PERLA is a global, double-blind, parallel phase II trial (NCT04581824) comparing efficacy and safety of anti–PD-1 antibodies dostarlimab and pembrolizumab, plus chemotherapy (DCT and PCT, respectively) as first-line treatment in patients with metastatic non-squamous NSCLC without known targetable genomic aberrations. Patients stratified by PD-L1 tumor proportion score and smoking status were randomized 1:1, receiving ≤35 cycles 500 mg dostarlimab or 200 mg pembrolizumab, ≤35 cycles 500 mg/m2 pemetrexed and ≤4 cycles cisplatin (75 mg/m2) or carboplatin (AUC 5 mg/ml/min) Q3W. Primary endpoint was overall response rate (ORR) (blinded independent central review). Secondary endpoints include progression-free survival (PFS) based on investigator assessment, overall survival (OS) and safety. Exploratory endpoints include ORR by PD-L1 subgroup and duration of response. PERLA met its pre-specified endpoint. ORR (n/N; 95% CI) is 45% (55/121; 36.4–54.8) for DCT and 39% (48/122; 30.6–48.6) for PCT (data cut-off: 07 July 23), numerically favoring dostarlimab in PD-L1-positive subgroups. Median PFS (months [95% CI]) is 8.8 (6.7–10.4) for DCT and 6.7 (4.9–7.1) for PCT (HR 0.70 [95% CI: 0.50–0.98]; data cut-off: 04 August 22). Median OS (months [95% CI]) is 19.4 (14.5–NR) for DCT and 15.9 (11.6–19.3) for PCT (HR 0.75 [95% CI: 0.53–1.05]) (data cut-off: 07 July 23). Safety profiles are similar between groups. In this study, DCT shows similar efficacy to PCT and demonstrates clinical efficacy as first-line treatment for patients with metastatic non-squamous NSCLC.