetaxel significantly improved overall survival (OS) vs docetaxel alone in patients with

494 | NSCLC, metastatic

adenocarcinoma non-small cell lung cancer (NSCLC) treated with one prior line of chemotherapy. No tumour or serum biomarkers are validated to predict nintedanib efficacy in this setting. Trial design: LUME-BioNIS (NCT02671422) is an ongoing, prospective, European, multicentre, non-interventional study (N300) investigating whether tumour-based genomic or proteomic alterations (6 clinical covariates) can predict OS in adults with advanced adenocarcinoma NSCLC initiating nintedanib þ docetaxel according to the nintedanib label. Tumour tissue obtained before first-line therapy will be used for biomarker analyses. To ensure sample quality and tumour content across all slides, the first, middle and last slides will be haematoxylin/eosin-stained and assessed by a certified pathologist for tumour content, extent of necrosis and immune cell infiltration. V Tumour DNA and RNA will be co-isolated from unstained slides (AllPrep DNA/RNA V V FFPE Kit) and quantitated with PicoGreen and RiboGreen reagents, respectively. Tumour DNA sequencing libraries will be prepared using a capture-based targeted gene panel covering whole exons of NSCLC-related genes (e.g. EGFR, KRAS, ALK, BRAF, PIK3CA, TP53) and nintedanib target genes (VEGFR1–3, FGFR1–3, PDFGR a/ V b) and analysed by Illumina next-generation sequencing (NGS). Transcriptomics analyses will be conducted to: (1) complement DNA analyses by providing further information on gene fusions; and (2) enable tumour classification into transcriptional subtypes. Tumour RNA libraries will be prepared and analysed by NGS or, if RNA quantity/quality is insufficient for sequencing, digital gene expression analysis V (nCounter Gene Expression Panels) will be performed. Unstained slides will also be analysed by immunohistochemistry for immune- and proliferation-related protein expression (PD-L1, Ki-67). The primary endpoint is OS, which will be analysed according to biomarker status. Clinical trial identification: NCT02671422 Legal entity responsible for the study: Boehringer Ingelheim Pharma GmbH & Co. KG Funding: Boehringer Ingelheim Pharma GmbH & Co. KG Disclosure: M. Reck: Author reports personal fees from Boehringer-Ingelheim, Hoffmann-La Roche, Lilly, MSD, BMS, AstraZeneca, Celgene, Merck and Pfizer. N. Morsli, K. Pietzko, T. Kitzing, J. Braunger: Author is an employee of BoehringerIngelheim. K.M. Kerr: Personal fees from Boehringer Ingelheim, during the conduct of the study. All other authors have declared no conflicts of interest. R

R

R

R

R

1383TiP

Blood first line ready screening trial (B-F1RST) and blood first assay screening trial (BFAST) enable clinical development of novel bloodbased biomarker assays for tumor mutational burden (TMB) and somatic mutations in 1L advanced or metastatic NSCLC

