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 Display session 3

1898 - Genomic correlates of response to anti-PDL1 Atezolizumab in non-small-cell lung cancer OAK and POPLAR trials


30 Sep 2019


Poster Display session 3



Tumour Site

Non-Small Cell Lung Cancer


Hari Singhal


Annals of Oncology (2019) 30 (suppl_5): v475-v532. 10.1093/annonc/mdz253


H. Singhal1, O. Ekinci2, C. Alcorn3, D.R. Gandara4

Author affiliations

  • 1 Diagnostics Information Solutions, Roche Molecular Systems, 94002 - Belmont/US
  • 2 Roche Diagnostics Information Solutions, Roche Molecular Systems, 95050 - Santa Clara/US
  • 3 Diagnostics Information Solutions, Roche Molecular Systems, 94402 - Belmont/US
  • 4 Internal Med: Hematology-oncology, University of California Davis Cancer Center, 95817 - Sacramento/US


Login to get immediate access to this content.

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

Abstract 1898


Limitations of integrating biological knowledgebases with genomics impedes the development of predictive correlates that would help in personalization of immunotherapy. Non-small cell lung cancer (NSCLC) patients treated with atezolizumab, an antiprogrammed death-ligand 1 (PDL1) antibody, have better overall survival when compared to patients receiving docetaxel chemotherapy in Poplar (Lancet, 2016) and Oak (Lancet, 2017). We hypothesized that patterns in the mutations of immune signatures would correlate with the immunotherapeutic effect of atezolizumab in NSCLC.


Sequencing data from Poplar (n = 277 patients) and Oak (n = 725 patients) trials was analyzed for understanding genomic alterations in relation to patient outcomes. Signatures from publicly-available knowledgebases were integrated with the genomics and clinicopathological data into an algorithm for identifying patterns correlative of immunotherapeutic response.


Patients benefitting from atezolizumab were more likely to have mutations in oncoimmunity-related genes which were significantly overlapping between the two trials. These overlapping genes were used to develop a de novo signature comprising of CDKN1A, ERRFI1, JAK2, NOTCH2, ACVR1B, NFKBIA, GNA13, MERTK, BTG1, CDKN1B, FOXP1, PDK1, ETV6, MLL2, SMAD3, DICER1 and BRCA2. Mutations in any of the genes within the signature identified patient subpopulations with higher immunotherapeutic response to Atezolizumab in both Oak (HR = 0.3, P = 1.8e-6 versus HR = 0.76, P = 5.5e-3 for entire cohort) and Poplar (HR = 0.18 and P = 3.6e-2 versus HR = 0.71, P = 0.023 for all patients) trials. Prediction efficiency of the signature was significantly better than established biomarkers such as PDL1 staining (Nature, 2014) in both Oak (HR: 0.58 and P = 1.8e-6) and Poplar (HR: 0.58 and P = 0.04) trials indicating that genomic correlates could add significant value to existing modalities in predicting immunotherapy response.


In summary, integration of biological knowledgebases with genomics suggests that a rheostat of mutational burden in non-overlapping genesets is associated with immunotherapeutic effect and identifies novel genomic correlates for response to Atezolizumab.

Clinical trial identification

NCT02008227; NCT01903993.

Editorial acknowledgement

Legal entity responsible for the study





D.R. Gandara: Research grant: Bristol-Myers Squibb, Roche-Genentech, Novartis, Merck; Consultancy: AstraZeneca, Celgene, CellMax Life, FujiFilm, Roche-Genentech, Guardant Health, Inviata, IO Biotech, Lilly, Liquid Genomics, Merck, Samsung Bioepis, Pfizer. All other authors have declared no conflicts of interest.

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.