Background: In CheckMate 025 (NCT01668784), nivolumab demonstrated superior OS (25.0 mo; 95% CI, 21.8–not estimable [NE]) vs everolimus (19.6 mo; 95% CI, 17.6–23.1) and a higher objective response rate (25% vs 5%, P< 0.001) in previously treated patients with aRCC (N Engl J Med 2015;373:1803–13). Here, we investigated OS benefit with nivolumab vs everolimus by tumor response. Methods: Patients were randomized 1:1 to nivolumab 3 mg/kg intravenously every 2 wk or everolimus 10 mg orally once daily. We conducted a landmark analysis of OS by best response (complete/partial response; responders), stable disease [SD], and progressive disease [PD]) within 4 mo of randomization. OS was assessed using the Kaplan–Meier method. Results: For the 410 and 411 patients randomized to nivolumab and everolimus, median time to response was 3.5 mo (range: 1.4–24.8) and 3.7 mo (1.5–11.2), respectively. By mo 4 with nivolumab, 19% of patients were responders, 30% had SD, and 40% had PD; with everolimus, 3% of patients were responders, 45% had SD, and 31% had PD. The remainder of patients had discontinued, died, or had an unevaluable response by 4 mo. Median (95% CI) OS is shown (table). For nivolumab, OS rates at 12 and 18 mo were 96% and 89% in responders, 91% and 81% in patients with SD, and 70% and 50% in patients with PD by 4 mo. For everolimus, OS rates at 12 and 18 mo were 93% and 77% in responders, 90% and 79% in patients with SD, and 54% and 35% in patients with PD by 4 mo. Analyses of each subgroup by disease features will be presented. Conclusions: Patients who had a best response of SD within 4 mo of starting treatment had a similar survival trajectory compared with responders. Patients with a best response of progression had an improved OS with nivolumab vs everolimus. Clinical trial information: NCT01668784