T.S.K. Mok1, S. Gadgeel2, E.S. Kim3, V. Velcheti4, S. Hu5, T. Riehl6, E. Schleifman7, S.M. Paul8, S. Mocci9, D.S. Shames7, S. Phan10, C. Yun11, M. Mathisen12, M. Kowanetz13, U. Sweere14, M.A. Socinski15 1 Clinical Oncology, Chinese University of Hong Kong, Hong Kong, China, 2Medical Oncoloy, University of Michigan, Ann Arbor, MI, USA, 3Medical Oncoloy, Levine Cancer Institute, Carolinas Health Care System, Charlotte, USA, 4Hematology and Oncology, Taussig Cancer Institute, Cleveland Clinic, Cleveland, OH, USA, 5Biostatistics, Genentech, Inc., South San Francisco, CA, USA, 6Product Development Clinical (Oncology), Genentech, South San Francisco, CA, USA, 7Oncology Biology Development, Genentech, South San Francisco, CA, USA, 8Oncology Biomarker Development, Genentech, South San Francisco, CA, USA, 9Product Development Oncology, Genentech, Inc., South San Francisco, CA, USA, 10USMA, Genentech, South San Francisco, CA, USA, 11USMA, Genentech, Inc., South San Francisco, CA, USA, 12Oncology Biomarker Development, Genentech, Inc., South San Francisco, CA, USA, 13Oncology Biology Development, Genentech, South San Francisco, CA, USA, 14PDG, F. Hoffmann-La Roche Ltd, Basel, Switzerland, 15Thoracic Oncology Program, Florida Hospital Cancer Institute, Orlando, FL, USA Background: Several clinical trials have confirmed the safety and efficacy of atezolizumab (atezo; anti–PD-L1) monotherapy in advanced NSCLC, including in PD-L1–selected 1L patients (pts). Independent of PD-L1 status, high TMB is associated with atezo efficacy. Alectinib is a potent, selective ALK/RET kinase inhibitor currently approved for NSCLC pts previously treated with crizotinib and is expected to have activity in 1L NSCLC pts with ALK or RET alterations. Currently, molecular diagnostics require tumor biopsies which can be difficult to obtain. Here we present trials in progress that aim to clinically evaluate and prospectively validate novel blood-based diagnostic assays that measure TMB in the blood (bTMB) and somatic mutations (e.g., ALK/RET), and to determine the efficacy and safety of 1L atezo or alectinib in NSCLC pts. Trial design: B-F1RST (NCT02848651) is a single-arm study to evaluate the efficacy and safety of atezo and the association between bTMB and efficacy in biomarkerunselected pts. BFAST is a screening and interventional umbrella trial for pts selected based on bTMB or somatic mutations. Eligible pts must have previously untreated, stage IIIB-IVB NSCLC of any histology and measurable disease per RECIST v1.1. Pts will continue treatment until disease progression (all arms) or loss of clinical benefit (atezo only). In B-F1RST, mandatory blood samples will be prospectively collected and retrospectively tested for bTMB. In BFAST, pre-enrollment screening will identify pts who harbor oncogenic somatic mutations (ALK/RET) or are bTMB þ (above a prespecified cutoff); pts will be assigned to the appropriate cohort based on screening results. Study treatments and key endpoints are shown in the table. Additional BFAST cohorts may be added in the future to address other somatic mutations.

Volume 28 | Supplement 5 | September 2017

abstracts

Annals of Oncology

Table: 1383TiP B-F1RST and BFAST Study Details Study

Treatment

Planned Enrollment, n

Primary Endpoints

Key Secondary Endpoints

B-F1RST Phase II

Atezo 1200 mg IV q3w

150

ORR per RECIST v1.1 (INV-assessed) for the efﬁcacy objective Relationship between PFS per RECIST v1.1 and various bTMB quantiles for the biomarker objective

PFS and DOR per RECIST v1.1 (INVassessed) OS

BFAST Phase II/III Cohort A ALKþ

Alectinib 600 mg PO bid

78

ORR per RECIST v1.1 (INV-assessed)

Cohort B RETþ

Alectinib 900 mg & 1200 mg dose escalation

52-62

ORR per RECIST v1.1 (INV-assessed)

Cohort C bTMBþ

Atezo 1200 mg IV q3w or platinum-based chemotherapya

440 (R, 1:1)

PFS per RECIST v1.1 (INV-assessed)

DOR, CBR and PFS per RECIST v1.1 (INV-assessed) ORR, DOR, CBR and PFS per RECIST v1.1 (IRF-assessed) OS DOR, CBR and PFS per RECIST v1.1 (INV-assessed) ORR, DOR, CBR and PFS per RECIST v1.1 (IRF-assessed) OS OS PFS, ORR and DOR per RECIST v1.1 (IRF-assessed) ORR and DOR per RECIST v1.1 (INV-assessed) 6and 12-month PFS rates

a

Cisplatin or carboplatin þ pemetrexed for non-squamous histology, and cisplatin or carboplatin þ gemcitabine for squamous histology. Administered per standard of care. INV, investigator; IRF, independent review facility; R, randomized; b TMB, blood Tumor Mutational Burden.

Clinical trial identification: NCT02848651, B-FAST NCT number available on poster Legal entity responsible for the study: F. Hoffmann-La Roche Ltd. Funding: F. Hoffmann-La Roche Ltd. Disclosure: T.S.K. Mok: Spt: AZ, BI, PFE, NV, SFJ, ROG, MSD, CLVS, BMS, Eisai, Taiho ROG/GNE, LLY, NV; Stock Sanomics AB: AZ, ROG/GNE, PFE, LLY, BI, CLVS, MSD, NV, SFJ, ACEA Bio, VRTX, BMS, GeneDecode, OGX, CELG, RXDX Bod: IASLC, CLCRF, CSCO, HKCTS. S. Gadgeel: Speaker’s bureau- Astra-Zeneca, Genentech/Roche Advisory Boards- Astra-Zeneca, Ariad, Pfizer, Bristol Myers- Squibb and Genentech/Roche. E.S. Kim: Consulting or advisory role: AstraZeneca, Celgene, BI, Eli Lily. V. Velcheti: Consulting/advisory role: Genentech, BMS, Merck, AstraZeneca, Foundation Medicine, Genoptix, Amgen. S. Hu, D.S. Shames: Genentech employee and Roche stock. T. Riehl, E. Schleifman, S. Mocci, S. Phan, M. Mathisen: Genentech employee. S.M. Paul: Genentech employee, Roche Stock, Travel, Accommodations, Expenses from Genentech. C. Yun: Employee of Genentech, Roche stock, Research funding from Genentech. M. Kowanetz: Genentech employee & Roche stock. U. Sweere: Roche employee. M.A. Socinski: Honoraria: Genentech Speakers Bureau: Genentech Research Funding: Genentech.

