Dr Rini on a Biomarker Analysis of KEYNOTE-426 Using Subtype Clustering RCC

Brian I. Rini, MD, FASCO discusses findings from an exploratory analysis of the phase 3 KEYNOTE-426 study.

Brian I. Rini, MD, FASCO, Ingram Professor of Medicine, Department of Medicine, Division of Hematology Oncology, Vanderbilt University, discusses findings from an analysis of subtype clustering from the phase 3 KEYNOTE-426 study (NCT02853331) evaluating pembrolizumab (Keytruda) plus axitinib (Inlyta) vs sunitinib (Sutent) for the first-line treatment of patients with advanced renal cell carcinoma (RCC).

At the 2024 ASCO Annual Meeting, Rini and colleagues presented findings from a biomarker analysis of KEYNOTE-426, which showed that T-cell–inflamed gene expression profile (GEP) was associated with improved clinical outcomes within the pembrolizumab plus axitinib arm.

As a part of the analysis, investigators analyzed patients by subtype clusters for patients with RCC derived from a previous analysis of the phase 3 IMmotion151 trial (NCT02420821), which evaluated atezolizumab (Tecentriq) plus bevacizumab (Avastin) compared with sunitinib in patients with locally advanced or metastatic RCC.

IMmotion151 established a 7-cluster analysis system based on RNA expression, which was used to evaluate biomarker data. The clusters include: angiogenic, angiogenic/stromal, stromal/proliferative, immune/proliferative, proliferative, and other.

Data from the analysis of KEYNOTE-426 showed that the distribution of patients by cluster was similar to IMmotion151 and other previous trials. Rini notes that although pembrolizumab plus axitinib improved overall response rate (ORR) across all cluster subtypes, the greatest advantage and the highest absolute ORR was observed in cluster 4—the immune/proliferative cluster. The smallest difference in ORR and the highest ORR to sunitinib was observed in the second cluster—the pure angiogenic cluster.

Rini explains that this analysis is another way that clinicians and researchers can understand and organize biomarker data, emphasizing that there is not one right or wrong way to use these analyses. An analysis utilizing cluster subtypes is another potential way to enhance the understanding of biomarker data in advanced RCC, Rini concludes.