Radboudumc datasets concerning computer-aided detection in colonoscopy

This collection contains the pseudonymized datasets underlying two international multicenter studies evaluating computer-aided detection (CADe) in colonoscopy. The studies assessed the effect of CADe on polyp detection and colonoscopy quality indicators in patients undergoing diagnostic, screening, and surveillance colonoscopy.

The first study is a randomized controlled trial evaluating CADe-assisted colonoscopy compared to conventional colonoscopy in an everyday clinical setting (Discovery II). The second study evaluated the effect of combining a mucosal exposure device (MED) with CADe on detection outcomes in a multicenter cohort compared to the historical CADe controls (Discovery III).

The collection includes colonoscopy-level datasets for both studies, along with SPSS and R scripts used for data preparation and statistical analyses. Variables cover patient demographics, colonoscopy indication, procedural characteristics, polyp detection and histopathology outcomes, and CADe-related detection parameters. Detailed study design and methods are described in the associated publications.