Tumor mutational burden as a predictive biomarker for molecularly matched therapy in two independent pan-cancer cohorts.

Abstract

3066 Background: Personalized cancer therapy based on molecular tumor aberrations has shown efficacy in a subset of tumors. Novel biomarkers are warranted to help extend this benefit to a larger patient population. Tumor mutational burden (TMB) is an established predictive biomarker for immune checkpoint inhibition. Its role for molecularly matched therapy is unknown. Methods: Comprehensive molecular profiling including whole-exome and RNA-sequencing was performed in 104 patients with advanced cancer within the DKTK MASTER program. 55 patients received a systemic therapy excluding immunotherapy. TMB and survival were analyzed in patients receiving molecularly matched therapy (n=34) or non-molecularly matched therapy (n=21) as well as in an independent published cohort of patients receiving molecularly matched (n=68) or non-molecularly matched (n=40) therapy (excluding immunotherapy). TMB and co-occurring driver mutations were analyzed in the DKTK MASTER cohort. Results: The median TMB of 1.68 mutations per megabase (mut/Mb) in the 34 patients receiving molecularly matched therapy was used to stratify patients into TMB-high and -low groups. Median overall survival (OS) and progression-free survival (PFS) were significantly shorter in the TMB-high than in the TMB-low group (OS: 4 months [95% CI, 3.3 to 7.6] versus 19 months [95%CI, 10.0 to not reached] p < 0.001, Hazard Ratio (HR) 4.4 [95%CI, 1.9 to 10.2], p<0.001; PFS 1.8 months [95%CI, 1.1 to 3.6] versus 8.5 months [95% CI, 2.8 to 17.0] p = 0.004, HR 2.9 [95%CI, 1.4 to 6.2], p=0.005). No significant differences were observed in patients receiving non-molecularly informed systemic therapy (OS HR 0.7 [95%CI, 0.2 to 2.6], p=0.635 and PFS HR 1.2 [95%CI, 0.4 to 3.5] p=0.701). In the validation cohort, shorter progression-free survival was also identified in the TMB-high group (median TMB cut-off of 4 mut/Mb; HR 2.2 [95%CI, 1.3 to 3.7] p=0.005) treated with molecularly matched therapy but not with non-molecularly informed therapy (HR 1.1 [95%CI, 0.5 to 2.2] p=0.848). TMB was associated with co-occurring driver mutations (n=104, r = 0.78 [95%CI, 0.68 to 0.85], p<0.001). Conclusions: High TMB was previously established as a positive predictive biomarker associated with efficacy of immune checkpoint inhibitor treatments. Our study reveals high TMB as a negative predictive biomarker associated with low efficacy of molecularly informed systemic therapy, potentially due to co-occurring driver mutations. Thus, TMB is a candidate predictive pan-cancer biomarker for combined precision oncology and immunotherapy programs.