Real-time automatic detection system increases colonoscopic polyp and adenoma detection rates: a prospective...

Area of Science:

• Gastroenterology
• Medical Imaging
• Artificial Intelligence

Background:

• Colonoscopy effectiveness in reducing colorectal cancer mortality is limited by adenoma miss rates.
• Improving adenoma detection rates (ADRs) is crucial for enhancing colonoscopy efficacy.

Purpose of the Study:

• To evaluate the impact of a deep learning-based automatic polyp detection system on polyp detection and ADR.
• To assess the system's performance in a real-time colonoscopy setting.

Main Methods:

• An open, non-blinded, prospective randomized trial comparing standard colonoscopy with AI-assisted colonoscopy.
• Patients were randomized to undergo colonoscopy with or without an AI system that provided real-time visual and auditory polyp detection alerts.
• The primary outcome measured was the adenoma detection rate (ADR).

Main Results:

• The AI system significantly increased ADR from 20.3% to 29.1% (p<0.001).
• The mean number of adenomas per patient increased from 0.31 to 0.53 (p<0.001), primarily due to a higher detection of diminutive adenomas.
• A significant increase in hyperplastic polyp detection was also observed with the AI system (114 vs 52, p<0.001).

Conclusions:

• An automatic polyp detection system using AI significantly enhances the detection of diminutive adenomas during colonoscopy.
• The AI system also led to an increased detection of hyperplastic polyps.
• The cost-benefit analysis of implementing this AI technology requires further investigation.