Related Experiment Video
Updated: Apr 16, 2026
04:09
Endoscopic Injection Sclerotherapy Assisted by Cyanoacrylate and Clips for Gastroesophageal Varices
Published on: June 13, 2025
1.7K
Factors that affect visibility during endoscopic hemostasis for upper GI bleeding: a prospective study.
Sungmo Jung1, Eun Hye Kim1, Ha Yan Kim2
1Division of Gastroenterology, Department of Internal Medicine, Severance Hospital, Institute of Gastroenterology, Yonsei University College of Medicine, Seoul, South Korea.
Gastrointestinal Endoscopy
|March 16, 2015
Summary
Predicting endoscopic visibility in upper GI bleeding (UGIB) is crucial. A classification and regression tree (CART) model using nasogastric tube aspirate appearance and endoscopy time can identify patients likely to have poor visibility during emergent procedures.
Area of Science:
- Gastroenterology
- Endoscopic Procedures
- Medical Imaging
Background:
- Adequate endoscopic visibility is critical for successful hemostasis in upper gastrointestinal bleeding (UGIB).
- Factors influencing endoscopic visibility during UGIB procedures remain incompletely understood.
Purpose of the Study:
- To identify independent factors affecting endoscopic visibility in UGIB.
- To develop a predictive model for identifying patients with suspected poor visibility before emergent endoscopy.
Main Methods:
- Prospective, observational study involving 121 patients with UGIB at a tertiary care hospital.
- Analysis of endoscopic visibility scores and development of a classification and regression tree (CART) model.
- Utilized nasogastric (NG) tube aspirate appearance and endoscopy duration (EGD time) as key predictors.
Main Results:
- Endoscopic retrograde cholangiopancreatography (EGD) time and NG tube aspirate characteristics (red blood, coffee grounds) were significant independent factors affecting visibility.
- The developed CART model, using NG tube aspirate and EGD time (8.5 hours), predicted poor visibility with 71.4% sensitivity and 86.4% specificity.
- The model was trained on 70 patients and validated on the remaining cohort.
Conclusions:
- A CART model effectively predicts patients likely to experience poor visibility during emergent UGIB endoscopy.
- This predictive capability can aid in optimizing procedural planning and resource allocation for UGIB management.