1384TiP

Prospective comparison of liquid biopsy to standard of care tissue testing in metastatic, non-squamous, non-small cell lung cancer (NSCLC) patients (pts)

Taus2, M. Majem Tarruella3, S. Viteri4, E. Carcereny Costa5, J. Garde6,7, R. Palmero1, A. E. Felip Font8, C. Cassidy9, D. Dix9, N. Karachaliou4, R. Rosell4,5 1 Catalan Institute of Oncology, L’Hospitalet, Spain, 2Del Mar Hospital, Barcelona, Spain, 3 Santa Creu i Sant Pau Hospital, Barcelona, Spain, 4Institute of Oncology Dr. Rosell (Dexeus University Hospital QuironSalud Group), Barcelona, Spain, 5Catalan Institute of Oncology, Germans Trias i Pujol Health Science Institute and Hospital, Badalona, Spain, 6 Arnau de Vilanova University Hospital, Valencia, Spain, 7Medica Scientia Innovation on University Hospital, Barcelona, Research-MEDSIR ARO, Barcelona, Spain, 8Vall d’Hebr Spain, 9Guardant Health, Inc., Redwood City, CA, USA, 10Oncology, Catalan Institute of Oncology, Germans Trias i Pujol Health Sciences Institute and Hospital, Badalona, Spain

non-inferior to tissue-based genotyping in detecting clinically actionable tumor biomarkers in pts with newly diagnosed metastatic non-squamous NSCLC. Trial design: This is a multi-center, prospective, single-cohort study. Major inclusion criteria are: (1) metastatic, biopsy proven, non-squamous NSCLC, (2) candidate for first line systematic therapy, (3) no prior targeted therapy. Peripheral blood (20 mL) is collected for cfDNA sequencing prior to and 2 weeks after treatment initiation and at the time of disease progression (maximum 12 months follow-up). cfDNA sequencing is performed using Guardant360, a comprehensive next-generation assay (Guardant Health, Inc., Redwood City, CA USA). Pre-treatment tissue samples archived at local sites are centrally analyzed. Pts are treated according to investigator standard of care criteria. Tumor response is assessed centrally per RECIST v1.1 on pts who received targeted therapy. Primary endpoint is the detection rate, in either blood or tissue, of a clinically actionable somatic biomarker, defined as mutations in EGFR, BRAF, MET and ERBB2, copy number of MET and rearrangement of the ROS1, RET and ALK genes. A total of 182 pts are needed to test a 10% non-inferiority margin. We assumed a 20% detection rate, with a 19% discordant pairs and 10% dropout rate. The one-sided asymptotic test has 90% power, at a nominal significance level of 5%. Secondary objectives are to compare turn-around time, time to treatment initiation, rates of insufficient tissue for testing or tumor not detected in cfDNA, and tumor response to targeted therapies. Genomically acquired resistance to targeted therapies is also investigated. Clinical trial identification: MedOPP125 (NCT number in progress) Legal entity responsible for the study: Medica Scientia Innovation Research-MEDSIR Funding: Guardant Health Inc. Disclosure: S. Viteri: Research: AbbVie, ARIAD, Astex, AZ/MedImmune, Boehringer, Clovis, CytRx, Daiichi Sankyo, GSK, Hanmi, Incyte, Merck, Novartis, Pfizer, Puma, Roche, Servier, Vaxon. Advisor: Boehringer, Clovis, Idea Pharma, Novartis, Promega Biotech, Roche, Targovax. E. Felip Font: Personal fees (Consulting fees) from: Boehringer Ingelheim, Eli Lilly, Pfizer, Roche and MSD, Astra Zeneca and Bristol Myers Squibb. All other authors have declared no conflicts of interest.

