Validation of Artificial Intelligence for Detecting Early Gastrointestinal Neoplasia Using Endoscopic Images and Videos

What this Trial Is About

Researchers are evaluating artificial intelligence AI algorithms to detect and characterize early gastrointestinal GI neoplasia. This observational study is investigator-initiated and conducted at multiple centers. It aims to validate and improve AI tools by comparing their performance against expert assessments and histological diagnoses, using endoscopic images and videos collected during routine clinical procedures. The study has two phases a Validation & Optimization phase using anonymized images and videos from a training library to refine the AI algorithm, followed by a Testing phase where AI performance is assessed prospectively during standard endoscopy procedures. No additional procedures, biopsies, or interventions are performed specifically for the study. Participants undergo standard gastroscopy andor colonoscopy for clinical reasons, during which endoscopic images and videos are collected. Researchers will measure outcomes such as the rate of neoplasia detection by AI and endoscopists at various time points up to one year and study completion. The study does not affect patient care and involves no extra procedures beyond routine clinical practice.

Evaluation of Neoplasia With Artificial Intelligence in Gastrointestinal Endoscopy