Background: Metastatic leiomyosarcomas have a poor prognosis, and currently doxorubicin alone is used as the standard first-line treatment. Doxorubicin combined with trabectedin has shown promising results in phase 1 and 2 studies. We aimed to identify and compare the progression-free survival of patients with metastatic or unresectable uterine or soft tissue leiomyosarcoma treated with doxorubicin and trabectedin combined as first-line therapy versus doxorubicin alone in a phase 3 trial. Methods: LMS-04 was a randomised, multicentre, open-label, superiority phase 3 trial, which included patients from 20 centres of the French Sarcoma Group (anticancer centers or hospitals with an oncological unit) in France. Eligible patients were aged 18 years or older, had an Eastern Cooperative Oncology Group performance status of 0-1, and had metastatic or relapsed unresectable leiomyosarcomas that had not previously been treated with chemotherapy. Patients were randomly assigned (1:1), by means of an interactive web response system (permuted blocks of different sizes from two to six), to receive either intravenous doxorubicin alone (75 mg/m2) once every 3 weeks for up to six cycles or of intravenous doxorubicin (60 mg/m2) plus intravenous trabectedin (1·1 mg/m2) once every 3 weeks up to six cycles followed by maintenance with trabectedin alone. Surgery for residual disease was allowed in both groups after six cycles of treatment. Randomisation was stratified by tumour location (uterine vs soft tissue) and disease (locally advanced vs metastatic). The primary endpoint was progression-free survival assessed by blinded independent central review and according to Response Evaluation Criteria in Solid Tumours 1.1 criteria. Efficacy analyses were performed on all randomly assigned patients, based on the intention-to-treat principle. The safety population included all randomly assigned patients who received at least one cycle of treatment. This trial is registered with ClinicalTrials.gov, NCT02997358, and is closed to enrolment. Findings: Between Jan 18, 2017, and March 21, 2019, 150 patients were enrolled (67 with uterine leiomyosarcomas and 83 with soft tissue leiomyosarcomas) and included in the intention-to-treat population: 76 in the doxorubicin alone group and 74 in the doxorubicin plus trabectedin group. The median duration of follow-up was 36·9 months (IQR 30·0-43·2) in the doxorubicine group and 38·8 months (32·7-44·2) in the doxorubicin plus trabectedin group. Median progression-free survival was significantly longer with doxorubicin plus trabectedin versus doxorubicin alone (12·2 months [95% CI 10·1-15·6] vs 6·2 months [4·1-7·1]; adjusted hazard ratio 0·41 [95% CI 0·29-0·58]; p<0·0001). The most common grade 3-4 adverse events were neutropenia (ten [13%] of 75 patients in the doxorubicin alone group vs 59 [80%] in the doxorubicin plus trabectedin group), anaemia (four [5%] vs 23 [31%]), thrombocytopenia (0 vs 35 [47%]), and febrile neutropenia (seven [9%] vs 21 [28%]). Nine (12%) patients in the doxorubicin alone group and 15 (201%) patients in the doxorubicin plus trabectedin group has serious adverse events. There was only one treatment-related death, reported in the doxorubicin alone group (cardiac failure). Interpretation: Doxorubicin plus trabectedin in first-line therapy was found to significantly increase progression-free survival in patients with metastatic or unresectable leiomyosarcomas compared with doxorubicin alone, despite a higher but manageable toxicity, and could be considered an option for the first-line treatment of metastatic leiomyosarcomas.