Endoscopic submucosal dissection (ESD) is increasingly performed by trainees under expert supervision, and appropriate case selection is crucial for safe procedure. Unlike in Western countries, gastric ESD is widely adopted in Japan and serves as the main training model. However, the number of gastric ESD cases is expected to decline owing to the widespread implementation of Helicobacter pylori eradication. This study aimed to identify factors associated with difficult cases in gastric ESD performed by trainees and to explore implications for ESD training.

We retrospectively analyzed consecutive gastric ESD procedures performed by trainees (fewer than 60 prior ESD cases) at a single center between September 2020 and September 2025. A difficult case was defined as any procedure having at least one of the following: (1) procedure time ≥120 minutes; (2) change of operator; or (3) occurrence of intraoperative perforation. Patient characteristics, lesion location, tumor size, and procedure-related factors were compared between difficult and non-difficult cases. Univariable analysis followed by multivariable logistic regression was performed to identify factors associated with difficult cases.

A total of 299 patients with 330 lesions performed by trainees were included. 30 patients had multiple lesions. The median age was 76 (range: 46–91) years, and 203 (68%) were men. Median tumor size was 9 (IQR: 6–15) mm. 38 lesions located in the upper part, 145 in the middle part, 147 in the lower part, respectively. Median procedure time was 62 (IQR: 34–110) minutes. Tumor invasion depth was M/SM1/SM2: 295/15/4 lesions, respectively. Overall, difficult cases were 101 cases (34%). 71 cases had a procedure time ≥120 minutes, 63 required an operator change, and 7 developed intraoperative perforation. Compared with non-difficult cases, difficult cases more frequently involved lesions in the upper part (24% vs 6%) and had significantly larger tumor size (10 mm vs 8 mm). In univariable analysis, location of the upper part and tumor size were associated with difficult cases, whereas there were no significant differences in age, number of cases with multiple lesions, depth of invasion, and lesion location in the cross-sectional circumference. In multivariable analysis, location of the upper part (odds ratio 4.4; 95% confidence interval 2.85–6.8; p<0.001) and tumor size (odds ratio 1.03; 95% confidence interval 1.00–1.05; p=0.047) remained independent predictors of difficult cases.

In gastric ESD performed by trainees, location of the upper part and larger tumor size were significantly associated with difficult cases. These difficult factors likely reflect technically demanding conditions, such as poor scope stability, the need for retroflexion, and extensive submucosal dissection.