Background: Targeted therapy improves clinical outcomes in pts with advanced NSCLC harboring specific genetic alterations. Guidelines recommend tissue-based assessment of markers, but this approach may be limited by access to sufficient tissue and tumor heterogeneity, making it difficult to identify all pts who may benefit from these treatments. An alternative approach is to assess somatic mutations in circulating cell-free tumor DNA (cfDNA). The aim of this study is to demonstrate that cfDNA is

Volume 28 | Supplement 5 | September 2017

doi:10.1093/annonc/mdx380 | 495

abstracts 1385TiP

Clinical Research platform Into molecular testing, treatment and outcome of non-Small cell lung carcinoma Patients (CRISP): a prospective German Registry in stage IV NSCLC AIO-TRK-0315)

F. Griesinger1, W. Eberhardt2, N. Marschner3, M. J€anicke4, A. Fleitz4, L. Spring4, J. Sahlmann5, A. Karatas6, A. Hipper6, W. Weichert7, M. Sebastian8, M. Thomas9 1 Department of Internal Medicine-Oncology, Pius Hospital Oldenburg, University of Oldenburg, Oldenburg, Germany, 2Department of Medical Oncology, University Hospital Essen Westdeutsches Tumorzentrum, Essen, Germany, 3Oncology and Haematology, Praxis fu¨r interdisziplin€ are Onkologie & H€ amatologie, Freiburg, Germany, 4 Clinical Epidemiology and Health Economics, IOMEDICO AG, Freiburg, Germany, 5 Data Management, Statistics & Medical Informatics, IOMEDICO AG, Freiburg, Germany, 6AIO-Studien-gGmbH, AIO-Studien-gGmbH, Berlin, Germany, 7Institute of Pathology, Technical University Munich (TUM), Munich, Germany, 8Department of Oncology and Haematology, Universit€ atsklinikum Frankfurt(Johannes-Wolfgang Goethe Institute), Frankfurt am Main, Germany, 9Medical Oncology, Thoraxklinik Heidelberg, Heidelberg, Germany Background: Treatment in NSCLC is quickly evolving and new agents make it to the routine practice at a rapid pace. Whether outcome and PRO data generated from clinical trials with often narrow inclusion and exclusion criteria will hold up in the routine practice is of high interest, especially due to the increasing costs of new drugs. Therefore registry data are of ever increasing importance to patients, physicians and reimbursement institutions. Trial design: We have started a prospective, clinical registry to document representative data on molecular testing, sequences of systemic therapies and other treatment modalities, and course of disease in patients with metastatic NSCLC in Germany (CRISP, NCT02622581). A particular focus is on molecular biomarker testing of patients before the start of first-line treatment. The data shall be used to assess the current state of care and to develop recommendations concerning topics that could be

496 | NSCLC, metastatic

Annals of Oncology improved. PRO assessment will provide large-scale data on quality of life and anxiety/ depression for real-life patients in routine practice. In addition, two questionnaires (concerning individual quality of life and patient-caregiver communication) will be validated in German patients with metastatic NSCLC. Furthermore CRISP will set up a decentral tissue annotation for future collaborative, investigational scientific biomarker testing. CRISP will be carried out in up to 150 representative cancer centers in all therapeutic sectors in Germany. More than 5000 patients will be recruited and followed up until death or for a maximum of 3 years. The first patient has been included in December 2015. Currently, 104 centers have been initiated, and 765 patients have been recruited. Preliminary data will be presented at the meeting in terms of molecular test rates, demographic data as well as treatment stratification in the 1st line setting. In conclusion: The registry CRISP will be the first to present representative real life data, covering all treatment settings of patients with NSCLC in Germany. CRISP is supported by AstraZeneca, Boehringer Ingelheim, BMS, Celgene, Lilly, MSD, Novartis, and Pfizer. Clinical trial identification: NCT02622581 Legal entity responsible for the study: AIO-Studien-gGmbH, Berlin Funding: AstraZeneca GmbH, Boehringer Ingelheim Pharma GmbH & Co. KG, Bristol-Myers Squibb GmbH & Co. KGaA, Celgene GmbH, MSD Sharp & Dohme GmbH, Lilly Deutschland GmbH, Novartis Pharma GmbH, and Pfizer Pharma GmbH. Disclosure: F. Griesinger: Advisory Board/Honoraria: Ariad, Astra-Zeneca, Boehringer-Ingelheim, Bristol-Myer-Squibb, Celgene, Clovis, Lilly, Merck-SharpDome, Novartis, Pfizer, Roche N. Marschner: ADB: Amgen, Roche Honoraria: Amgen, Celgene, Roche research grants: Amgen, Celgene, stock ownweship/leadership position: iOMEDICO AG. M. Sebastian: Advisory boards: BMS, MSD, Roche, Novartis, AstraZeneca, Boehringer, Celgene, Lilly, Pfizer. M. Thomas: Honoraria/AD Boards: MSB, BMS, Lilly, Astrazeneca, Roche, Pfizer, Celgene, Novartis. All other authors have declared no conflicts of interest.

