Background
Phase III double-blind RCT, N=685 patients with previously untreated unresectable locally advanced, recurrent, or metastatic biliary tract cancer (intrahepatic CCA, extrahepatic CCA, gallbladder cancer, ampullary cancer). Stratified by disease status (unresectable vs recurrent), tumor origin, and geographic region. Enrolled 2018–2020 across 105 sites globally.
Interventions and follow up
Arm A: Durvalumab 1500 mg q3wk + gemcitabine 1000 mg/m² + cisplatin 25 mg/m² (days 1 and 8 q3wk) × up to 8 cycles, then durvalumab 1500 mg q4wk maintenance until progression
Arm B: Placebo + gemcitabine + cisplatin × up to 8 cycles, then placebo maintenance
Primary endpoint: Overall survival (OS)
Median follow-up: 24.1 months
Arm B: Placebo + gemcitabine + cisplatin × up to 8 cycles, then placebo maintenance
Primary endpoint: Overall survival (OS)
Median follow-up: 24.1 months
Results
OS: 12.9 vs 11.3 months, HR 0.80 (95% CI 0.66–0.97), P=.021
PFS: 7.2 vs 5.7 months, HR 0.75 (95% CI 0.63–0.89), P<.001
24-month OS rate: 24.9% vs 10.4%
PFS: 7.2 vs 5.7 months, HR 0.75 (95% CI 0.63–0.89), P<.001
24-month OS rate: 24.9% vs 10.4%
Adverse events
Overall: Grade ≥3 AEs 75.7% vs 77.8% (Arm A vs B); treatment discontinuation due to AE 8.9% vs 7.0%.
Immune-mediated: Any grade 12.7% vs 4.7%; grade ≥3: 4.7% vs 0.3%. Most common: hypothyroidism, hepatitis, rash.
Immune-mediated: Any grade 12.7% vs 4.7%; grade ≥3: 4.7% vs 0.3%. Most common: hypothyroidism, hepatitis, rash.
Conclusions
Durvalumab plus gemcitabine-cisplatin significantly improved OS compared to gemcitabine-cisplatin alone (HR 0.80, P=.021) in previously untreated advanced biliary tract cancer, with a striking improvement in 24-month OS rate (24.9% vs 10.4%). This was the first positive immunotherapy trial in BTC and established a new standard of care.
Key Limitations
Absolute OS improvement is modest (1.6 months median); survival advantage becomes more pronounced at landmark analyses (24-mo OS rate nearly doubled). PD-L1 status was not prospectively validated as a predictive biomarker — benefit appeared across PD-L1 subgroups, limiting biomarker-guided patient selection. Trial included ampullary cancers alongside cholangiocarcinoma and gallbladder cancer (heterogeneous). The maintenance phase (single-agent durvalumab post-chemotherapy) complicates interpretation of relative contributions of the chemo+IO combination vs. IO maintenance. Concurrent KEYNOTE-966 trial (pembrolizumab+gem/cis) also positive — no head-to-head comparison exists between these two regimens. IDH1-mutated, FGFR2-rearranged, and HER2-amplified patients may benefit from targeted therapies rather than or after immunotherapy-based first-line; biomarker interplay not addressed.
Clinical Context
FDA approved durvalumab (Imfinzi) + gemcitabine+cisplatin for unresectable/metastatic BTC in September 2022 — the first immunotherapy approval in biliary cancer. Concurrently, KEYNOTE-966 (pembrolizumab+gem/cis; OS HR 0.83) was published in Lancet 2023 and also FDA-approved. ESMO guidelines endorse both as preferred first-line options. In clinical practice, durvalumab+gem/cis and pembrolizumab+gem/cis are considered interchangeable; choice is often based on institutional preference and access. Gem+cis alone (ABC-02 era) is no longer the preferred first-line standard in most guidelines.