Risk Prediction for Delayed Bleeding After Endoscopic Submucosal Dissection in Colorectal Precancerous Lesions Using
Zunan Wu1,2, Lei Yuan2,3,4, Shengjie Zhang1,2
1Department of Gastroenterology, Renmin Hospital of Wuhan University, Wuhan, Hubei, China.
Delayed bleeding after endoscopic submucosal dissection (ESD) is common. A new XGBoost model identifies risk factors like male gender, age, lesion size, procedure time, and location to predict post-ESD bleeding in colorectal precancerous lesions.
Area of Science:
- Gastroenterology
- Endoscopic Surgery
- Oncology
Background:
- Delayed bleeding is a significant complication following endoscopic submucosal dissection (ESD).
- Identifying patients at high risk for this complication is crucial for effective management.
Purpose of the Study:
- To identify risk factors for delayed bleeding after ESD in patients with colorectal precancerous lesions.
- To develop a predictive model for delayed bleeding post-ESD.
Main Methods:
- Retrospective, multicenter case-control study involving 956 patients.
- Utilized univariate and multivariate analyses to identify independent risk factors.
- Developed an XGBoost model for risk stratification.
Main Results:
- 7.5% of patients experienced delayed bleeding post-ESD.
- Independent risk factors identified: male gender, age ≥ 65, lesion size ≥ 30 mm, procedure time ≥ 120 minutes, rectal location.
- The XGBoost model differentiated low-risk (4.0% bleeding) and high-risk (13.1% bleeding) groups.
Conclusions:
- An XGBoost-based risk prediction model accurately forecasts delayed bleeding risk after colorectal ESD.
- The model aids in stratifying patients into low and high-risk categories for better clinical decision-making.
More Related Videos
03:05Author Spotlight: Advancing Early Detection and Treatment of Gastrointestinal Tumors
Published on: February 16, 2024
04:09Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
Published on: October 10, 2018
