Artificial Intelligence Automated Assessment of Colonoscopy Quality Metrics

Area of Science:

• Gastroenterology
• Medical Technology
• Artificial Intelligence

Background:

• Measuring colonoscopy quality in routine practice is challenging due to provider variability.
• Existing methods for assessing colonoscopy performance are difficult to implement consistently.
• There is a need for objective and reliable tools to evaluate colonoscopy quality.

Purpose of the Study:

• To introduce and evaluate an artificial intelligence tool (AI-CQ) for measuring colonoscopy quality.
• To compare the AI-CQ's assessment of quality metrics with manual measurements.
• To determine the feasibility and accuracy of using AI for colonoscopy quality assessment.

Main Methods:

• Colonoscopy procedures were recorded and analyzed using the AI-CQ tool.
• Quality metrics, including adenoma detection rate (ADR) and withdrawal time (WT), were calculated by AI and compared to manual assessments.
• The study included a large cohort of colonoscopists and procedures (18,597 total procedures).

Main Results:

• AI-CQ showed strong correlations with manual measurements for insertion time (r=0.60) and withdrawal time (r=0.91).
• AI-polyps per colonoscopy (AI-PPC) correlated well with ADR (0.54) and serrated detection rate (0.67).
• The AI-CQ accurately measured polypectomy rates, including cold snare polypectomy (84.0%).

Conclusions:

• The AI-CQ tool can accurately measure key colonoscopy quality metrics from recorded videos.
• This AI-driven approach provides a novel and feasible method for reliable colonoscopy quality assessment.
• The AI-CQ tool has the potential to improve standardization and quality improvement in colonoscopy practices.