Biomarker analysis of the phase 3 KEYNOTE-426 study of pembrolizumab (P) plus axitinib (A) versus sunitinib (S) for advanced renal cell carcinoma (RCC).
Brian I. Rini, Elizabeth R. Plimack, V.P. Stus, Rustem Gafanov, Tom Waddell, Dmitry Nosov, Frédéric Pouliot, B. Yа. Alekseev, Denis Soulières, Sérgio Jobim Azevedo, Delphine Borchiellini, Raymond S. McDermott, Rodolfo F. Perini, Julia F. Markensohn, L. Rhoda Molife, Yiwei Zhang, Michael Nebozhyn, Andrey Loboda, Amir Vajdi, Thomas Powles
Abstract
4505 Background: P + A improved OS, PFS, and ORR over S in 1L advanced RCC in KEYNOTE-426 (NCT02853331).Here, we present exploratory biomarker results including RNAseq, WES, and PD-L1. Methods: Patients (pts) with treatment-naive advanced RCC were randomly assigned 1:1 to P + A or S. Association between T-cell–inflamed gene signature (Tcell inf GEP), angiogenesis gene signature (RNAseq), and PD-L1 CPS (22C3 IHC) with clinical outcomes were tested at prespecified α=0.05. Other RNA signatures (Cristescu et al. Clin Cancer Res. 2022. 2022;28:1680) and molecular subtypes, based on clustering identified from IMmotion151 (Motzer et al. Cancer Cell. 2020;38:803), were tested at prespecified α = 0.10 after multiplicity adjustment. DNA mutations ( VHL, PBRM1, SETD2, and BAP1) by WES were tested at prespecified α = 0.10 after multiplicity adjustment. Results: Of 861 pts, 369 (P + A) and 361 (S) had archival samples for RNAseq; 347 (P + A) and 351 (S) had WES samples. PD-L1 CPS was negatively associated with OS ( P=0.013) for S. There was a strong positive association of Tcell inf GEP with OS ( P=0.003), PFS ( P<0.0001), and ORR ( P<0.0001) for P + A. Angiogenesis was positively associated with OS ( P=0.013) for P + A; there was a strong positive association with OS ( P<0.0001), PFS ( P<0.001), and ORR ( P=0.002) for S. For other RNA signatures, positive association with mMDSC was found for PFS ( P=0.018) and ORR ( P=0.093) with P + A. For S, positive association was found with hypoxia (OS, P=0.034; ORR, P=0.071) and negative associations with MYC (OS, P<0.001; PFS, P=0.012) and proliferation (OS, P=0.002). Across all molecular clusters, ORR favored P + A over S, with the highest P + A ORR in the immune/proliferative cluster (Table). By WES, PBRM1 mutation had positive association with ORR ( P=0.004) and PFS ( P=0.079) for P + A. For S, positive associations were observed with OS for VHL ( P=0.073) and PBRM1 ( P=0.001) mutations and a negative one observed for BAP1 mutation ( P=0.046). P + A improved ORR over S regardless of mutational status. Conclusions: There was a strong relationship of Tcell inf GEP with clinical outcomes with P + A. Angiogenesis was positively associated with outcomes with S and only with OS with P + A. Further understanding the role of the immune microenvironment in combination therapy will be critical to advance treatment strategies. Clinical trial information: NCT02853331 . [Table: see text]