Long-Term Performance of Prognostic Models for Advanced Renal Cell Carcinoma in the Era of Improved Survival With Immune Checkpoint Inhibitors

PURPOSE: In the era of prolonged survival for advanced renal cell carcinoma (aRCC) with standard-of-care first-line therapy now including immune checkpoint inhibitor, re-evaluation of the Memorial Sloan Kettering Cancer Center (MSKCC) and International Metastatic RCC Database Consortium (IMDC) prognostic models is overdue. METHODS: Data from 1,052 patients with aRCC treated on the CheckMate-214 phase III randomized trial with first-line nivolumab + ipilimumab or sunitinib were analyzed after minimum 5 years of follow-up. The end point was overall survival (OS). To investigate long-term prognostication with each treatment approach, model performance based upon continuous risk score was assessed in a time-dependent manner of increasing 6-month intervals and globally over full follow-up, using discrimination concordance (c)-indices. RESULTS: With time-dependent assessment, the IMDC and MSKCC models maintained their performance over approximately 2 years from sunitinib initiation (c ≥0.69 through 18-24 months); thereafter, the models' performances with long-term OS attenuated. Over full follow-up, the models' discrimination was c = 0.66 (95% CI, 0.658 to 0.664) and c = 0.64 (95% CI, 0.640 to 0.645), respectively, for the sunitinib group. After nivolumab + ipilimumab initiation, the IMDC and MSKCC models' global discrimination was c = 0.63 (95% CI, 0.628 to 0.634) and c = 0.61 (95% CI, 0.607 to 0.614), respectively. The models' performances were attenuated in the short term (c ranging 0.64-0.69 through 18-24 months) and the long term. CONCLUSION: This retrospective analysis of the CheckMate-214 trial, in which nivolumab + ipilimumab improved survival versus sunitinib with 48% and 37% of patients, respectively, surviving beyond 5 years, confirmed the strength of the models' prognostication for the early years after first-line sunitinib initiation continuing to stratify three prognostic categories, but also diminished discrimination among long-term survivors and with initiation of nivolumab + ipilimumab. As novel treatments are developed and patients with aRCC live longer, new models to estimate long-term prognosis are needed.