Effectiveness of Artificial Intelligence in Screening Esophagogastroduodenoscopy

Area of Science:

• Gastroenterology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Retrospective study evaluating the EW10-EG01 artificial intelligence (AI) application for screening esophagogastroduodenoscopy (EGD).
• AI tools are increasingly being explored to enhance diagnostic accuracy in medical procedures.

Purpose of the Study:

• To investigate the clinical utility of the EW10-EG01 AI system in improving the detection of upper gastrointestinal tumors during EGD.
• To compare the tumor detection rates between EGD performed with and without AI assistance.

Main Methods:

• Analysis of 7,655 subjects undergoing EGD for medical checkups.
• Comparison of upper gastrointestinal tumor diagnoses between EGD with (3,841 subjects) and without (3,814 subjects) the AI system.
• Evaluation of detection rates based on endoscopist experience with the AI system.

Main Results:

• EGD with AI resulted in more frequent biopsy procedures and longer examination times.
• A higher number of upper gastrointestinal tumors were detected with AI (1.02%) compared to without AI (0.63%), though not statistically significant (P=0.062).
• A significant increase in the detection rate of esophageal and gastric tumors was observed with AI (0.78%) versus without AI (0.37%) (P=0.017).

Conclusions:

• The EW10-EG01 AI system demonstrates potential usefulness in screening EGD.
• AI assistance led to a significantly higher detection rate for esophageal and gastric tumors.
• The AI system improved tumor detection across different levels of endoscopist experience.