Background: In many countries, gastric cancer is not diagnosed until an advanced stage. An Internet-based elearning system to improve the ability of endoscopists to diagnose gastric cancer at an early stagewas developed and was evaluated for its effectiveness. Methods: The study was designed as a randomized controlled trial. After receiving a pre-test, participants were randomly allocated to either an e-learning or non-e-learning group. Only those in the e-learning group gained access to the e-learning system. Two months after the pre-test, both groups received a post-test. The primary endpoint was the difference between the two groups regarding the rate of improvement of their test results. Findings: 515 endoscopists from 35 countries were assessed for eligibility, and 332 were enrolled in the study, with 166 allocated to each group. Of these, 151 participants in the e-learning group and 144 in the non-elearning group were included in the analysis. The mean improvement rate (standard deviation) in the elearning and non-e-learning groups was 1·24 (0·26) and 1·00 (0·16), respectively (P b 0·001). Interpretation: This global study clearly demonstrated the efficacy of an e-learning system to expand knowledge and provide invaluable experience regarding the endoscopic detection of early gastric cancer (R000012039).