Key Points:
- An AI-assisted image-enhanced endoscopy system accurately assessed vascular healing and predicted long-term clinical relapse in patients with ulcerative colitis (UC), based on data from a study of a novel investigational tool.
- Patients with vascular healing in all segments had a lower clinical relapse rate (3%) compared to those with vascular activity (23.9%).
- The AI system, called EndoBRAIN-UC, was trained on images from five different scopes and developed to differentiate between vascular activity and vascular healing, allowing for the prediction of relapse.
- In a prospective cohort study of 100 patients with UC in clinical remission, the clinical relapse rate was higher in the vascular active group identified by AI compared to the vascular healing group.
- The diagnostic output of the AI was evaluated independently of the endoscopist’s experience, showing consistent results.
- AI-assisted colonoscopy is still in early stages, and while initial results are promising, widespread implementation into routine clinical practice may take 10-20 years, according to experts.
Implications:
- AI-assisted image-enhanced endoscopy systems like EndoBRAIN-UC hold potential for improving the assessment of UC patients in clinical remission, aiding in the prediction of long-term clinical outcomes.
- The ability to differentiate between vascular activity and healing using AI may allow for more personalized treatment strategies and better management of UC patients.
- Continued research and clinical data are necessary to validate the effectiveness and reliability of AI-assisted colonoscopy systems before widespread adoption in routine clinical practice.