Multicenter IGoMIPS Registry-based analysis of robotic versus laparoscopic approach to left pancreatectomy for pancreatic cancer

: Robotic and laparoscopic minimally invasive left pancreatectomy are increasingly adopted for pancreatic ductal adenocarcinoma (PDAC), but comparative oncologic outcomes, especially margin status, remain uncertain. This retrospective multicenter study utilized the prospectively maintained IGoMIPS registry to identify adult patients (≥ 18 years) who underwent minimally invasive left pancreatectomy (laparoscopic, LLP, and robot-assisted, RLP) for histologically confirmed PDAC of the body and tail between September 2019 and May 2025. The primary endpoint was margin positivity (R-positive vs. R0); secondary endpoints included perioperative and pathological outcomes and exploratory overall survival (OS) and disease-free survival (DFS). Univariable and multivariable logistic regression identified predictors of R-positive resection, and Kaplan-Meier analysis compared OS and DFS between approaches. In addition to conventional regression, multiple sensitivity analyses were performed to address treatment selection and confounding. Overall, 645 patients were included, with 410 (64%) undergoing LLP and 235 (36%) RLP. Baseline characteristics were largely similar, except that neoadjuvant treatment was more common in the LLP group (22% vs.13%; p = 0.004). RLP was linked to a lower conversion rate (5% vs.19%, p < 0.001) and longer operative time (median 310 vs.260 min, p < 0.001). The R0 resection rate was higher with RLP than with LLP (81% vs.68%, p = 0.001), while tumor size, pT/pN stage, lymph node yield, lymph node ratio, and lymphovascular invasion were similar. Overall morbidity, Clavien-Dindo grade ≥3 complications, readmission, mortality, and length of stay showed no significant differences between groups. In multivariable analysis of predictors of positive margin, the robotic approach was associated with lower R-positive resection (OR 0.49, 95% CI [0.32-0.75], p = 0.001). In the overlap-weighted cohort, positive-margin rates were 30.2% for LLP and 19.2% for RLP (OR 0.55, 95% CI 0.30-0.99; p = 0.047), but the association was not reproduced in center-fixed-effect, mixed-platform-center, restricted-cohort, or era-stratified analyses. At a median follow-up of 9 months, overall survival (OS) and disease-free survival (DFS) were comparable (OS log-rank p = 0.948; DFS log-rank p = 0.765). RLP was associated with a higher crude R0 rate, and the overall overlap-weighted estimate remained borderline significant. However, this association was not robust to center-sensitive, restricted-cohort, or era-stratified analyses. The data therefore do not establish an independent oncological advantage of the robotic platform; the observed difference may reflect case mix, institutional practice, procedural complexity, and temporal adoption. Perioperative safety and exploratory survival were comparable.