Randomized comparative evaluation of endoscopic submucosal dissection self-learning software in France and Japan

Background and study aim: Endoscopic submucosal dissection (ESD) is currently the reference method to achieve an en bloc resection for large lesions; however, the technique is difficult and risky, with a long learning curve. In order to reduce the morbidity, training courses that use animal models are recommended. Recently, self-learning software has been developed to assist students in their training. The aim of this study was to evaluate the impact of this tool on the ESD learning curve. Methods: A prospective, randomized, comparative study enrolled 39 students who were experienced in interventional endoscopy. Each student was randomized to one of two groups and performed 30 ESDs of 30 mm standardized lesions in a bovine colon model. The software group used the self-learning software whereas the control group only observed an ESD procedure video. The primary outcome was the rate of successful ESD procedures, defined as complete en bloc resection without any perforation and performed in less than 75 minutes. Results: A total of 39 students performed 1170 ESDs. Success was achieved in 404 (70.9 %) in the software group and 367 (61.2 %) in the control group (P = 0.03). Among the successful procedures, there were no significant differences between the software and control groups in terms of perforation rate (22 [4.0 %] vs. 29 [5.1 %], respectively; P = 0.27) and mean (SD) procedure duration (34.1 [13.4] vs. 32.3 [14.0] minutes, respectively; P = 0.52). For the 30th procedure, the rate of complete resection was superior in the software group (84.2 %) compared with the control group (50.0 %; P = 0.01). Conclusion: ESD self-learning software was effective in improving the quality of resection compared with a standard teaching method using procedure videos. This result suggests the benefit of incorporating such software into teaching programs.