Background
Phase III double-blind, placebo-controlled RCT of 1069 patients with previously untreated, unresectable, locally advanced or metastatic biliary tract cancer, measurable disease per RECIST 1.1, and ECOG 0-1.
Interventions and follow up
Arm A: Pembrolizumab 200 mg IV Q3W (max 35 cycles) + gemcitabine 1000 mg/m2 IV days 1 and 8 Q3W (no max) + cisplatin 25 mg/m2 IV days 1 and 8 Q3W (max 8 cycles)
Arm B: Placebo (max 35 cycles) + gemcitabine and cisplatin (same schedule)
Primary endpoint: Overall survival
mFollow up: 25.6 mo (IQR 21.7–30.4)
Arm B: Placebo (max 35 cycles) + gemcitabine and cisplatin (same schedule)
Primary endpoint: Overall survival
mFollow up: 25.6 mo (IQR 21.7–30.4)
Results
mOS: 12.7 mo (Arm A) vs 10.9 mo (Arm B); HR 0.83, 95% CI 0.72–0.95; P=.0034
mPFS: 6.5 mo vs 5.6 mo; HR 0.87, 95% CI 0.77–0.99
ORR: 29% vs 29% (similar)
mPFS: 6.5 mo vs 5.6 mo; HR 0.87, 95% CI 0.77–0.99
ORR: 29% vs 29% (similar)
Adverse events
Grade 3-4 events (A vs B): 79% vs 75%
Treatment-related deaths (A vs B): 6% vs 9%
Immune-mediated AEs (any grade, A vs B): 22% vs 13%; grade 3-4 immune-mediated events 7% vs 4%
Treatment-related deaths (A vs B): 6% vs 9%
Immune-mediated AEs (any grade, A vs B): 22% vs 13%; grade 3-4 immune-mediated events 7% vs 4%
Conclusions
Adding pembrolizumab to gemcitabine and cisplatin significantly improved OS versus chemotherapy alone in advanced biliary tract cancer, with a manageable and consistent safety profile.
Key Limitations
Modest absolute OS gain (~1.8 mo) with no improvement in ORR; no predictive biomarker identified; benefit broadly consistent but small, raising cost-effectiveness questions.
Clinical Context
KEYNOTE-966 led to FDA (2023) approval of pembrolizumab plus gemcitabine/cisplatin for locally advanced or metastatic biliary tract cancer, alongside durvalumab/gem-cis (TOPAZ-1). ESMO endorses adding a PD-(L)1 inhibitor to gem-cis as a 1L standard. Molecular profiling (FGFR2, IDH1, HER2, MSI) remains essential for later-line targeted options.