Position Paper: IOlBD Evidence-based Consensus on the Use of Artificial Intelligence for Assessment of Endoscopic and Histologic Endpoints in Clinical Trials of Inflammatory Bowel Disease - PubMed
4 days ago
- #Inflammatory Bowel Disease
- #Clinical Trials
- #Artificial Intelligence
- Central reading of endoscopy and histopathology is the current standard for assessing disease activity in IBD clinical trials but has limitations like variability, delays, and cost.
- AI and ML can improve accuracy, efficiency, and reproducibility in assessing endoscopic and histologic endpoints in IBD trials.
- The IOIBD conducted a review and formulated 36 survey statements, with consensus reached on 28 statements related to endoscopy, pathology, and trial design.
- Experts agreed AI-based central reading could enhance diagnostic accuracy, speed up processes, reduce costs, and improve reproducibility.
- Combining human and AI assessments was preferred over replacing human readers entirely.
- Key limitations include insufficient validation, generalizability concerns, and reliance on human-annotated training datasets.
- The consensus supports integrating AI/ML into IBD trial central reading while maintaining human oversight.
- Further research is needed on validation, regulatory frameworks, and multi-modal integration for broader adoption.