NCT04487626

Development and Validation of an Artificial Intelligence System for Bowel Preparation Quality Scoring

• Status: Completed
• Phase: N/A
• Study type: Interventional
• Enrollment: 100 (actual)
• Sponsor: National Cancer Center, Korea · Other Government
• Sex: All
• Age: 20 Years – 80 Years
• Healthy volunteers: Not accepted

Summary

The purpose of this study is to develop and validate the accuracy and reliability of an artificial intelligence(AI) system for bowel preparation quality scoring based of Boston Bowel preparation Scoring(BBPS). Then evaluate whether this AI system can help endoscopists to improve the quality of colonoscopy in clinical practice.

Detailed description

Bowel preparation is one of the most important factors which decide the quality of colonoscopy. Adequate bowel preparation is essential to guarantee a clear vision of colonic mucosa, complete inspection of entire colon and improves the adenoma detection rates(ADRs). Therefore, the quality of bowel preparation should be evaluated accurately and objectively. However, current bowel preparation quality scales depend on memories and subjective decision of endoscopists. Recently, articifial intelligence system based on deep learning algorithm has been used widely in medical fields. But, few studies have been reported to evaluate the performance of AI system based on deep learning in bowel preparation quality scoring. This study aims to develop and train an AI system to assess bowel preparation quality using the BBPS, and validate the AI system to improve the quality of colonoscopy.

Conditions

• Colonic Disease

Interventions

Type	Name	Description
DEVICE	AI	After taking standard bowel preparation regimen, patients receive colonoscopy under the AI system device. During the withdrawal phase, bowel preparation quality is assessed by AI-associated scoring system. The withdrawal time is targeted at least 6min. Every detected polyp will be removed and obtained for pathological assessment.

Timeline

• Start date: 2020-07-01
• Primary completion: 2020-10-01
• Completion: 2020-12-30
• First posted: 2020-07-27
• Last updated: 2021-03-03

Locations

1 site across 1 country: South Korea