Abstract

Background: First-line nivolumab plus ipilimumab has shown improved overall survival in patients with advanced non-small-cell lung cancer (NSCLC). We aimed to investigate whether the addition of a limited course (two cycles) of chemotherapy to this combination would further enhance the clinical benefit. Methods: This randomised, open-label, phase 3 trial was done at 103 hospitals in 19 countries. Eligible patients were aged 18 years or older with treatment-naive, histologically confirmed stage IV or recurrent NSCLC, and an Eastern Cooperative Oncology Group performance status of 0–1. Patients were randomly assigned (1:1) by an interactive web response system via permuted blocks (block size of four) to nivolumab (360 mg intravenously every 3 weeks) plus ipilimumab (1 mg/kg intravenously every 6 weeks) combined with histology-based, platinum doublet chemotherapy (intravenously every 3 weeks for two cycles; experimental group), or chemotherapy alone (every 3 weeks for four cycles; control group). Randomisation was stratified by tumour histology, sex, and PD-L1 expression. The primary endpoint was overall survival in all randomly assigned patients. Safety was analysed in all treated patients. Results reported here are from a pre-planned interim analysis (when the study met its primary endpoint) and an exploratory longer-term follow-up analysis. This study is active but no longer recruiting patients, and is registered with ClinicalTrials.gov, number NCT03215706. Findings: Between Aug 24, 2017, and Jan 30, 2019, 1150 patients were enrolled and 719 (62·5%) randomly assigned to nivolumab plus ipilimumab with two cycles of chemotherapy (n=361 [50%]) or four cycles of chemotherapy alone (n=358 [50%]). At the pre-planned interim analysis (median follow-up 9·7 months [IQR 6·4–12·8]), overall survival in all randomly assigned patients was significantly longer in the experimental group than in the control group (median 14·1 months [95% CI 13·2–16·2] vs 10·7 months [9·5–12·4]; hazard ratio [HR] 0·69 [96·71% CI 0·55–0·87]; p=0·00065). With 3·5 months longer median follow-up (median 13·2 months [IQR 6·4–17·0]), median overall survival was 15·6 months (95% CI 13·9–20·0) in the experimental group versus 10·9 months (9·5–12·6) in the control group (HR 0·66 [95% CI 0·55–0·80]). The most common grade 3–4 treatment-related adverse events were neutropenia (in 24 [7%] patients in the experimental group vs 32 [9%] in the control group), anaemia (21 [6%] vs 50 [14%]), diarrhoea (14 [4%] vs two [1%]), increased lipase (22 [6%] vs three [1%]), and asthenia (tjree [1%] vs eight [2%]). Serious treatment-related adverse events of any grade occurred in 106 (30%) patients in the experimental group and 62 (18%) in the control group. Seven (2%) deaths in the experimental group (acute kidney failure, diarrhoea, hepatotoxicity, hepatitis, pneumonitis, sepsis with acute renal insufficiency, and thrombocytopenia; one patient each) and six