Abstract LB-339: Biomarkers predictive of response to pembrolizumab in head and neck cancer (HNSCC)

Background: Biomarkers predictive of response to anti-PD-1 therapy include tumor mutational burden (TMB) and inflammatory biomarkers such as PD-L1 expression and T-cell activated gene expression profile (GEP). This study evaluated relationships between TMB, PD-L1 expression, and GEP and response to pembrolizumab in HNSCC patients in the KEYNOTE (KN)-012 and KN-55 trials. Methods: Data were combined from 261 HNSCC patients in KN-012 (NCT01848834, subsets of B1 [PD-L1 + , ≥1%, modified proportion score or interface pattern, QualTek IHC; n=34] and B2 [n=73] cohorts) and KN-055 (NCT02255097, platinum/cetuximab resistant, n=154) who had TMB data available by whole exome sequencing. Of these, 258 patients had PD-L1 data (PD-L1 IHC using 22C3 pharmDx, CPS readout) and 240 patients had GEP score (RNA analyzed on NanoString nCounter). HPV +/- status was assessed by p16 IHC and WES methods. Statistical testing of relationships between biomarkers and best overall response (BOR) by logistic regression and progression free survival (PFS) by Cox regression was prespecified and performed in a blinded manner. Results: TMB, PD-L1 CPS and GEP were each significantly associated with BOR in the overall population (p - and 80 HPV + patients. TMB and GEP were each highly associated with BOR regardless of HPV - (p=0.0034 and 0.0084) and HPV + (p=0.0446 and 0.0008) status, respectively. PD-L1 CPS showed similar associations with BOR in HPV - (p=0.0649) and HPV + (p=0.0003) patients. TMB showed no correlation with PD-L1 (r = -0.047) and GEP (r = -0.135); whereas PD-L1 CPS and GEP were moderately correlated (r = 0.469), consistent with an interferon-γ induced T-cell activated microenvironment including PD-L1 induction. TMB, PD-L1 CPS and GEP remained significantly predictive for BOR when assessed individually in joint models of TMB and PD-L1, and TMB and GEP (all p - (n=201) and HPV + (n=58) patients identified by p16 IHC, indicating the feasibility of using the WES method. Conclusion: TMB, PD-L1 and T-cell inflamed GEP were independently predictive of response to pembrolizumab in HNSCC patients, in general regardless of HPV status. When used alone or jointly, these biomarkers may have utility in characterizing responses to anti PD-1 therapies and novel cancer regimens in HNSCC. Citation Format: Tanguy Y. Seiwert, Robert Haddad, Joshua Bauml, Jared Weiss, David G. Pfister, Shilpa Gupta, Ranee Mehra, Iris Gluck, Hyunseok Kang, Francis Worden, J. Paul Eder, Makoto Tahara, Barbara Burtness, Stephen V. Liu, Andrea Webber, Lingkang Huang, Robin Mogg, Razvan Cristescu, Jonathan Cheng, Laura Q. Chow. Biomarkers predictive of response to pembrolizumab in head and neck cancer (HNSCC) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr LB-339.