Abstract:Objective To analyze the risk factors of the difficulty of endoscopic resection (ER) in patients with gastric gastrointestinal stromal tumors (gGISTs), and to establish and verify a predictive model of the difficulty of RE. Methods A total of 318 patients with ER-gGISTs in our hospital from January 2013 to December 2023 were retrospectively analyzed. The patients were divided into a training group (n = 222) and a validation group (n = 96) in a ratio of 7 ∶ 3. The patients were further divided into a difficult group and a non-difficult group according to surgical difficulty. Clinical data were analyzed and compared between the two groups. Lasso regression and multivariate logistic regression were performed in the training group to screen independent risk factors for surgical difficulty and construct a nomogram. The nomogram was then validated. Results After Lasso regression and multivariate logistic regression screening, the results showed that tumor growth pattern, tumor diameter, surgical experience, neutrophil to lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR) were independent factors influencing the difficulty of ER of gGISTs (P<0. 05). Based on the above influencing factors, a multi-factor predictive nomogram model of the difficulty of ER was constructed. The AUCs of the predictive nomogram model were 0. 933 (95% CI:0. 901 ~ 0. 965) and 0. 861 (95% CI: 0. 775 ~ 0. 947) in the training group and the validation group, respectively. The calibration curves showed a good fit between the predicted and measured values. The clinical decision curve indicated a net return rate within the threshold range of 0. 1 ~ 0. 9. Conclusions Tumor growth pattern, tumor diameter, operator experience, NLR and PLR are correlated with the difficulty of ER of gGISTs. The constructed nomogram model has a high discrimination and calibration in predicting the difficulty of the ER of gGISTs. It is helpful for surgical operation and postoperative management.