Immunotherapies have shown clinical activity in patients with advanced biliary tract cancer, for which outcomes remain poor despite standard of care treatment with gemcitabine and cisplatin. We aimed to evaluate gemcitabine and cisplatin plus durvalumab with or without tremelimumab as first-line treatment in patients with advanced biliary tract cancer. This open-label, single-centre, phase 2 study was conducted at Seoul National University Hospital. Eligible patients were treatment-naïve adults aged 18 years or older with histologically proven unresectable or recurrent biliary tract cancer, at least one measurable lesion based on the Response Evaluation Criteria in Solid Tumors (version 1.1), an Eastern Cooperative Oncology Group performance status of 0 or 1, life expectancy of 12 weeks or longer, and adequate healthy organ and bone marrow function. Initially, all patients received one 3-week cycle of gemcitabine (1000 mg/m2) and cisplatin (25 mg/m2) on day 1 and 8 followed by gemcitabine and cisplatin plus durvalumab (1120 mg) and tremelimumab (75 mg) on day 1 of each cycle, starting with the second cycle (chemotherapy followed by chemotherapy plus durvalumab and tremelimumab group). Following protocol amendment, patients were recruited to receive gemcitabine and cisplatin plus durvalumab, starting on day 1 of the first cycle (chemotherapy plus durvalumab group) or gemcitabine and cisplatin plus durvalumab and tremelimumab also from day 1 of the first cycle (chemotherapy plus durvalumab and tremelimumab group) in parallel and allocated using a random block method. Assessors and patients were not masked to the treatment group. The primary endpoint was objective response rate, assessed in the efficacy population (ie, patients who were treated at least until the first tumour response assessment). This study is registered with ClinicalTrials.gov, NCT03046862 (active). Between March 2, 2017, and Feb 13, 2020, 128 patients were enrolled (32 in the chemotherapy followed by chemotherapy plus durvalumab and tremelimumab group, 49 in the chemotherapy plus durvalumab group, and 47 in the chemotherapy plus durvalumab and tremelimumab group). Four patients (two in the chemotherapy followed by chemotherapy plus durvalumab and tremelimumab group and two in the chemotherapy plus durvalumab group) were excluded and 124 were evaluable for tumour response. The median duration of follow-up was 48·2 months (IQR 41·5-49·4) for the chemotherapy followed by chemotherapy plus durvalumab and tremelimumab group, 26·6 months (19·0-27·9) for the chemotherapy plus durvalumab group, and 24·2 months (20·7-31·7) for the chemotherapy plus durvalumab and tremelimumab group. 82 (66%) of 124 patients achieved an objective response (15 [50%] of 30 in the chemotherapy followed by chemotherapy plus durvalumab and tremelimumab group, 34 [72%] of 47 in the chemotherapy plus durvalumab group, and 33 [70%] of 47 in the chemotherapy plus durvalumab and tremelimumab group). The most common grade 3 and 4 adverse events were decreased neutrophil count (67 [53%] of 126), anaemia (50 [40%]), and decreased platelet count (24 [19%]), with no unexpected safety events. No adverse events leading to discontinuation or death occurred. Gemcitabine and cisplatin plus immunotherapy showed promising efficacy and acceptable safety in patients with biliary tract cancer. Gemcitabine and cisplatin plus durvalumab are being evaluated in the phase 3, TOPAZ-1 study (NCT03875235) as first-line treatment in patients with advanced biliary tract cancer. AstraZeneca; National Research Foundation of Korea (Grant No. 2021R1A2C2007430).