Volume 28 | Supplement 5 | September 2017

Annals of Oncology 28 (Supplement 5): v497–v501, 2017 doi:10.1093/annonc/mdx382

PALLIATIVE CARE 1386O

Cancer cachexia (CAX), anorexia and muscle wasting (sarcopenia) assessment in non-small cell lung cancer (NSCLC): an observational study in 531 patients

D. Debieuvre1, H. Morel2, P-J. Souquet3, V. Surmont4, D. Planchard5, F. Bonnetain6, I. Krakowski7, H. Gaudin8, S. Antoun9 1 Department of Pneumology, Hoˆpital Emile Muller, Mulhouse, France, 2Department of Pneumology, C.H.R. Orleans - La Source, Orleans, France, 3Department of Pneumology, Centre Hospitalier Lyon Sud, Pierre Be´nite, France, 4Department of Thoracic Oncology, Gent University Hospital, Gent, Belgium, 5Department of Medical Oncology, GUSTAVE ROUSSY, Villejuif, France, 6Methodological and Quality of Life Unit in Oncology (INSERM UMR 1098), CHRU Besanc¸on Jean Minjoz, Besanc¸on, France, 7De´partement Interdisciplinaire de Soins de Support pour le Patient en Oncologie. (DISSPO-CARE), Institut Bergonie´, Bordeaux, France, 8Medical Affairs, Chugai Pharma France, Paris, France, 9Unite´ de Me´decine aigue en oncologie, Institut Gustave Roussy, Villejuif, France Background: Since publications highlighting the role of sarcopenia, weight loss (WL) is no longer the corner stone of malnutrition assessment. An international consensus proposed in 2011 a definition and a staging of CAX, mainly based on WL, sarcopenia, inflammation and anorexia (Fearon). We initiated this study to fill the gap of epidemiological data on CAX in NSCLC in France and Belgium. Methods: This cross-sectional, prospective, multicentric study was conducted in Patients (pts) with NSCLC regardless of the tumor stage and the treatment line. Skeletal muscle mass (SMM) was assessed by analyzing L3 CT-scan image. Pts completed Anorexia/CAX subscale of FAACT and EORTC QLQ-C30 health related quality of life (QoL) questionnaires. Primary endpoint was the frequency of CAX according to Fearon criteria. Secondary endpoints were the frequency and the characteristics of the other stages of CAX focusing on early and discrete malnutrition changes (pre-CAX). Results: 539 NSCLC pts were recruited within 3 months in 2016 by 56 sites, analysis population was of 531 pts and 312 had SMM assessment. Median age was 66 years, 66.5% were males, 79.9% were PS < 2, and the tumor stage was mainly IIIB-IV (87.3%). 38.7% of pts had CAX, 33.8% pre-CAX and 0.9% refractory CAX. CAX was associated with molecular tumor profiles: 23.9% in patients EGFR, ALK, ROS1, BRAF or HER2 positive, 41.4% in K-RASþ and 43.2% with no molecular abnormality (p ¼ 0.003). Interestingly, the more advanced the CAX stage is, the poorer the score of functional scale (except cognitive) of the QoL questionnaire (p < 0.0001). Sarcopenia was present in 66.7% of CAX pts and 68.5% of pre-CAX pts (all without WL or WL 2%). Notably, 25.8% of pre-CAX pts had only sarcopenia with limited WL (2%) and no anorexia (questioning the mechanisms of sarcopenia). In pts with limited WL (2%), the loss of appetite was associated with sarcopenia in 44% of the cases. Conclusions: This is the first study showing an association between molecular abnormality in NSCLC and cachexia. It has also shown that it may be useful to detect sarcopenia in pts with limited WL ( 5% change. Results: 92,155 patients were included (median age 73 yo, IQR 62-81; 61.9% male; 53.0% metastatic). The prevalence of the ACCEoL was 71.1%, 69.9% in metastatic patients vs. 72.6% in others (p < 0.001), and varying by primary cancer from 62.7% in breast to 79.3% in haematological (p < 0.001). The most prevalent individual indicators were > 14 days in hospital (42.7%; 42.3% in metastatic) and surgery (27.8%; 26.4% metastatic). The least prevalent were permanent tracheostomy (0.1%) and percutaneous gastrostomy (0.3%). Primary outcome remained stable overtime and despite some individual indicators showed statistically significant changes in study timeframe, none of these had > 5% change. Conclusions: Surprisingly, we found unchanged trends of high ACCEoL among adult patients and no clinically meaningful difference for metastatic disease group. A lack of integrated palliative care, even with growing resources in the timeframe analysed, suggest that these have not been enough to reduce ACCEoL. The reduced ACCEoL in patients who died with slow progressive cancers (e.g. breast) suggests that better knowledge of disease trajectories can contribute towards reducing ACCEoL. Legal entity responsible for the study: N/A Funding: Calouste Gulbenkian Foundation, Liga Portuguesa Contra o Cancro N ucleo Regional do Sul Disclosure: All authors have declared no conflicts of interest. Keywords: aggressiveness of care, cancer, end-of-life, palliative care

