BACKGROUND: PD-1 inhibitor plus chemotherapy had been shown to be an effective first-line treatment for patients with metastatic non-small-cell lung cancer (NSCLC). However, there was no robust evidence showing a PD-L1 inhibitor combined with chemotherapy benefited patients with squamous and non-squamous NSCLC. GEMSTONE-302 aimed to evaluate the efficacy and safety of a PD-L1 inhibitor, sugemalimab, plus chemotherapy for patients with metastatic squamous or non-squamous NSCLC. METHODS: This randomised, double-blind, phase 3 trial was done in 35 hospitals and academic research centres in China. Eligible patients were aged 18-75 years, had histologically or cytologically confirmed stage IV squamous or non-squamous NSCLC without known EGFR sensitising mutations, ALK, ROS1, or RET fusions, no previous systemic treatment for metastatic disease, and an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1. Patients were randomly assigned (2:1) to receive sugemalimab (1200 mg, intravenously, every 3 weeks) plus platinum-based chemotherapy (carboplatin [area under the curve (AUC) 5 mg/mL per min, intravenously] and paclitaxel [175 mg/m, intravenously] for squamous NSCLC, or carboplatin [AUC 5 mg/mL per min, intravenously] and pemetrexed [500 mg/m, intravenously] for non-squamous NSCLC; sugemalimab group) or placebo plus the same platinum-based chemotherapy regimens for squamous or non-squamous NSCLC as in the sugemalimab group; placebo group) for up to four cycles, followed by maintenance therapy with sugemalimab or placebo for squamous NSCLC, and intravenous sugemalimab 500 mg/m or matching placebo plus pemetrexed for non-squamous NSCLC. Randomisation was done by an interactive voice-web-response system via permuted blocks (block size was a mixture of three and six with a random order within each stratum) and stratified by ECOG performance status, PD-L1 expression, and tumour pathology. The investigators, patients, and the sponsor were masked to treatment assignment. The primary endpoint was investigator-assessed progression-free survival in the intention-to-treat population. Safety was analysed in all patients who received at least one treatment dose. Results reported are from a prespecified interim analysis (ie, when the study met the primary endpoint) and an updated analysis (prespecified final analysis for progression-free survival) with a longer follow-up. This study is registered with ClinicalTrials.gov (NCT03789604), is closed to new participants, and follow-up is ongoing. FINDINGS: Between Dec 13, 2018, and May 15, 2020, 846 patients were assessed for eligibility; 367 were ineligible, and the remaining 479 patients were randomly assigned to the sugemalimab group (n=320) or placebo group (n=159). At the preplanned interim analysis (data cutoff June 8, 2020; median follow-up 8·6 months [IQR 6·1-11·4]), GEMSTONE-302 met its primary endpoint, with significantly longer progression-free survival in the sugemalimab group compared with the placebo group (median 7·8 months [95% CI 6·9-9·0] vs 4·9 months [4·7-5·0]; stratified hazard ratio [HR] 0·50 [95% CI 0·39-0·64], p<0·0001]). At the final analysis (March 15, 2021) with a median follow-up of 17·8 months (IQR 15·1-20·9), the improvement in progression-free survival was maintained (median 9·0 months [95% CI 7·4-10·8] vs 4·9 months [4·8-5·1]; stratified HR 0·48 [95% CI 0·39-0·60], p<0·0001). The most common grade 3 or 4 any treatment-related adverse events were neutrophil count decreased (104 [33%] of 320 with sugemalimab vs 52 [33%] of 159 with placebo), white blood cell count decreased (45 [14%] vs 27 [17%]), anaemia (43 [13%] vs 18 [11%]), platelet count decreased (33 [10%] vs 15 [9%]), and neutropenia (12 [4%] vs seven [4%]). Any treatment-related serious adverse events occurred in 73 (23%) patients in the sugemalimab group and 31 (20%) patients in the placebo group. Any treatment-related deaths were reported in ten (3%) patients in the sugemalimab group (pneumonia with respiratory failure in one patient; myelosuppression with septic shock in one patient; pneumonia in two patients; respiratory failure, abdominal pain, cardiac failure, and immune-mediated pneumonitis in one patient each; the other two deaths had an unspecified cause) and in two (1%) patients in the placebo group (pneumonia and multiple organ dysfunction syndrome). INTERPRETATION: Sugemalimab plus chemotherapy showed a statistically significant and clinically meaningful progression-free survival improvement compared with placebo plus chemotherapy, in patients with previously untreated squamous and non-squamous metastatic NSCLC, regardless of PD-L1 expression, and could be a newfirst-line treatment option for both squamous and non-squamous metastatic NSCLC. FUNDING: CStone Pharmaceuticals. TRANSLATION: For the Chinese translation of the abstract see Supplementary Materials section.