Artificial intelligence-enhanced video-based assessment of surgical quality for training in laparoscopic right hemicolectomy: The "Marginal Gains" pilot study.

INTRODUCTION

The study aims to propose a standardised workflow with critical views for surgical quality assessment (SQA) in laparoscopic right hemicolectomy (LRH), to disseminate it through a "Marginal Gains" course, and to evaluate its impact through artificial intelligence (AI) enhanced video-based assessment (VBA).

MATERIALS AND

METHODS

Expert colorectal surgeons proposed a protocol for SQA in LRH based on evidence and consensus. A course ("Marginal Gains") comprising remote e-learning and on-site clinical immersion was organised to disseminate the proposed approach to LRH. Videos of procedures performed by participants before and after the course were analysed by experts (SQA items) and AI (workflows). Descriptive and inferential statistic was used to study the applicability of the proposed protocol and the impact of the course.

RESULTS

A protocol with 21 SQA items over 9 phases for LRH was proposed. Four surgeons successfully completed the pilot "Marginal Gains" course. Across the 8 videos uploaded, VBA showed that the proposed SQA items were appliable in 82.7 % (139/168 items) of the cases. Three out of 4 of the participants had higher SQA scores after the course, with an overall improvement of 30 % (20.75 ± 13.2 vs 32.75 ± 2.99 points; p = 0.126). All participants performed intracorporeal anastomosis after the course, with a significant quality improvement (1.5 ± 1.73 vs 3.75 ± 0.5 points; +56 %; p = 0.046). Overall, mean operative times increased by 00:23:38 after the course (01:36:03 ± 00:10:43 vs 01:59:41 ± 00:48:02; p = 0.465).

CONCLUSION

This study advocates for a paradigm shift in surgical education and practice by proposing, piloting, and measuring the impact of a structured, step-based approach to LRH.