A real-time coach for colonoscopy

The quality of a colonoscopy depends not only on finding abnormalities, but also on how completely and carefully the examination is performed. A newly proposed mucosa touch rate measures how often the endoscope tip is too close to the intestinal wall.

The system identifies these moments in video and highlights contact regions in red, producing feedback while the procedure is underway rather than after the fact.

In prospective clinical validation reported by the authors, real-time feedback improved polyp detection, with particularly notable benefits for less experienced endoscopists. This is an early clinical study, not a replacement for training or judgment, but it shows how AI can make quality visible at the moment it matters. Of course, the home test is easier, if your doctor thinks you're a good candidate for it.