Effect of artificial intelligence implementation to the latest generation 4K colonoscopy

Area of Science:

• Gastroenterology
• Medical technology
• Artificial Intelligence

Background:

• Colonoscopy is crucial for colorectal cancer (CRC) screening, with key quality indicators including adenoma detection rate (ADR), cecal intubation rate (CIR), withdrawal time (WT), and bowel preparation (Boston Bowel Preparation Scale; BBPS).
• Limitations in detecting suspicious lesions during colonoscopy highlight the need for advanced technologies.
• Next-generation endoscopes with 4K resolution and AI-powered computer-aided detection (CADe) show promise for improving colonoscopy quality.

Purpose of the Study:

• To evaluate the impact of AI-driven CADe systems integrated with the latest generation of 4K endoscopes on colonoscopy quality indicators.
• To assess changes in polyp detection rate (PDR), adenoma detection rate (ADR), advanced adenoma detection rate (AADR), and mean polyp per patient (MPP) score.

Main Methods:

• A retrospective analysis of 2,000 colonoscopies was conducted, divided into two groups: 1,000 with the Olympus Endo-Aid CADe AI system and 4K endoscope, and 1,000 without CADe.
• Key metrics assessed included PDR, ADR, AADR, and MPP.
• Comparative analysis was performed for the overall procedure and individual bowel segments, with adjustments for bowel preparation quality.

Main Results:

• Overall PDR, ADR, and AADR showed no significant differences between the AI and non-AI groups.
• The mean polyp per patient (MPP) score was significantly higher in the AI group (1.26 vs. 0.85, P<0.001).
• PDR significantly increased in the left colon (29.3% vs. 18.0%, P<0.001) with AI, and MPP was higher in both the left and right colon segments. These findings persisted after adjusting for bowel preparation quality.

Conclusions:

• While AI-CADe did not improve overall polyp detection rates, it significantly increased the number of polyps detected per patient, especially in the left colon.
• The study suggests AI's potential role in enhancing colonoscopy quality, particularly with advanced endoscopic technology.
• Further prospective randomized controlled trials using the newest generation scopes are recommended to fully elucidate AI's role in high-resolution colonoscopy.