Computer-aided Detection of Colorectal Polyps

Development and Validation of a New Artificial Intelligence System for Automated Detection of Colorectal Polyps During Colonoscopy

In this observational pilot study, we assess the diagnostic performance of an artificial intelligence sytem for automated detection of colorectal polyps.

During standard colonoscopy, a substantial number of colorectal polyps can be missend. As shown in a recent meta-analysis, miss rates for adenomas can reach up to 26%. In this study, it is tested whether an artificial intelligence system that highlights colorectal polyps during screening or surveillance colonoscopy in real time can lead to an increased detection of colorectal polyps during the examination.

Inclusion Criteria:

• Screening or surveillance colonoscopy

Exclusion Criteria:

• known or suspected inflammatory bowel disease
• uncontrolled coagulopathy
• known polyps or referral for polypectomy