Evaluating the learning curve of robotic radical antegrade modular pancreatosplenectomy: A retrospective cohort study

Radical antegrade modular pancreatosplenectomy (RAMPS), a new surgical approach for pancreatic ductal adenocarcinoma of the body and tail, has become increasingly accepted and performed in recent years. Robotic surgery has advantages over open and laparoscopic surgeries in terms of surgical vision and instrument flexibility. However, the lack of comprehension of the learning curve has limited its generalization. This study aimed to evaluate the learning curve of robotic posterior RAMPS.Patients who underwent robotic posterior RAMPS between February 2017 and April 2021 at our institution were included in this study. Data on patient characteristics, perioperative outcomes, and pathological outcomes were summarized and analyzed. The cumulative sum (CUSUM) method was used to assess the learning curve and inflection points based on operation time and estimated blood loss.One hundred consecutive patients who underwent robotic posterior RAMPS were enrolled. The median operation time was 235.0 (interquartile range [IQR], 210.0-270.0) min, and the estimated blood loss was 210.0 (IQR, 165.0-245.0) mL. The grade 3/4 Clavien-Dindo complication rate was 8% (8/100). According to the CUSUM plot, the inflection points of the learning curve were 25 and 65 cases, dividing the case series into the learning (1-25 cases), plateau (26-65 cases), and maturation (66-100 cases) phases. The operation time was relatively high in the learning phase, reached a plateau between 25 and 65 cases (270.0 min vs. 220.0 min, p < 0.01), and decreased significantly in the maturation phase (p < 0.01). Estimated blood loss improved in the maturation phase compared to the learning phase (150.0 vs. 245.0 mL, p < 0.01). No significant differences in conversion rate, complications, or mortality were observed among the three phases.The inflection points of the learning and plateau phases were the 25th and 65th cases, respectively. Robotic RAMPS is safe and feasible even in the learning phase.