Artificial intelligence to improve polyp detection and screening time in colon capsule endoscopy

Area of Science:

• Gastroenterology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Colon capsule endoscopy (CCE) offers a minimally invasive alternative to traditional colonoscopy.
• Reviewing CCE videos is time-consuming and requires expert interpretation to detect polyps and other abnormalities.
• Efficient polyp detection in CCE is crucial for timely diagnosis and patient management.

Purpose of the Study:

• To develop and evaluate an AI-Tool, a web application utilizing a Convolutional Neural Network (CNN), to assist in CCE video analysis.
• To assess the impact of the AI-Tool on the efficiency and accuracy of polyp detection compared to conventional methods.
• To prioritize images with a high probability of containing polyps for expedited review.

Main Methods:

• Development of a multi-platform web application, AI-Tool, integrating a CNN for polyp detection in CCE videos.
• Comparative study involving 3 experts reviewing 18 CCE videos using both the AI-Tool and traditional RAPID Reader Software v9.0.
• Quantitative analysis of reviewing time and polyp detection sensitivity between the two methods.

Main Results:

• The AI-Tool reduced CCE video reviewing time by a factor of 6 compared to the classical linear review method.
• Polyp detection sensitivity increased from 81.08% with the traditional method to 87.80% when using the AI-Tool.
• The AI-Tool effectively prioritized images with a high likelihood of containing polyps.

Conclusions:

• The AI-Tool significantly enhances the efficiency and accuracy of CCE interpretation.
• AI-assisted review of CCE videos can lead to faster diagnosis and potentially improve patient outcomes.
• This AI-Tool represents a valuable advancement in gastrointestinal diagnostics, optimizing the use of expert time.