Published online Sep 15, 2022. doi: 10.4251/wjgo.v14.i9.1823
Peer-review started: April 19, 2022
First decision: May 11, 2022
Revised: May 14, 2022
Accepted: August 14, 2022
Article in press: August 14, 2022
Published online: September 15, 2022
With the development of endoscopic techniques, endoscopic submucosal dissection (ESD) has been widely used in the treatment of early gastric cancer (EGC); however, due to the wide range of ESD peeling, deep lesion peeling, difficult operations, and relatively high risk of complications such as bleeding and perforation, a personal predictive model of the outcome is necessary.
A personalized and effective prediction method of the outcomes of ESD for EGC is urgently needed in clinical practice.
This study aimed to build a personalized prediction model that may provide a theoretical basis for the prevention of ulcers among EGC patients after ESD.
A total of 196 EGC patients who received ESD treatment in our hospital from March 2019 to March 2021 were enrolled in our study. The general information of the patients, pathological features and endoscopic features were analyzed, and multivariate logistic regression analysis was performed to evaluate their predictive value.
After LASSO regression analysis and validation, clopidogrel medication history, lesion diameter, convergent folds, and mucosal discoloration were the 4 independent variables that predicted postoperative ulceration. Receiver operating characteristic curve analysis showed that the AUC of the risk prediction model for ulcers after ESD in patients with EGC was 0.916 (95%CI 0.865-0.967). Compared to each of the four indicators alone, their combined prediction model should have significantly increased accuracy for the prediction of ulcer occurrence after ESD for EGC patients.
A LASSO regression-based ulcer risk prediction model that included clopidogrel medication history, lesion diameter, convergent folds, and mucosal discoloration was built for EGC.
A large sample size should be used to validate the prediction model in future studies.