Track 09 of 9
AI in GI
Polyp detection, histology prediction and decision support.
AI is now embedded in routine endoscopy, with CADe systems (GI Genius, ENDO-AID, CAD EYE) demonstrating consistent adenoma detection rate (ADR) gains in randomized trials and real-world registries. CADx optical biopsy is approaching a credible diagnose-and-leave or resect-and-discard strategy for diminutive polyps. The track covers AI for Barrett’s neoplasia, capsule endoscopy reading, ulcerative colitis endoscopic Mayo scoring, and histopathology deep-learning models for dysplasia and MSI prediction. Implementation sessions address regulatory pathways (FDA SaMD, EU MDR), bias and generalizability, prospective ADR drift studies, and integration with electronic health records.
Focus areas
- CADe systems: GI Genius, ENDO-AID, CAD EYE ADR outcomes
- CADx and resect-and-discard for diminutive polyps
- AI for Barrett’s dysplasia and early esophageal neoplasia
- Deep learning in capsule endoscopy and small-bowel imaging
- Automated UCEIS and Mayo endoscopic scoring
- Histopathology AI: dysplasia, MSI and HER2 prediction
- Regulatory pathways (FDA SaMD, EU MDR) and post-market surveillance