Adoption of Robotic Pancreatoduodenectomy: Initial Experience With the First 100 Consecutive Cases

Despite the growing adoption of robotic pancreatoduodenectomy (R-PD), there remains a paucity of literature examining the learning curve associated with R-PD and the comparison of outcomes between early and later phases of institutional experience in R-PD. Accordingly, this study analyses 100 consecutive R-PD cases at a tertiary centre in Singapore, comparing outcomes across the initial and subsequent 50 cases to assess the effect of the learning curve, and benchmarking low-risk patients against international open PD (O-PD) standards.

This study is a post hoc review of 210 consecutive patients who underwent minimally-invasive PD (MI-PD) between 2016 and 2025, of which 100 patients who underwent MI-PD with robotic assistance (R-PD) were included. Baseline characteristics and perioperative outcomes were identified, and univariate analyses were conducted to evaluate differences between the first 50 cases (Group 1) and subsequent 50 cases (Group 2). Benchmarking was also performed for low-risk cases against known O-PD cut-offs.

Comparison between Groups 1 and 2 demonstrated a significant decrease in major postoperative complications (46% vs. 24%; p = 0.04), clinically significant post-operative pancreatic fistula (Grade B/C POPF) (39.1% vs. 18.2%; p = 0.05), POPF graded Clavien-Dindo Grade 3 and above (21.7% vs. 4.5%; p = 0.03) and delayed gastric emptying (46% vs. 14%; p = 0.00). Following risk stratification, Group 2 achieved ten of 11 benchmarks, compared with five of 11 in Group 1.

This study supports the safe and effective implementation of R-PD within a high-volume hepatopancreatobiliary programme. Outcomes improved substantially after 50 cases, with low-risk patients in Group 2 achieving nearly all international O-PD benchmarks.