Loading...
There is considerable interest in computer-aided detection (CADe) systems to improve detection of colorectal neoplasms during colonoscopy. While several systems are available, there are no data from large randomized controlled trials (RCTs) regarding their efficacy in real-time colonoscopy.
In this single-center study in China, over 1000 patients (mean age, 50 years) were randomized to CADe or standard colonoscopy. The CADe system was developed using deep learning architecture and had favorable performance characteristics and processing speed in preliminary testing. Eight endoscopists ranging in seniority level performed the procedures. Most patients underwent colonoscopy for symptomatic indications, and the withdrawal time in procedures wit…