Video Grading of the Pancreatico-Jejunostomy in Robotic Pancreatoduodenectomy to Determine the Learning Curve and the Risk of Postoperative Pancreatic Fistula

Abstract

Introduction: Video grading may be used to assess the surgical learning curve and could predict complications. Predicting the risk of postoperative pancreatic fistula (POPF) could guide treatment, quality control, or be used for benchmarking however a cut-off variable regarding surgical expertise is lacking. We graded the pancreatico-jejunostomy of robotic pancreatoduodenectomy (RPD) to assess learning curve and the risk of POPF. Methods: We included patients who underwent RPD for all indications from a recently completed multicenter training program in RPD (LAELAPS-3). Technical performance was graded by two blinded graders using the Objective Structured Assessment of Technical Skills (OSATS) score, ranging from 12-60. Cumulative sum (CUSUM) analysis was used to assess a learning curve inflection point. Logistic regression was used to determine the cut-off OSATS score to predict the increased chance of POPF grade B/C. Results: Videos of 153 patients from four centers were assessed. POPF B/C occurred in 35 patients (22.9%). Based on uaFRS criteria, 72 (47.1%) patients had intermediate or high-risk anastomoses. Median OSATS was 48 [interquartile range 41–52]. CUSUM analysis of OSATS found an inflection point after 33 procedures. An OSATS score below 49 was associated with a higher risk of POPF, OR 2.27 (1.04–4.99, p=0.041). Conclusion: This multicenter study demonstrates that video grading of the pancreatico-jejunostomy can determine both the learning curve and the risk of POPF after RPD.