Oops, you're using an old version of your browser so some of the features on this page may not be displaying properly.

MINIMAL Requirements: Google Chrome 24+Mozilla Firefox 20+Internet Explorer 11Opera 15–18Apple Safari 7SeaMonkey 2.15-2.23

Poster session 05

1325P - TRIDENT: Machine learning (ML) multimodal signatures to identify patients that would benefit most from tremelimumab (T) addition to durvalumab (D) + chemotherapy (CT) with data from the POSEIDON trial

Date

14 Sep 2024

Session

Poster session 05

Topics

Tumour Site

Non-Small Cell Lung Cancer

Presenters

Ferdinandos Skoulidis

Citation

Annals of Oncology (2024) 35 (suppl_2): S802-S877. 10.1016/annonc/annonc1602

Authors

F. Skoulidis1, S. Jabbour2, E.B. Garon3, P. Iyengar4, G. Scagliotti5, L. Ferrer6, G. Etchepare7, O. Gallinato8, J. Faure9, P. Bernard10, T. Colin6, P. Menu11, Y. Lin12, L. Cai13, J. Faria14, A. Remorino15, L. Luciani-Silverman16, K.P. Miller17, D. Dellamonica18, Y. Zhang19

Author affiliations

  • 1 Department Of Thoracic/head And Neck Medical Oncology, The University of Texas M.D. Anderson Cancer Center, 770304000 - Houston/US
  • 2 Department Of Radiation Oncology, Rutgers Cancer Institute of New Jersey, 08901 - New Brunswick/US
  • 3 Department Of Medicine, David Geffen School of Medicine at UCLA, 90404 - Los Angeles/US
  • 4 Radiation Oncology, Memorial Sloan Kettering Cancer Center, 10065 - New York/US
  • 5 Medical Affairs, AstraZeneca, 10123 - Torino/IT
  • 6 Multimodal R&d, SOPHiA GENETICS, 33600 - Pessac/FR
  • 7 Multimodal R&d, SOPHiA GENETICS, 02116 - Boston/US
  • 8 Data Science, SOPHiA GENETICS, 33600 - Pessac/FR
  • 9 Image Processing Team, SOPHiA GENETICS, 33170 - Gradignan/FR
  • 10 Multimodal Data Engineering, SOPHiA GENETICS, 33110 - LE BOUSCAT/FR
  • 11 Product And Medical Department, SOPHiA GENETICS, 1180 - Rolle/CH
  • 12 Biostatistics, AstraZeneca, 94080 - South San Francisco/US
  • 13 Oncology Biometrics, AstraZeneca, 20878 - Gaithersburg/US
  • 14 Oncology Data Science, Oncology R&d, AstraZeneca, CB2 8PA - Cambridge/GB
  • 15 Oncology Digital - Precision Healthcare, AstraZeneca, 08028 - Barcelona/ES
  • 16 Us Medical Affairs, AstraZeneca, 20878 - Gaithersburg/US
  • 17 Oncology Business Unit, AstraZeneca, 20878 - Gaithersburg/US
  • 18 Digital Healthcare Delivery, AstraZeneca, 6340 - Zug/CH
  • 19 Digital Health Transformation, R&d, AstraZeneca, Shanghai/CN

Resources

Login to get immediate access to this content.

If you do not have an ESMO account, please create one for free.

Abstract 1325P

Background

The POSEIDON phase 3 trial showed an improvement in overall survival (OS) with T+D+CT vs CT in 1L mNSCLC. A post-hoc analysis showed a trend for sustained OS benefit with T+D+CT vs CT in patients with STK11, KEAP1 or KRAS mutations. These patients may require novel combinations (eg. incl. CTLA4) that might enhance responses to PD(L)-1 inhibitor-based regimens.

Methods

POSEIDON’s clinical and laboratory data in the T+D+CT and D+CT arms were combined with radiomic data from computed tomography scans via a segmentation algorithm to generate the T_full set and with available genomic data for NSq cases, to generate the T_NSq set. ML was used to train models to identify patients predicted to derive higher OS benefit from addition of T to D+CT. Nested cross-validation was used to assess performance.

Results

Comparison of OS for T+D+CT vs D+CT within the T_full set yielded a HR of 0.88 (95% CI: 0.73-1.05). A model based on clinical+radiomics data yielded an improved HR of 0.75 (95% CI: 0.58-0.98) for patients with scores >median. Neutrophil/Lymphocyte ratio, radiomics and albumin contributed most to the predictive power. Within the T_NSq set, a model based on clinical+genomic data yielded a HR of 0.56 (95% CI: 0.33-0.97) for patients with scores >median, as compared to 0.88 (95% CI: 0.68-1.12) in the total T_NSq set. EGFR (non activating), FGFR3, CDKN2A, KRAS, STK11 mutations contributed most to the predictive power. Table: 1325P