1388PD

Open-label randomized study of individualized pharmacokinetically (PK)-guided dosing versus body surface area (BSA) dosing of paclitaxel (PTX) in advanced Non-Small Cell Lung Cancer (NSCLC) NCT02058433

S.J. Salamone1, J. Zhang2, H. Qi2, H. Ni2, Y. Li1, J.B. Courtney1, C. Zhou3 Research and Development, Saladax Biomedical, Inc., Bethlehem, PA, USA, 2 Department of Oncology, Shanghai Pulmonary Hospital, Shanghai, China, 3 Department of Oncology, Shanghai Pulmonary Hospital, Tongji University, Shanghai, China

1

Background: Variability of PTX exposure using BSA dosing is well documented and often leads to severe toxicities. While carboplatin is dosed to obtain a specific exposure, paclitaxel is conventionally dosed by BSA, leading to a wide range of exposure. This study compared PTX PK-guided dosing to BSA dosing in a PTX-carboplatin regimen treating stage IIIB/IV NSCLC. This is the final analysis of interim results presented at ASCO 2015 (Poster #375). ClinicalTrials.gov Identifier NCT02058433. Methods: 309 patients with stage IIIB/IV NSCLC were randomized to receive up to 4 cycles of first line 3-weekly carboplatin (AUC 5) and a PTX dose of 175 mg/m2 (Arm A), or a PTX PK-guided dose (Arm B) to achieve a time above a PTX plasma concentration of 0.05mM (Tc>0.05) for 26 to 31 hours. Response was classified according to Response Evaluation Criteria in Solid Tumors Group. PTX concentrations were

C European Society for Medical Oncology 2017. Published by Oxford University Press. V

All rights reserved. For Permissions, please email: [email protected].

abstracts measured by immunoassay; Tc>0.05 was calculated with PK software. The primary endpoint was reduction of grade 4 hematological toxicities. Results: There were 164 patients in Arm A and 155 patients in Arm B, with 191 males and 128 females participating. PK-guided dose adjustment resulted in doses that were widely distributed 73 – 175 mg/m2, and statistically lower than in the BSA arm (by 24%, p < 0.001). Compared to Arm A, PK-guided dosing significantly reduced grade 4 neutropenia by 35% (p ¼ 0.002, 23% vs. 16%) over 4 cycles. The incidence of severe (grade 3) neutropenia was also significantly reduced by 25% in Arm B over all cycles (p ¼ 10mg/L and albumin levels 1 hospitalization (43.3%). Most measures showed no statistically significant time trend. Conclusions: In a setting in early stages of pediatric palliative care development, we found that eight in ten children dying from cancer experience ACcEoL in their last month of life. This estimate is higher than those found in countries in more advanced developmental stages and may indicate a need to increase paediatric palliative care availability. The findings also prompt healthcare professionals to reflect on their current practice, balancing treatments and hospitalisations with patients’ quality of life in the days they have to live. Legal entity responsible for the study: N/A Funding: None Disclosure: All authors have declared no conflicts of interest. Keywords: aggressiveness of care, cancer, children, End of life

