Validation of a 15-gene prognostic signature in metastatic clear cell renal cell carcinoma.

578 Background: Our group previously developed and validated a 15-gene prognostic signature (15G score) in localized clear cell renal cell carcinoma (ccRCC). Giving its ability to differentiate aggressive and indolent disease in the localized setting, we sought to apply the 15G score to patients with metastatic ccRCC to evaluate the 15G score’s prognostic performance across different treatments. Methods: We analyzed 2,121 patients from four major metastatic ccRCC trial datasets: IMmotion 150 (IMM150), Immotion 151 (IMM151), Javelin Renal 101 (JR101), and Checkmate 009, 010, and 025 (CM009/010/025). Treatment and control arms from all trials were included and grouped within treatment groups which included atezolizumab, atezolizumab plus bevacizumab, sunitinib, avelumab plus axitinib, nivolumab, and everolimus. The 15G score, originally derived and established to be prognostic in localized ccRCC, was calculated using z-score transformed RNA-seq data from each arm of the included trials to classify patients into high-risk versus low-risk using a previously derived threshold. We tested the association of 15G score high-risk with progression free survival (PFS) in each treatment group using Cox proportional hazards analyses with and without controlling for MSKCC risk. Results: On univariable Cox analysis, a 15G high-risk score was associated with significantly worse PFS in five of the seven treatment groups (Table). Worse PFS was observed for atezolizumab, atezolizumab plus bevacizumab, sunitinib, and everolimus (HR [95% CI]: 2.8 [1.6-4.9],1.7 [1.4-2.2], 1.7 [1.4-2.0], and 1.5 [1.1-2.2]. Within IMM151 and CM009/010/025, where risk group data were available, 15G score high-risk was independently associated with worse PFS while controlling for MSKCC risk group for atezolizumab plus bevacizumab, sunitinib, and everolimus (HR [95% CI]: 1.5 [1.2-2.0], 1.7 [1.4-2.2], 1.6 [2.3], p<0.05). Within CM009/010/025, where overall survival (OS) data were available, 15G score high-risk was associated with worse OS for everolimus (HR [95% CI]: 1.5 [1.0-2.2]) but did not reach significance for nivolumab (HR [95% CI]: 1.4 [0.9-2.0]). Conclusions: The 15G score was independently prognostic in metastatic ccRCC among a diverse cohort of patients receiving different systemic therapy regimens. PFS associated with 15G score high-risk compared to low-risk from univariable Cox analysis within each treatment group. Treatment Group Trials 15G scoreLow-risk (n) 15G scoreHigh-risk (n) p-value PFS HR (95% CI) Immunotherapy Atezolizumab IMM150 45 41 <0.001 2.8(1.6, 4.9) Nivolumab CM009/010/025 74 106 0.06 1.4(1.0, 1.9) Combination Atezolizumab+Bevacizumab IMM151/150 218 277 <0.001 1.7(1.4, 2.2) Avelumab+Axitinib JR101 152 202 0.09 1.3(1.0, 1.8) TKI / MTORi Sunitinib IMM150/IMM151/JR101 370 507 <0.001 1.7(1.4, 2.0) Everolimus CM009/010/025 51 78 0.03 1.5(1.1, 2.2)