Abstract
Active surveillance (AS) may be used in the management of metastatic renal cell carcinoma (mRCC), but consensus regarding its application is lacking. We report an exploratory analysis of prospectively collected specimens prespecified in the only modern clinical trial evaluating AS in mRCC. Whole-exome and RNA sequencing were performed for patients providing consent to identify putative biomarkers associated with time on AS (TAS), the primary endpoint. Log-rank tests and multivariable Cox proportional-hazards models were used for analyses. Patients with mutations in either TP53 or SMARCA4 tumor suppressor genes had shorter TAS (7.5 vs 14.2 mo; log-rank p = 0.004). While these patients exhibited features of aggressive disease clinically, the two-gene model was independently predictive in multivariable analyses (hazard ratio 3.30, 95% confidence interval 1.07-10.18; p = 0.038). In conclusion, insight into the underlying RCC biology improves patient selection for AS. If validated, this two-gene model could help in stratifying patients with mRCC and identifying those who are poor candidates for AS. PATIENT SUMMARY: In this study, we analyzed tumors from patients with metastatic kidney cancer enrolled in a clinical trial of imaging surveillance. We found that tumors with mutations in either the TP53 or SMARCA4 gene progressed faster than tumors without these mutations. Thus, patients harboring mutations in these genes may not be good candidates for AS.