Learning Curve of Robotic Pancreaticoduodenectomy with Portal-Superior Mesenteric Vein Resection for Pancreatic...

Background: Pancreaticoduodenectomy (PD) with portal-superior mesenteric vein (PV-SMV) resection is increasingly performed in borderline-resectable periampullary cancer. While conventional PD is the reference standard, robotic PD (RPD) may improve operative ergonomics and recovery; its performance and learning curve in PV-SMV resection remain unclear. Materials and Methods: We retrospectively reviewed consecutive patients undergoing PD with PV-SMV resection at a single tertiary center by a single surgeon (July 2016-September 2022). Twenty-seven patients met the inclusion criteria and were grouped as conventional PD (n = 14) or RPD (n = 13). To assess the learning curve, RPD cases were stratified as early (cases 1-3) versus late (cases 4-13). Primary outcomes were operative time and blood loss; secondary outcomes included 90-day morbidity/mortality, R0 margin, lymph node yield, length of stay, readmission, and overall survival. Results: Baseline characteristics were comparable between conventional PD and RPD. Median operative time was longer with RPD vs. conventional PD (624.0 [IQR 579.0-794.0] vs. 529.5 [456.5-636.5] mins; p = 0.024). Median blood loss trended lower with RPD (350.0 [200.0-1950.0] vs. 1455.0 [630.0-2940.0] mL; p = 0.254). Rates of clinically relevant complications (including POPF, DGE, and hemorrhage), R0 resection (69% vs. 64%), lymph node retrieval, length of stay, 90-day readmission, 90-day mortality, and overall survival were similar between conventional PD and RPD. Within RPD, operative time and blood loss improved from the early to late phases (794.0 → 601.5 min; 1950.0 → 275.0 mL), consistent with a learning-curve effect, though not statistically significant in this small cohort. Conclusions: In selected patients, RPD with PV-SMV resection is feasible and achieves oncologic and short-term clinical outcomes comparable to conventional PD, with evidence of efficiency gains as experience accrues. These findings support structured training and case accumulation for the safe adoption of complex robotic pancreatic surgery.