Development and validation of a predictive scoring model for post-ESD coagulation syndrome in elderly with colorectal LSTs
摘要
Post-endoscopic submucosal dissection (ESD) electrocoagulation syndrome (PEECS) is a known complication in colorectal laterally spreading tumors (LSTs), but its risk factors in elderly patients remain unclear. This study aimed to develop and validate a risk-stratification scoring system for PEECS in elderly ESD patients.
MethodsA multicenter retrospective study (2020–2025) enrolled 506 elderly patients with colorectal LSTs undergoing ESD, randomly allocated to training (TC, n = 354) and validation (VC, n = 152) cohorts (7:3). Synthetic minority over-sampling technique (SMOTE) was used in the TC to identify risk factors and construct a predictive model, which was validated in the VC.
ResultsThe incidence of post-ESD PEECS was 8.1% (41 cases). After applying the SMOTE, multivariate analysis identified sex, lesion with fibrosis, and intraoperative bleeding as independent risk factors. The scoring system assigned: 1 point for female sex, 3 points for lesion with fibrosis, and 2 points for intraoperative bleeding. In the VC, the model demonstrated an area under the curve (AUC) of 0.921, with a specificity of 0.949 and an accuracy of 0.908. Following risk stratification, the low-risk group (0–4 points) showed a PEECS incidence of 10.1% in the TC and 5.1% in the VC, while the high-risk group (5–6 points) exhibited 94.5% in the TC and 53.3% in the VC.
ConclusionIn elderly patients with colorectal LSTs undergoing ESD, female sex, fibrotic lesions, and intraoperative bleeding were identified as independent predictors of PEECS. The proposed scoring system demonstrated good discriminatory ability in both cohorts and may be useful for risk stratification in this population.