Abstract 29: The recently approved high-TMB criteria may introduce a sex bias in response to PD1 inhibitors

Abstract Motivation and research question: The U.S. Food and Drug Administration recently approved treatment with pembrolizumab, an immune checkpoint inhibitor (ICI) targeting PD1 (anti-PD1), for all advanced solid tumors with high tumor mutational burden (TMB), defined as ≥10 mutations/Megabase (mut/Mb). Recent studies have suggested that strength of immune selection, biomarkers of outcome, TMB levels and response to ICI treatment may differ between male and female cancer patients in some tumor types. This motivated us to examine what are the sex-specific implications, if any, of the ≥10 mut/Mb threshold on selecting patients for ICI treatment. Data & Methods: We analyzed the largest ICI cohort publicly available to date (Samstein et al. 2019), including 1,070 patients treated with anti-PD1/PD-L1 monotherapy, 99 treated with anti-CTLA4 and 255 treated with an ICI combination. We focused on the nine solid tumor types with TMB and clinical response data (median follow-up of 19 months) for at least 50 patients. Results: 1.We observed a significant difference in the median TMB across sex among melanoma patients (n=130, female vs male median TMB=8.36 vs 11.81, P<0.02), in concordance with previous observations.2.Notably, we found that the FDA-approved threshold of ≥10 mut/Mb to select patients would result in an unwarranted sex bias in melanoma patients. This threshold successfully identifies female melanoma patients with better overall survival (hazard ratio (HR) =0.19, P<0.03) but fails to do so in male patients (HR=0.94, P<0.85), resulting in a five times higher HR in males vs females at that threshold level (interaction between sex and TMB P<0.03). This bias in effect size by sex is confirmed in two additional melanoma cohorts treated with anti-PD1 (Hugo et al. 2016, N=38, P<0.024; Liu et al. 2016, N=114, P<0.03; and P<0.01 for all three melanoma datasets combined, N=312). We do not observe this differential effect for anti-CTLA4 treatment (P<0.14, N=174) or for anti-PD1 + anti-CTLA4 combination (P<0.8, N=115) or for non-ICI treatments (P<0.4, N=322).3.Analysis of additional eight cancer types whose data are available in (Samstein et al 2019) points to glioblastoma (female vs male HR=0.87 vs 0.5) and cancer of unknown origin (female vs male HR= 1.03 vs 0.15), which show marked differences in the HR between male and female patients when applying the FDA threshold for TMB. However, despite these large effects, they do not achieve statistical significance, probably due to the small size of these datasets. Conclusions: The FDA-approved criteria of 10 mutations/Mb could serve as an informative biomarker for stratifying female melanoma patients but is inadequate for stratifying male patients for anti-PD1 treatment. Our results indicate that its usage is likely to introduce a sex bias in additional cancer types, which will be highly important to carefully test further in larger datasets. Citation Format: Neelam Sinha, Sanju Sinha, Kuoyuan Cheng, Sanna Madan, Alejandro Schaffer, Kenneth Aldape, Ayelet Erez, Brid M. Ryan, Eytan Ruppin. The recently approved high-TMB criteria may introduce a sex bias in response to PD1 inhibitors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 29.