Do tissue-based biomarkers in metastatic clear cell renal cell carcinoma perform better when derived from a metastatic site?

569 Background: Gene expression and mutation analyses in metastatic clear cell renal cell carcinoma (ccRCCs) have attempted to identify tumors that differentially respond to tyrosine kinase inhibitors (TKIs) and immunotherapy (IO) – broadly an angiogenic and an immunogenic clade. However, ccRCC is genomically heterogeneous and sequencing a single-region of a primary tumor does not necessarily represent the metastatic clone. We re-analyzed the biomarker results of the Phase III IMmotion151 trial, stratifying patients into primary-derived and metastasis-derived sequencing, to assess if metastasis-derived sequencing performed better in differentiating response to therapy. Methods: We replicated the IMmotion151 biomarker analyses on transcriptomic clusters and mutations on all patients and then stratified to 198 patients with metastasis-derived and 625 patients with primary-derived sequencing. We assessed progression-free survival (PFS) with Cox proportional hazards while controlling for IMDC risk group for atezolizumab and bevacizumab (atezo + bev) versus sunitinib based on previously-described angiogenic groupings: Cluster 1 (Angiogenic/Stromal), Cluster 2 (Angiogenic), and PBRM1 mutant tumors; and immunogenic groupings: Cluster 4 (T-effector/Proliferative), Cluster 5 (Proliferative), PBRM1 wild-type and CDKN2A -loss tumors. Results: PFS favored Sunitinib in angiogenic Cluster 2 and Clusters 1+2 when sequencing was derived from the metastatic site (HR 3.39, p=0.011; HR 2.91, p=0.005) but not the primary (HR 1.04 p=0.82; HR 1.02, p=0.88) (Table). A similar trend was observed for other angiogenic groups (Cluster 1 & PBRM1 mutant), but did not reach significance. PFS favoring atezo + bev for immunogenic groupings remained significant when derived from the primary. Conclusions: Angiogenic signatures distinctly favored sunitinib in advanced ccRCC when sequencing was derived from the metastasis but not the primary tumor. Immunogenic signatures favored atezo + bev even when derived from the primary site. A prevalent angiogenic signature throughout the primary tumor, that is now always present in the metastatic clone, may explain these results. Cox proportional hazards results for Atezo + Bev versus Sunitinib corrected for IMDC risk-group. HR >1 favors Sunitinib, HR <1 favors Atezo + Bev. Marker All patientsN= 823 for RNA-seqN= 715 patients for mutations Metastasis-Derived Sequencing OnlyN= 198 for RNA-seqN= 152 patients for mutations Primary Derived Sequencing OnlyN= 625 for RNA-seqN= 563 patients for mutations DFS- HR P- value DFS- HR P- value DFS- HR P- value Markers Favoring Angiogenic Response / Sunitinib Cluster 1- Angio/Stromal(n=98) 1.21 0.48 3.10 0.14 1.02 0.95 Cluster 2- Angio(n=245) 1.22 0.25 3.33 0.024 1.04 0.82 Cluster 1 + 2(n=343) 1.20 0.22 2.66 0.011 1.02 0.88 PBRM1 mutant(n=333) 1.05 0.73 1.72 0.13 0.86 0.33