Identification of Patients for Adjuvant Therapy After Resection of Carcinoma of the Extrahepatic Bile Ducts: A Propensity Score-Matched Analysis

Resectability rates for extrahepatic cholangiocarcinoma have increased over time, but long-term survival after resection alone with curative intent remains poor. Recent series suggest improved survival with adjuvant therapy. Patient subsets benefiting most from adjuvant therapy have not been clearly defined. Patients with extrahepatic cholangiocarcinoma who underwent resection with curative intent and received adjuvant therapy (chemotherapy ± radiotherapy) or surgery alone (SA) were identified in the U.S. National Cancer Data Base (2004–2014). Cox regression identified covariates associated with overall survival (OS). Adjuvant therapy and SA cohorts were matched (1:1) by propensity scores based on the survival hazard in Cox modeling. Overall survival was compared by Kaplan–Meier estimates. Of 4872 patients, adjuvant chemotherapy was used frequently for 2416 (49.6%), often in conjunction with radiotherapy (RT) (n = 1555, 64.4%). Adjuvant chemotherapy with or without RT was used increasingly for cases with higher T classification [reference: T1–2; T3: 1.36; 95% confidence interval (CI), 1.19–1.55; T4: 1.77; 95% CI 1.38–2.26], nodal positivity [odds ratio (OR), 1.26; 95% CI 1.01–1.56], lymphovascular invasion (OR 1.21; 95% CI 1.01–1.46), or margin-positive resection (OR 1.85; 95% CI 1.61–2.12), and was associated with significant improvements in OS for each high-risk subset in the propensity score-matched cohort. Adjuvant therapy was associated with improved median OS for hilar tumors (40.0 vs 30.6 months; p = 0.025) but not distal tumors (33.0 vs 30.3 months; p = 0.123). Chemoradiotherapy was associated with superior outcomes compared with chemotherapy alone in the subset of margin-positive resection [hazard ratio (HR), 0.63; 95% CI 0.42–0.94]. Adjuvant multimodality therapy is associated with improved survival for patients with resected extrahepatic cholangiocarcinoma and high-risk features.