Background: The role of somatic genomic mutations in predicting outcomes in clear cell renal cell carcinoma (ccRCC) is poorly defined. Both ipilimumab/nivolumab (IO/IO) and several combinations of VEGFR-targeted tyrosine kinase inhibitor (TKI/IO) are preferred first-line treatment options for metastatic ccRCC. The impact of somatic mutations on PFS-2 for IO combination therapies was evaluated in this retrospective analysis. Methods: One-hundred forty-one patients with ccRCC had sequencing with MSK-IMPACT and initiated 1^st line combination IO at Memorial Sloan Kettering (MSK) Cancer Center between 1/1/2014 and 12/30/2020. Treatment groups are defined as IO/IO (n=76) and TKI/IO (n=65). PFS-2 is defined as time from start of 1^st line to progression on next therapy, or death, and is estimated using Kaplan-Meier method. PFS-2 association with each mutation was assessed via log-rank test and a model-based approach. Restricted mean survival time (RMST) up to 48 months was calculated for PFS-2 and modeled using a generalized linear model adjusted for IMDC risk. To test for heterogeneity of effect among treatment groups, an interaction test between somatic mutation and treatment group is performed in the model. Results: VHL, PBRM1, SETD2, mTOR pathway, BAP1, and TP53 were the most common somatic mutations. Only BAP1 was significantly associated with PFS-2: median PFS-2 was 44 months for patients without BAP1 mutation and 17 months for patients with BAP1 mutation (log-rank p=0.006). After adjusting for IMDC risk group, 48 month RMST for PFS-2 was 13 months lower in patients with BAP1 mutation. An interaction test performed in the RMST model showed no significant variation in the treatment effect of IO/IO or TKI/IO based on the presence or absence of BAP1 mutation (p=0.97), although sample size limits power to detect differences. Results for other somatic mutations are shown. Conclusions: PFS-2 was significantly inferior for ccRCC patients with somatic BAP1 mutation treated with IO-based therapy. Interaction test did not suggest a significant advantage for IO/IO or TKI/IO for mutation-present or -absent groups. Clinical stratification remains key for selecting TKI/IO or IO/IO for patients.

*NE, not estimated. *Calculated after 48 months of follow-up and adjusted for IMDC risk in a generalized linear model (PROC RMSTREG, SAS v9.4).