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

Area of Science:

• Medical Training
• Artificial Intelligence in Medicine
• Gastroenterology

Background:

• Esophagogastroduodenoscopy (EGD) is crucial for diagnosing upper gastrointestinal issues.
• Traditional EGD training for new endoscopists faces challenges in efficiency and consistency.
• This study investigates an AI-assisted system, EndoAdd, to enhance EGD training.

Purpose of the Study:

• To evaluate the effectiveness of the AI-assisted EndoAdd system in improving novice endoscopist performance.
• To assess the impact of real-time AI feedback on reducing blind spots during EGD procedures.
• To determine if AI assistance can improve lesion detection and photodocumentation quality in EGD training.

Main Methods:

• A randomized controlled trial involving eight novice endoscopists was conducted.
• Participants were divided into an AI-assisted group (EndoAdd) and a control group (traditional training).
• The EndoAdd system provided real-time feedback on blind spots and photodocumentation quality.

Main Results:

• The EndoAdd system demonstrated high accuracy (98.0%) and AUC (0.984).
• The AI-assisted group showed a significant reduction in blind spots and improved photodocumentation.
• Lesion detection rates were higher in the EndoAdd group, with no increase in examination time.

Conclusions:

• The AI-assisted EndoAdd system effectively enhanced novice endoscopist performance in EGD.
• EndoAdd significantly reduced blind spots and improved lesion detection during training.
• AI systems hold promise for accelerating endoscopy education and improving procedural outcomes.