Expert-AI Collaborative Training for Novice Endoscopists: A Path to Enhanced Efficiency

Abstract

Background: Esophagogastroduodenoscopy (EGD) is essential for diagnosing upper gastrointestinal disorders. Traditional training for novice endoscopists is often inefficient and inconsistent. This study evaluates the effectiveness of an AI-assisted system (EndoAdd) in improving EGD training. Methods: In a randomized controlled trial, eight novice endoscopists were assigned to either the EndoAdd group or a control group (traditional training). The EndoAdd system provided real-time feedback on blind spots and photodocumentation. Primary outcomes were the number of blind spots, with secondary outcomes including examination time, lesion detection, and photodocumentation completeness. Results: The EndoAdd system exhibited an overall accuracy of 98.0% and a mean area under the curve (AUC) of 0.984. The EndoAdd group had significantly fewer blind spots, improved photodocumentation, and a higher lesion detection rate. Examination time was reduced without compromising diagnostic accuracy. Conclusions: The AI-assisted EndoAdd system improved novice endoscopist performance, reducing blind spots and enhancing lesion detection. AI systems like EndoAdd show potential in accelerating endoscopy training and improving procedural quality.

Keywords

• artificial intelligence
• novice endoscopists
• training