OBJECTIVESIn the randomized phase III TASUKI-52 trial, nivolumab with carboplatin, paclitaxel, and bevacizumab significantly prolonged the progression-free survival (PFS) of treatment-naive patients with advanced or recurrent nonsquamous non-small cell lung cancer (NSCLC). Here, we report the long-term outcomes of patients treated with nivolumab plus carboplatin, paclitaxel, and bevacizumab with 3 years of follow-up.METHODSPatients with stage IIIB/IV or recurrent nonsquamous NSCLC without sensitizing EGFR, ALK, or ROS1 mutations were randomized (1:1) to receive either nivolumab or placebo, in addition to carboplatin, paclitaxel, and bevacizumab, every 3 weeks. Treatment was continued for a maximum of six cycles. The endpoints included PFS, overall survival (OS), and safety. Exploratory analyses included efficacy and safety in subgroups.RESULTSA total of 550 patients were randomized to the nivolumab arm (n = 275) and placebo arm (n = 275). At the minimum follow-up of 36.1 months, PFS was consistently longer in the nivolumab arm than in the placebo arm (median, 10.6 vs. 8.2 months; hazard ratio [HR], 0.59; 95 % confidence interval [CI], 0.47-0.73; P < 0.0001), with PFS rates of 20.2 % vs. 4.9 %. The median OS was 31.6 months (95 % CI, 26.8-36.5) in the nivolumab arm and 24.7 months (95 % CI, 21.1-28.0) in the placebo arm (HR, 0.71; 95 % CI, 0.57-0.88), with OS rates of 44.2 % and 32.3 %, respectively. Of note, PFS and OS favored the nivolumab arm across patients with different PD-L1 expression levels, and regardless of baseline brain metastasis status. Grade 3-4 treatment-related adverse events occurred in 76.2 % and 74.9 % of the patients in the nivolumab and placebo arms, respectively, while no new safety concerns were identified.CONCLUSIONNivolumab, in addition to carboplatin, paclitaxel, and bevacizumab, remained to demonstrate significantly longer PFS and long-term OS benefit compared with placebo in the first-line treatment of patients with nonsquamous NSCLC. The extended follow-up identified no new safety signals.