Immunotherapy Boost for Bile Duct Cancer Survival Holds Firm After Three Years

Biliary tract cancer is one of the most unforgiving diagnoses in oncology. These malignancies, which arise in the bile ducts and gallbladder, are diagnosed in roughly 50,000 people each year in the United States and about 210,000 worldwide, yet fewer than one in twenty patients with metastatic disease survives five years. For decades the only first-line option for advanced disease was a chemotherapy doublet of gemcitabine and cisplatin, a regimen that bought time but rarely transformed outcomes. Now, extended results from the largest global phase 3 trial of a PD-1 inhibitor in this disease confirm that the survival gains achieved by adding immunotherapy to that chemotherapy backbone are not a statistical flash in the pan but a durable shift in the natural history of the illness.

The trial, known as KEYNOTE-966, enrolled 1,069 patients between October 2019 and June 2021 across 175 medical centres in the Asia-Pacific region, Europe, North America and South America. Participants had untreated, histologically confirmed unresectable locally advanced or metastatic biliary tract cancer, including extrahepatic cholangiocarcinoma, gallbladder cancer and intrahepatic cholangiocarcinoma. They were randomly assigned in equal measure to receive either pembrolizumab or a saline placebo alongside standard gemcitabine and cisplatin, with neither patients nor treating investigators knowing which arm they occupied. Pembrolizumab, an antibody that blocks the programmed cell death protein 1, or PD-1, receptor on T cells, is designed to release a molecular brake that tumours exploit to evade immune destruction.

When the protocol-specified final analysis was reported, with a median follow-up of 25.6 months, the pembrolizumab combination had delivered a statistically significant and clinically meaningful improvement in overall survival: a median of 12.7 months compared with 10.9 months for chemotherapy plus placebo, corresponding to a 17 percent reduction in the risk of death. The new report, published in eClinicalMedicine, extends that picture with a median study follow-up of 36.6 months, roughly eleven months of additional observation. At the updated data cutoff of November 2023, median overall survival remained 12.7 months in the pembrolizumab group versus 10.9 months in the placebo group, with a hazard ratio of 0.86, indicating that the survival advantage persisted essentially undiminished over time.

The longer follow-up also sharpened the portrait of long-term survivors. At 24 months, an estimated 25 percent of patients in the pembrolizumab arm were alive compared with 19 percent in the control arm; at 36 months the figures were 13 percent and 11 percent respectively. Approximately 15 percent of participants in both groups remained alive beyond three years, a striking finding in a disease where metastatic five-year survival has historically languished below 5 percent. The investigators caution that the number of patients still at risk in the tail of the survival curves is small, making precise comparisons at these late time points difficult, but the persistence of a hazard ratio favouring pembrolizumab throughout follow-up suggests the treatment benefit is genuinely preserved rather than an artefact of early curve separation.

Progression-free survival, the time until disease worsens or death, also continued to favour the immunotherapy combination, with a median of 6.5 months versus 5.6 months and a hazard ratio of 0.85. Interestingly, the objective response rate, the proportion of patients whose tumours shrank measurably, was identical between arms at 29 percent, a reminder that in immunotherapy the most meaningful gains often come not from shrinking more tumours but from controlling them for longer. That interpretation is borne out by the duration of response: among patients who did respond, responses lasted a median of 8.3 months with pembrolizumab versus 6.9 months with placebo, and at 24 months, 18 percent of responses in the pembrolizumab arm were still ongoing compared with just 7 percent in the placebo arm. The number of patients completing the full 35 cycles, roughly two years, of pembrolizumab also grew since the final analysis, while no comparable increase occurred in the placebo group.

Safety data accumulated over the extended follow-up brought no unpleasant surprises. Nearly all patients in both arms experienced some adverse event, most commonly the neutropenia and anaemia expected from cisplatin-based chemotherapy, and the rates of severe grade 3 or 4 events remained stable at around 70 percent in each group. Immune-mediated adverse events, the characteristic toxicities of checkpoint inhibitors, arose in 25 percent of pembrolizumab recipients versus 15 percent of placebo recipients, with hypothyroidism and pneumonitis the most frequent. Six additional patients in the pembrolizumab arm experienced severe immune-mediated events during the extended follow-up, including gastritis, haemolytic anaemia and single cases of colitis, hepatitis, hypophysitis and pneumonitis, but the overall profile remained comparable to the final analysis, with no new safety signals identified. Deaths attributable to treatment-related events were unchanged at 2 percent in the pembrolizumab group.

