Background and study aims We developed an e-learning program for endoscopic diagnosis of invasion depth of early gastric cancer (EGC) using a simple diagnostic criterion called non-extension sign, and the contribution of self-study quizzes to improvement of diagnostic accuracy was evaluated. Methods We conducted a prospective randomized controlled study that recruited endoscopists throughout Japan. After completing a pretest, the participants watched video lectures and undertook post-test 1. The participants were then randomly allocated to either the self-study or non-self-study group, and participants in the first group completed the self-study program that comprised 100-case quizzes. Finally, participants in both groups undertook post-test 2. The primary endpoint was the difference in post-test 2 scores between the groups. The perfect score for the tests was set as 100 points. Results A total of 423 endoscopists completed the pretest and were enrolled. Post-test 1 was completed by 415 endoscopists and 208 were allocated to the self-study group and 207 to the non-self-study group. Two hundred and four in the self-study group and 205 in the non-self-study group were included in the analysis. Video lectures improved the mean score of post-test 1 from 72 to 77 points. Participants who completed the self-study quizzes showed significantly better post-test 2 scores compared with the non-self-study group (80 vs. 76 points, respectively, P < 0.0001). Conclusions Our e-learning program showed that self-study quizzes consolidated knowledge of the non-extension sign and improved diagnostic ability of endoscopists for invasion depth of EGC.