ABSTRACT
BACKGROUND: Endoscopic submucosal dissection (ESD) is a reliable method to resect early esophageal cancer. Esophageal stricture is one of the major complications after ESD of the esophagus. Steroid prophylaxis for esophageal strictures, particularly local injection of triamcinolone acetonide (TA), is a relatively effective method to prevent esophageal strictures. However, even with steroid prophylaxis, stenosis still occurs in up to 45% of patients. Predicting the risk of stenosis formation after local TA injection would enable additional interventions in risky patients. AIM: To identify the predictors of esophageal strictures after steroids application. METHODS: Patients who underwent esophageal ESD and steroid prophylaxis and who were comprehensively assessed for lesion- and ESD-related factors at Southeast University Affiliated Zhongda Hospital between February 2018 and March 2023 were included in the study. The univariate and multivariate regression analyses were conducted to identify the predictors of stricture among patients undergoing steroid prophylaxis. RESULTS: A total of 120 patients were included in the analysis. In the oral prednisone and oral prednisone combined with local tretinoin injection groups, the stenosis rates were 44/53 (83.0%) and 56/67 (83.6%), respectively. Among them, univariate analysis showed that the lesion circumference (P = 0.01) and submucosal injection solution (P = 0.04) showed significant correlation with the risk of stenosis formation. Logistic regression analyses were then performed using predictors that were significant in the univariate analyses and combined with known predictors from previous reports, such as additional chemoradiotherapy and tumor location. We identified a lesion circumference < 5/6 (OR = 0.19; P = 0.02) and submucosal injection of sodium hyaluronate (OR = 0.15; P = 0.03) as independent predictors of on esophageal stricture formation. CONCLUSION: Steroid prophylaxis effectively prevents stenosis. Moreover, the lesion circumference and submucosal injection of sodium hyaluronate were independent predictors of esophageal strictures. Additional interventions should be considered in high-risk patients.
ABSTRACT
OBJECTIVE: To develop binary and quaternary classification prediction models in patients with severe acute pancreatitis (SAP) using machine learning methods, so that doctors can evaluate the risk of patients with acute respiratory distress syndrome (ARDS) and severe ARDS at an early stage. METHODS: A retrospective study was conducted on SAP patients hospitalized in our hospital from August 2017 to August 2022. Logical Regression (LR), Random Forest (RF), Support Vector Machine (SVM), Decision Tree (DT), and eXtreme Gradient Boosting (XGB) were used to build the binary classification prediction model of ARDS. Shapley Additive explanations (SHAP) values were used to interpret the machine learning model, and the model was optimized according to the interpretability results of SHAP values. Combined with the optimized characteristic variables, four-class classification models, including RF, SVM, DT, XGB, and Artificial Neural Network (ANN), were constructed to predict mild, moderate, and severe ARDS, and the prediction effects of each model were compared. RESULTS: The XGB model showed the best effect (AUC = 0.84) in the prediction of binary classification (ARDS or non-ARDS). According to SHAP values, the prediction model of ARDS severity was constructed with four characteristic variables (PaO2/FiO2, APACHE II, SOFA, AMY). Among them, the overall prediction accuracy of ANN is 86%, which is the best. CONCLUSIONS: Machine learning has a good effect in predicting the occurrence and severity of ARDS in SAP patients. It can also provide a valuable tool for doctors to make clinical decisions.