Data set description and baseline characteristics Dataset Analysis set N Age, median (range) Sex n (%) Histology n (%) PD-L1 status n (%)
POSEIDON P_full ITT 676 64.0 (27-87) M: 522 (77.2)F: 54 (22.8) NSq: 423 (62.6)Sq: 252 (37.3)Other: 1 (0.1)

Conclusions

The ML multimodal models yielded signatures identifying patients that would benefit from T addition to D+CT in 1L mNSCLC: NSq histology and an EGFRwt, FGFR3wt, CDKN2Awt and KRAS and STK11 mutation signature drive higher benefit. Limitations of this analysis include its post hoc nature and need for external validation (planned). TRITON (NCT06008093) study is further evaluating STK11/KEAP1/KRASm mNSCLC population with T+D+CT vs pembrolizumab+CT.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

AstraZeneca.

Funding

AstraZeneca.

Disclosure

F. Skoulidis: Financial Interests, Personal, Invited Speaker: ESMO, AACR, IASLC, Japanese Lung Cancer Society, Medscape LLC, Intellisphere LLC, VSPO McGill Université de Montreal, MJH Life Sciences, IDEOlogy Health, MI&T, PER LLC, CURIO LLC, DAVA Oncology, BMS and RV Mais Promocao Eventos LTDS; Financial Interests, Institutional, Research Grant: Amgen, Revolution Medicines, Mirati Therapeutics, Boehringer Ingelheim, Merck & Co, and Novartis; Financial Interests, Institutional, Research Grant, Spouse: Almmune, Genentech; Financial Interests, Personal, Speaker, Consultant, Advisor: Amgen Inc., Revolution Medicines, Novartis, BridgeBio, BeiGene, BergenBio, AstraZeneca, Guardant Health Inc., Calithera Biosciences, Tango Therapeutics, Merck Sharp & Dohme, Roche, Novocure, Hookipa Pharma; Financial Interests, Institutional, Speaker, Consultant, Advisor, Spouse: Genentech, Novartis; Financial Interests, Personal, Stocks or ownership: BioNTech SE, Moderna Inc.; Financial Interests, Institutional, Trial Chair: Pfizer. S. Jabbour: Financial Interests, Personal, Advisory Board: AstraZeneca, Merck; Financial Interests, Personal, Expert Testimony: Dechert LLP; Financial Interests, Institutional, Funding: Merck, Adlai, BeiGene, Guardant; Non-Financial Interests, Institutional, Local PI: Merck; Financial Interests, Personal, Speaker, Consultant, Advisor: Ideology, BeiGene, MJH LifeScience, Binaytara. E.B. Garon: Financial Interests, Institutional, Advisory Role: AbbVie; ABL-Bio; AstraZeneca, Boehringer-Ingelheim; Bristol Myers Squibb; Dracen Pharmaceuticals; EMD Serono; Eisai; Eli Lilly; Gilead, GSK; Merck; Natera; Novartis; Personalis; Regeneron; Sanofi; Shionogi; Xilio; Zymeworks; Financial Interests, Institutional, Research Grant: ABL-Bio; AstraZeneca; Bristol Myers Squibb; Daiichi Sanko; Dynavax Technologies; Eli Lilly; EMD Serono; Genentech; Gilead; Iovance Biotherapeutics, Merck; Mirati Therapeutics; Neon; and Novartis. P. Iyengar: Financial Interests, Personal, Advisory Board: AstraZeneca, Novocure, BioConvergent; Financial Interests, Institutional, Funding: Incyte. G. Scagliotti: Financial Interests, Personal, Speaker, Consultant, Advisor, Independent Consultant: AstraZeneca. L. Ferrer, G. Etchepare, O. Gallinato, T. Colin, P. Menu: Financial Interests, Personal, Full or part-time Employment: Sophia Genetics; Financial Interests, Personal, Stocks/Shares: Sophia Genetics. J. Faure, P. Bernard: Financial Interests, Personal, Full or part-time Employment: Sophia Genetics. Y. Lin: Financial Interests, Personal, Full or part-time Employment: Novartis. L. Cai, J. Faria D. Dellamonica, Y. Zhang: Financial Interests, Personal, Full or part-time Employment: AstraZeneca; Financial Interests, Personal, Stocks/Shares: AstraZeneca. A. Remorino, K.P. Miller: Financial Interests, Personal, Full or part-time Employment: AstraZeneca. L. Luciani-Silverman: Financial Interests, Personal, Full or part-time Employment: AstraZeneca; Financial Interests, Personal, Stocks/Shares: AstraZeneca, BMS.

This site uses cookies. Some of these cookies are essential, while others help us improve your experience by providing insights into how the site is being used.

For more detailed information on the cookies we use, please check our Privacy Policy.

Customise settings
  • Necessary cookies enable core functionality. The website cannot function properly without these cookies, and you can only disable them by changing your browser preferences.