Machine Learning Applications in Endoscopic Assessments

This document is a research report detailing the advancements in machine learning (ML) applications for gastrointestinal (GI) endoscopic assessments, specifically focusing on inflammatory bowel disease (IBD). It presents five abstracts from the ACG 2024 conference that highlight the potential of ML models to enhance clinical research and address unmet patient needs. The report outlines various methodologies, including systematic reviews and retrospective analyses, that evaluate the efficacy of ML in predicting endoscopic scores and improving treatment selection. Key findings indicate that ML models can accurately assess endoscopic responses and remission rates, with reported accuracies ranging from 56.8% to 96%. The document also discusses the collaboration between Iterative Health and academic institutions to refine endoscopic assessments, aiming to improve long-term patient outcomes. The overall goal is to leverage real-world evidence and advanced AI capabilities to transform traditional endpoints in GI clinical research.