doi:10.1093/annonc/mdx382 | 501

Annals of Oncology 28 (Supplement 5): v502–v506, 2017 doi:10.1093/annonc/mdx383

PREVENTION AND SCREENING 1401PD

Fluctuating cancer screening uptake in France: results of the 5th EDIFICE survey

J. Viguier1, S. Couraud2, L. Greillier3, J-Y. Blay4, C. Touboul5, C. Lhomel6, A. Cortot7, J-F. Morere8, F. Eisinger9 1 Public Health and Care, The French National Cancer Institute (INCa), BoulogneBillancourt, France, 2Respiratory Diseases and Thoracic Oncology, Centre Hospitalier Lyon Sud, Pierre Be´nite, France, 3Multidisciplinary Oncology and Therapeutic Innovations, Hopital St. Marguerite Assistance Publique Hopitaux de Marseille, Marseille, France, 4Medical Oncology, Centre Leon Berard, Lyon, France, 5Statistics, Kantarhealth, Paris, France, 6Medical, Roche - France, Boulogne-Billancourt, France, 7 Pneumology-Oncology, Hospital Albert Calmette, Lille, France, 8Medical Oncology, Hopital Paul Brousse, Villejuif, France, 9Aix Marseille Univ, INSERM, SESSTIM, Institute Paoli Calmettes, Marseille, France Background: The EDIFICE nationwide surveys assess attitudes to cancer screening in France. All 5 self-reported surveys (2005, 2008, 2011, 2014 and 2016) focused on breast (BC), colorectal (CRC), prostate cancer (PC) screening; the 4th and 5th editions also included cervical (CC) and lung cancer (LC) screening. Methods: The 5th survey recruited a representative sample of 1299 subjects (men [M], women [F]; age, 50-74 y; no history of cancer) and focused on target populations of the national screening programs for BC and CRC (50-74 y), and on specific subpopulations for PC (M, 50-75 y), CC (F, 50-65 y) and LC (M and F, 55-74 y) screening. Participants were questioned about uptake of at least 1 lifetime screening test and compliance to recommended intervals. Data analysis encompassed nationwide screening programs, opportunistic screening, and vulnerability (assessed by the EPICES score). Results: Rates for at least 1 lifetime BC screening test (screening rate) were 93%/94%/ 95%/97%/97% in 2005/2008/2011/2014/2016, respectively. In line with recommendations, 75%/83%/83%/81%/75% women reported having had a mammogram in the past 2 years (compliance), with a significant drop in 2016 vs 2014 (P¼0.02). Vulnerability had a negative impact on compliance in 2016, though not previously. For CRC, screening rates were 25%/38%/59%/60%/64%. Compliance (FOBT or FIT in the past 2 years) increased steadily from 7% (2005) to 33% (2014), and rose significantly to 38% in 2016 (P¼0.02). The rise was mainly observed in the 50-54 y age group, among men, and in non-vulnerable subjects. In 2016, a significant drop in overall CC screening uptake was observed (99% in 2014 vs 94% in 2016, P18 years. Unaffected carriers were recommended PTG and preoperative esophagogastroduodenoscopy with random gastric biopsies (RGB). Each PTG specimen was wholly sectioned (median # cassettes: 203) to look for occult cancer and histopathology was compared to RGB findings. Results: A novel pathogenic variant in CDH1 c.48G>A (p.Q16Q) was identified in 28 family members, 16 male/12 female. Prior to variant identification, 6 obligate carriers were diagnosed with an advanced DGC, median age 56 (53-62) years and all died of the disease. After genetic testing, 8 asymptomatic carriers were found early-stage DGC in the PTG specimen, median age 25 (19-59) years. Age-specific frequency of DGC in carriers according to PTG is shown in the Table.

Table: 1407P

Age 1406P

Increased mutation burden in high-risk lung tissues: Toward precision cancer risk diagnosis

