Laparoscopic versus robotic pancreaticoduodenectomy for distal cholangiocarcinoma after learning curves of surgeons: a multicenter propensity score-matched study.

Background: Pancreaticoduodenectomy (PD) is the only potentially curative treatment for distal cholangiocarcinoma (DCC). This multicenter propensity score matching (PSM) study aimed to compare the perioperative and oncological outcomes of laparoscopic PD (LPD) and robotic PD (RPD) after the learning curve of surgeons. Methods: Consecutive patients with DCC who underwent curative LPD or RPD at eight Chinese centers between January 2016 and December 2022 were included. PSM was performed to minimize selection bias. Univariate and multivariate logistic regression analyses were used to identify independent prognostic factors for textbook outcome (TO) in these patients. Results: Overall, 529 patients who underwent PD for DCC were included, of which 251 underwent LPD and 278 underwent RPD. After PSM, 227 patients were enrolled into each group. There were no significant differences in estimated blood loss (EBL), lymph node harvest, intraoperative transfusion, vascular resection, R0 resection, severe complications, readmission, 30-day mortality, or long-term survival between the two groups. However, after the learning curve, RPD had a perioperative advantage over LPD, especially in terms of operation time (270 vs. 300 min, P<0.001). Similar conclusions were drawn in the subgroup analysis. Multivariable analysis showed that comorbidities (P=0.001), main pancreatic duct (MPD) >3 mm (P=0.001), and operative time >360 min (P=0.006) were significantly associated with TO. Conclusions: After the surgeon's learning curve, the feasibility and safety of LPD and RPD for DCC patients are comparable. Randomized controlled trials (RCTs) should be performed to confirm these findings.