Relationship of tumor mutational burden (TMB) to immunotherapy response in metastatic renal cell carcinoma (mRCC).

Abstract

662 Background: Multiple studies across histologies (e.g., lung cancer, melanoma and bladder cancer) have linked TMB to response to checkpoint inhibition (CPI). However, the association in patients with mRCC remains unclear. Methods: From an institutional database, consecutive patients (pts) with mRCC who had comprehensive genomic profiling (CGP) done in the course of routine clinical care were assessed. Data pertaining to systemic therapy was collected, and for each patient, duration of therapy (DOT) on 1st immunooncology agent (DOT-IO) was computed. Furthermore, best response data (using a RECIST v1.1 defintion) was collected, and patients were characterized as either having clinical benefit (CB; defined as complete/partial response or stable disease) or no clinical benefit (NCB; progressive disease). CGP was performed on hybridization-captured, adaptor ligation based libraries to a mean coverage depth of 718X for up to 315 cancer-related genes plus introns from up to 28 genes frequently rearranged in cancer. TMB was estimated using the cumulative number of non-variant of unknown significance (non-VUS) and VUS genomic alterations (GAs). The correlation between TMB and DOT-IO as well as CB rate was assessed. Results: Of 31 patients (26:5 M:F) who received IO, 25 pts and 6 pts had clear cell and non-clear cell (papillary = 4; sarcomatoid = 1; chromophobe = 1) histologies, respectively. The most common 1st IO agent rendered was nivolumab (77%), and this was administered most commonly in the post-2nd-line setting (60%). The median number of GAs in the cohort was 10, with the most frequent GAs encountered being VHL (21%), TERT (10%) and PBRM1 (10%). When segregated in groups by estimate of TMB, DOT-IO was similar in high TMB and low TMB pts (71 days and 70 days, respectively; P = 0.39). The median estimated TMB was similar in pts with CB and NCB (11 and 10, respectively; P = 0.84). Conclusions: While previous studies in other histologies have suggested an association between TMB and IO response, our study fails to confirm these findings in mRCC. Given our limited sample size, this association warrants exploration in larger series.