E. Kubo1, H. Takeshima1, S. Yamashita1, N. Motoi2, T. Ushijima1 1 Division of Epigenomics, National Cancer Center Research Institute, Tokyo, Japan, 2 Department of Pathology and Clinical Laboratories, National Cancer Center Central Hospital, Tokyo, Japan Background: Mutations are believed to accumulate in normal tissues at extremely low levels as a result of exposure to various carcinogenic factors. The degree of accumulation, namely mutation burden, is likely to be associated with cancer risk. However, owing to the limits of current detection methods for such extremely low frequency mutations, the mutation burden present in normal human lung tissues has been unclear. To overcome this limitation, we established a novel method for the quantification of extremely low frequency mutations in DNA samples. Using this method, we aimed to reveal the presence of mutation burden in normal lung tissues and its association with cancer risk. Methods: Somatic mutations were quantified in normal lung tissues without smoking history (n ¼ 11) (“entirely normal lung tissues”:G1), normal lung tissues with smoking history (n ¼ 11) (“smoking-exposed normal tissues”:G2), and non-cancerous lung tissues of patients with lung cancer and smoking history (n ¼ 11) (“smoking-exposed non-cancerous tissues”:G3). A sequence library (15,724 bases of 291 regions of 55 cancer-related genes) was prepared by multiplex PCR using 100 DNA molecules. Libraries were sequenced using a next generation sequencer. Results: The mutation burden in G3 (2.7 6 0.8 105 mutations/base) was significantly higher than that in G1 (1.8 6 0.5 105 mutations/base) (p ¼ 0.0189). Accumulation of somatic mutations tended to be associated with increased cancer risk (OR ¼ 3.75; 95% CI ¼ 0.54–26.046). C>T mutations were significantly more frequent in G2 and G3 than in G1, which is in accordance with reported mutation signatures in cancer tissues [Alexandrov et al., Science, 354:2016]. GCC>GTC and CCC>CTC mutations, signatures of exposure to the nitrosamines contained in tobacco smoke, were significantly enriched in G2 and G3.

Volume 28 | Supplement 5 | September 2017

Change of natural history of hereditary diffuse gastric cancer after identification of a novel CDH1 mutation

10-20 21-30 31-40 41-50 51-60 61-70

DGC n¼6

5 1

No PTG

PTG

n¼8

n ¼ 20

Cumulative frequency

DGC n¼8

Cumulative frequency

62% 75%

1 3 1 1 1 1

5% 20% 25% 30% 35% 40%

Histopathological RGB and PTG correlation was performed in 17 carriers attended at our institution (May 2013-Sept 2015). Median age at PTG was 34 (19-63) years. All preoperative RGB were negative, but one, which identified a single milimetric DGC foci. PTG specimens revealed one Tis and six T1a DGC, conferring RGB a predictive negative value (PNV) of 66% for DGC. Stage IA DGC had a median of 2.8 foci/gastrectomy, localized in the body (83%) and atrium (17%), with average diameter 0.73 mm and ecadherin expression in 100% of the foci. No severe postoperative morbidity was recorded after a median follow-up of 29 (16-44) months. Conclusions: PTG has changed the natural disease history in c.48G>A CDH1 carriers. Endoscopic RGB showed a low PNV for DGC and PTG is still highly recommended. More reliable screening methods are required in order to delay PTG in CDH1-mutation carriers. Legal entity responsible for the study: Vall d’Hebron Hospital Funding: None Disclosure: All authors have declared no conflicts of interest.

doi:10.1093/annonc/mdx383 | 503

abstracts 1409P

Cervical cancer screening in France: recent change in behaviors

T. de la Motte Rouge1, J. Viguier2, J-F. Morere3, L. Brignoli-Guibaudet4, J-Y. Blay5, C. Lhomel6, F. Eisinger7 1 Medical Oncology, Centre Eugene - Marquis, Rennes, France, 2Medical Oncology, CHRU Bretonneau, Tours, France, 3Medical Oncology, Hopital Paul Brousse, Villejuif, France, 4Statistics, Kantarhealth, Paris, France, 5Medical Oncology, Centre Leon Berard, Lyon, France, 6Medical, Roche - France, Boulogne-Billancourt, France, 7Aix Marseille Univ, INSERM, SESSTIM, Institute Paoli Calmettes, Marseille, France Background: Cervical cancer (CC) is the fourth most common cancer in women in France. Human papillomavirus vaccination and screening are complementary secondary prevention measures against CC. Screening by conventional Pap smear is recommended every three years for women aged 25-65y. Methods: The EDIFICE nationwide observational surveys assess population attitudes to cancer screening in general. Representative samples of the French population aged 50-75 years are interviewed by phone using the quota method. Although the French CC screening program covers all women aged 26-65y, the present analysis pertains to a subpopulation aged 50-65y (N ¼ 356 in 2014 and N ¼ 460 in 2016). Interviewees, with no personal history of cancer, were asked if they had ever had a smear test during a gynecological exam. The date of the last test was noted. Data analysis focused on age group, socioprofessional categories (SPC) and social vulnerability (defined by the EPICE score). Results: In 2016, 94% of interviewees reported at least one lifetime smear test vs. 99% in 2014 (P