Background The existing ex vivo models of endoscopic submucosal dissection (ESD) cannot simulate intraoperative hemorrhage well. We aimed to establish an ESD training method by applying an ex vivo training model with continuous perfusion (ETM-CP). Methods Four training sessions were conducted for 25 novices under the guidance of 2 experts. Eventually, 10 novices completed ESD operations on a total of 89 patients after the training. The resection effectiveness, resection speed, complication rate, and novice performance before and after the training were compared. The data regarding the effects of the training and the model were gathered through a questionnaire survey. Results In terms of the simulation effect of the model, ETM-CP was evaluated as similar to the live pig in all aspects (P > 0.05). The questionnaire analysis revealed that the ESD theoretical knowledge, skill operation, and self-confidence of novices were improved after the training (P < 0.05). The resection time per unit area had a correlation with the number of training periods (rs = – 0.232). For novice performance, the resection time per unit area was shortened (P < 0.05). There was no difference in patient performance between the novice group and the expert group after the training in terms of en bloc resection, R0 resection, complication rate, endoscopic resection bleeding (ERB) score, muscularis propria injury (MPI) score, and resection time per unit area (P > 0.05). Conclusion The ETM-CP is effective for ESD training.

Background Linked color imaging (LCI) can improve the diagnostic rate of Helicobacter pylori (H. pylori) infection-related gastritis, and Kyoto Classification of Gastritis contributes to the diagnosis of H. pylori infection under endoscopy. The present study aims to combine LCI with Kyoto Classification of Gastritis to establish a training model for general practitioners (GPs) with a special interest in digestive diseases. Methods Gastroenterology sub-specialty training was conducted for GPs. After analysis by expert endoscopists, a simplified Kyoto Classification of Gastritis was generated to prepare the questionnaire. The questionnaire containing images in white light imaging (WLI) mode and LCI mode was used to assess the trainees' ability to determine the status of H. pylori infection by observing endoscopic mucosal manifestations before and after learning the Kyoto Classification of Gastritis. SPSS 26.0 software was used for statistical analysis. Results The analysis of expert endoscopists on the Kyoto Classification of Gastritis showed that gastroscopic mucosal manifestations including mucosal atrophy, hematin, mucosal swelling, and sticky mucus had an impact on the judgment of H. pylori infection (P < 0.05). After training, the total questionnaire score of GPs was improved (3.48 vs 4.45, P < 0.05), and there was no difference in the questionnaire completion time between GPs and standard gastroenterologists (SGs). After training, the questionnaire score of GPs based on images in WLI and LCI modes was improved. The score of the LCI mode was higher than that of the WLI mode (LCI: 1.45 vs 2.14, P < 0.05, WLI: 2.04 vs 2.31, P = 0.355). Conclusions Kyoto Classification of Gastritis combined with LCI improves the ability of GPs to diagnose H. pylori infection through endoscopic images.