Biomarker-defined outcomes in metastatic clear cell renal cell carcinoma (mccRCC) receiving first-line (1L) immune checkpoint inhibitor-based therapy.

4546 Background: Dual immune checkpoint inhibitor (IO-IO) and immune checkpoint inhibitor plus VEGF receptor tyrosine kinase inhibitor (IO-TKI) regimens are approved 1L treatments for mccRCC, yet treatment selection remains empiric in the absence of prospective comparisons and validated predictive biomarkers. While prior retrospective studies have linked genomic alterations such as PBRM1 with favorable immunotherapy outcomes (Miao et al., 2018), the impact of these alterations across IO-IO vs. IO-TKI regimens is not well characterized. We evaluate PD-L1 expression and tumor genomic alterations in relation to clinical outcomes among patients receiving 1L IO-based therapy. Methods: Adults with mccRCC diagnosed through 2/4/2025 were identified in ConcertAI’s Patient360, a U.S. de-identified database integrating health records, claims, genomics, and mortality data. Patients initiating 1L IO-IO (ipilimumab + nivolumab) or IO-TKI (cabozantinib + nivolumab, pembrolizumab + axitinib, or lenvatinib + pembrolizumab) were included. Molecular testing performed prior to or during 1L therapy was collected. PD-L1 positivity was defined as ≥1% tumor cell staining. Overall survival (OS) and progression-free survival (PFS) were estimated via Kaplan-Meier methods; associations by treatment and biomarker status were assessed using Cox models. Results: Among 4,845 patients with mccRCC, 3,225 patients had 1L treatment data, of whom 1,007 (31.2%) received IO-based therapy. IO-based therapy included IO-IO (508 patients; 50.4%) and IO-TKI (499 patients; 49.6%). Among patients receiving IO-based therapy, median age was 62 years, and 63.6% had de novo metastatic disease. Tumor genomic alterations were identified in 125 patients, including VHL (77.6%), PBRM1 (50.4%), BAP1 (28.8%), and SETD2 (4.8%). Patients with any alterations in VHL , PBRM1 , or BAP1 demonstrated significantly longer median PFS with IO-TKI compared with IO-IO (37.9 vs. 18.4 months, respectively; p=0.04). In gene-specific analyses, similarly large numerical differences in PFS were observed with IO-TKI vs. IO-IO amongst patients with VHL (38.0 vs. 18.4 months) and PBRM1 (34.4 vs. 18.4 months) alterations. In the overall cohort, median PFS was 32.2 vs. 20.8 months (p=0.06), and median OS was 58.4 vs. 56.7 months (p=0.60) for IO–TKI and IO–IO, respectively. Among 102 patients with available PD-L1 status, 25.5% were PD-L1-positive and 74.5% were PD-L1-negative. Among PD-L1-positive patients, OS and PFS did not significantly differ between regimens. Conclusions: Among patients with mccRCC receiving 1L IO-based therapy, those harboring VHL , PBRM1 , or BAP1 alterations were collectively associated with numerically longer PFS when treated with IO-TKI. These findings represent one of the first and largest biomarker-stratified comparisons of outcomes between IO-TKI and IO-IO regimens.