The updated results land alongside three-year data from the rival TOPAZ-1 trial, which tested a different checkpoint inhibitor, durvalumab, an antibody targeting PD-L1 rather than PD-1, on the same chemotherapy backbone. TOPAZ-1 reported a hazard ratio of 0.74 for overall survival, with medians of 12.9 versus 11.3 months, after a median follow-up of 41.3 months. The two trials differ in instructive ways: TOPAZ-1 limited gemcitabine to eight cycles whereas KEYNOTE-966 allowed it to continue until progression or intolerance, and KEYNOTE-966 was larger, with 1,069 participants compared with 685, including substantially more patients enrolled outside Asia. Notably, the survival benefit in KEYNOTE-966 was consistent in the large non-Asian subgroup, strengthening confidence that the finding generalises across diverse populations and healthcare systems.

Quality-of-life analyses, reported previously from the trial, add an important dimension: adding pembrolizumab to the chemotherapy regimen did not produce a clinically meaningful deterioration in patient-reported functioning or well-being. The authors of the new analysis acknowledge a caveat here, namely that existing quality-of-life instruments were largely designed for cytotoxic drugs and may not fully capture the distinct experience of immune-mediated toxicities, and they call for better standardisation of patient-reported outcome assessment in future biliary tract cancer trials so that survival gains can be weighed against the lived experience of treatment.

The study is not without limitations. Because the primary survival endpoint was already met at the final analysis, the updated comparisons are descriptive and exploratory, with no formal hypothesis testing or alpha allocation. The shrinking at-risk population with longer follow-up makes late survival estimates imprecise, and the biological characteristics that distinguish the long-term survivors, whether tumour molecular features, disease burden, or responses to subsequent therapies, were not specifically evaluated. The authors argue that characterising this favourable subgroup could inform future patient selection and point the way toward rational combinations with novel agents, an urgent priority given how poor the prognosis remains even with immunotherapy on board.

Nevertheless, the message for clinical practice is clear and now firmly anchored in three years of evidence: pembrolizumab plus gemcitabine and cisplatin stands as a validated standard first-line treatment for previously untreated unresectable or metastatic biliary tract cancer. Together with TOPAZ-1, the trial establishes that immune checkpoint blockade has earned a permanent place in the frontline management of one of medicine’s deadliest cancers, converting a uniformly grim prognosis into one where a meaningful minority of patients can now anticipate years, not months, of survival.

Subject of Research: Long-term phase 3 trial results of pembrolizumab combined with gemcitabine and cisplatin chemotherapy for advanced biliary tract cancer

Article Title: Pembrolizumab plus gemcitabine and cisplatin for advanced biliary tract cancer: longer-term follow-up from the randomised, double-blind, placebo-controlled, phase 3 KEYNOTE-966 trial

Article References: Ueno, M., Yoo, C., Ren, Z., Furuse, J., Kelley, R. K., Chan, S. L., Edeline, J., Klümpen, H.-J., Yau, T., Verslype, C., Ozaka, M., Bouattour, M., Park, J. O., Vogel, A., Valle, J. W., Starkopf, L., Malhotra, U., Siegel, A. B., Qin, S., … Elias, R. (2026). Pembrolizumab plus gemcitabine and cisplatin for advanced biliary tract cancer: longer-term follow-up from the randomised, double-blind, placebo-controlled, phase 3 KEYNOTE-966 trial. eClinicalMedicine, 100, Article 104213. https://doi.org/10.1016/j.eclinm.2026.104213

Image Credits: AI Generated

DOI: 10.1016/j.eclinm.2026.104213

Keywords: biliary tract cancer, cholangiocarcinoma, pembrolizumab, KEYNOTE-966, immunotherapy, PD-1 inhibitor, gemcitabine, cisplatin, phase 3 trial, overall survival, checkpoint inhibitor, oncology