Real-world efficacy of gemcitabine plus cisplatin with immune checkpoint inhibitor and with S-1 in biliary tract cancer: A multicenter retrospective study.

550 Background: Biliary tract cancer (BTC) is an aggressive malignancy with limited treatment options and a poor prognosis. Immune checkpoint inhibitor combined with gemcitabine and cisplatin (GC + ICI) demonstrated overall survival (OS) improvement for unresectable biliary tract cancer (BTC) in the TOPAZ-1 and KEYNOTE-966 trials. In Japan, gemcitabine, cisplatin, and S-1 (GCS) also showed superiority over GC. However, the optimal first-line regimen across real-world settings remains uncertain. We conducted a multicenter retrospective study to evaluate the efficacy and safety of GC+ICI and GCS in routine practice. Methods: We retrospectively collected data from seven Japanese institutions in patients with BTC treated with GC+ICI (durvalumab or pembrolizumab) or GCS as first-line chemotherapy between April 2019 and January 2025. Inclusion criteria were as follows: an ECOG performance status of 0–2, no prior systemic therapy for biliary tract cancer, and adequate organ function. Primary endpoints were progression-free survival (PFS) and OS, with overall response rate (ORR) and disease control rate (DCR) as secondary endpoints. Survival outcomes were estimated using the Kaplan–Meier method. Safety was evaluated according to CTCAE v5.0. Results: A total of 148 patients were included: 91 in the GC+ICI group (median follow-up 15.1 months) and 57 in the GCS group (median follow-up 26.5 months). In the GC+ICI group, 92% of patients received durvalumab and 8% received pembrolizumab. Median age were 73 (40–84) and 69 (42–83) years, respectively. ECOG PS 0/1/2 were 64%/34%/2% in GC+ICI and 58%/40%/2% in GCS. Tumor sites differed substantially: intrahepatic/perihilar/extrahepatic/gallbladder/ampulla were 18%/23%/22%/33%/0% in GC+ICI and 40%/18%/9%/28%/5% in GCS. ORR was 34% with GC+ICI and 33% with GCS, while DCR was 76% 74%. Median PFS was 6.9 months (95% CI 6.0–7.3) with GC+ICI and 5.5 months (95% CI 3.5–8.3) with GCS. Median OS was 14.3 months (95% CI 10.6–16.4) for GC+ICI and 9.9 months (95% CI 8.5–13.5) for GCS. Grade 3–4 adverse events occurred in 50% and 44% of patients in the GC+ICI and GCS groups, respectively. In the GC+ICI group, immune-related adverse events were observed in 19.8% of patients, with 4.4% experiencing grade 3–4 events. Treatment discontinuation due to toxicity was observed in 12% and 9%. Conclusions: The present multicenter retrospective study demonstrated that the GC+ICI and GCS regimens achieved comparable objective response rates. There was a tendency toward improved survival in the GC+ICI group. Given the consideration of discrepancies in patient characteristics and treatment era, the findings of this study offer substantial real-world evidence to inform first-line treatment strategies for patients with unresectable BTC.