Clinical Prognostic Factors and Survival Risk Stratification for Advanced Biliary Tract Cancer Treated with Gemcitabine-Based Palliative Chemotherapy: A Real-World Retrospective Study

BACKGROUND: Although gemcitabine-based palliative chemotherapy remains widely used for advanced biliary tract cancer (BTC), practical pretreatment prognostic factors are needed. This study identified baseline prognostic factors associated with overall survival (OS) and explored a simple risk stratification approach for 12-month survival in advanced BTC patients treated with gemcitabine-based chemotherapy. METHODS: This retrospective cohort study included advanced BTC patients treated with gemcitabine-based palliative chemotherapy between 2011 and 2025. Baseline clinical and laboratory variables were collected at treatment initiation. The Kaplan-Meier method estimated OS. Univariable and multivariable Cox proportional hazards regression analyses identified prognostic factors. A post hoc exploratory risk score was developed using routinely available factors independently associated with OS. RESULTS: A total of 154 patients were included, gemcitabine plus cisplatin was administered to 95 patients, whereas 59 received gemcitabine plus carboplatin. Median OS was 9.43 months. Multivariable analysis showed that ECOG performance status ≥ 2 (adjusted HR, 5.68; 95% CI, 2.52-12.81), alkaline phosphatase (ALP) ≥ 2 × ULN (adjusted HR, 1.53; 95% CI, 1.01-2.33), and neutrophil-to-lymphocyte ratio (NLR) ≥ 3 (adjusted HR, 1.51; 95% CI, 1.03-2.20) were associated with worse OS. An exploratory score assigning one point to each factor stratified patients into low-, intermediate-, and high-risk groups. The estimated 12-month OS rates were 59.6%, 40.4%, and 10.8%, respectively. CONCLUSIONS: Poor performance status, elevated ALP, and elevated NLR were associated with worse OS in advanced BTC patients receiving gemcitabine-based palliative chemotherapy. However, the associations for ALP and NLR were modest and should be interpreted cautiously. A simple exploratory score based on routinely available factors demonstrated distinct 12-month survival across risk groups and may help inform prognostic discussions in routine practice. This approach should be considered hypothesis-generating, and external validation is warranted.