External validation of MAGIC-D model in Taiwanese patients with advanced biliary tract cancer receiving durvalumab and chemotherapy

BACKGROUND: The MAGIC-D score is a prognostic index developed to stratify outcomes in patients with advanced biliary tract cancer (BTC) treated with gemcitabine-cisplatin-durvalumab (GCD). We conducted an external validation study to evaluate the prognostic performance of MAGIC-D in a real-world Taiwanese cohort. MATERIAL AND METHODS: We reviewed patients with unresectable or metastatic BTC receiving first-line GCD at four centers between 2021 and 2025. Baseline clinical, laboratory, treatment, and safety data were collected. Patients were classified into low-, intermediate-, and high-risk groups using MAGIC-D. Overall survival (OS) and progression-free survival (PFS) were estimated using Kaplan-Meier analysis with hazard ratios (HRs) from Cox regression. Model discrimination was assessed using area under the curve (AUC). RESULTS: Of 172 patients, MAGIC-D classified 48 (27.9%) as low-risk, 80 (46.5%) intermediate-risk, and 44 (25.6%) as high-risk. Median OS and PFS were 13.8 and 5.1 months, respectively for the entire cohort. Median OS by risk group was 23.2 months (low), 16.3 months (intermediate; HR 1.95, 95% CI 1.04-3.67), and 4.9 months (high; HR 7.12, 95% CI 3.70-13.7). PFS times were 6.7, 5.1, and 2.2 months, respectively. MAGIC-D achieved AUCs of 0.755 and 0.749 for 6- and 12-months survival prediction, respectively, outperforming neutrophil-to-lymphocyte ratio and other baseline factors. CONCLUSION: In this first external validation in Asian BTC patients, MAGIC-D effectively stratified patients by survival and treatment response. MAGIC-D may aid in risk-adapted patient counseling, follow-up planning, and clinical trial stratification for GCD-treated BTC.