Résumé
Objective: To determine whether extending resections to trisectionectomies (TS) improves long-term outcomes in Bismuth-Corlette (BC) type II-III perihilar cholangiocarcinomas (PHC). Background: Extended (segment 1) hemihepatectomy (EH) is the standard surgical approach for BC type II-III PHC. Methods: This nationwide French retrospective study compared patients undergoing right/left-sided TS and EH for BC II-III PHC between 2010 and 2020. The primary endpoint was overall survival (OS) with secondary endpoints including disease-free survival, R0-resection rate and postoperative outcomes. Propensity score and inverse probability treatment weighting were applied to balance confounders. Kaplan-Meier curves and log-rank tests were used for survival analysis, with adjusted hazard ratios (aHRs) derived from mixed-effect multivariable Cox models clustered within hospitals. Results: Among 443 patients (median age: 66, 57.1% men), 286 (64.6%) underwent TS and 157 (35.4%) EH. TS achieved more frequently R0 resection (83.9% vs 75.8%, P = 0.037), had more frequent grade B-C liver failure (23.4% vs 11.3%, P = 0.002) but similar 90-day severe morbidity (Clavien-Dindo >= 3) and mortality rates with EH (51.6% vs 50.7%, P = 0.85 and 12.9% vs 14.0%, P = 0.55). TS significantly improved OS and disease-free survival [52.4 months (42.5-104.5) vs 31 months (24.3-42.3), P < 0.001 and 39.9 months (26.2-66.2) vs 24.3 months (20.7-36.3), P = 0.01] and was independently associated with improved OS in the original cohort [aHR: 0.63 (0.46-0.88), P = 0.007], inverse probability treatment weighting pseudo-population [aHR: 0.63 (0.41-0.95), P = 0.029] and across different BC subtypes and side of resection. Conclusions: TS improves oncological outcomes compared with EH in BC II-III PHC resection and should be strongly considered when feasible.