7th Global Gastroenterology and Artificial Intelligence Summit (Live)

12 - Akshintala

Pdf Summary

This presentation outlines recent advances and future directions of Artificial Intelligence (AI) in Gastroenterology (GI), emphasizing its impact on endoscopy, risk prediction, workflow automation, and decision support. Over the past decade, AI-related GI publications have risen exponentially, notably in colonoscopy for polyp detection and classification, achieving accuracies up to 96.4%. Applications include Barrett’s esophagus diagnosis, biliary stricture identification, and capsule endoscopy that significantly reduces reading time.<br /><br />However, real-world evaluations reveal challenges, such as limited sensitivity and specificity improvements in computer-aided diagnostics (CADx) for small adenomas. AI also extends to Endoscopic Ultrasound (EUS) for structure identification, mass detection, classification, and pathology correlation through texture analysis, aiding pancreatic cancer prediction with models demonstrating up to 90% sensitivity and specificity in early detection.<br /><br />Emerging technologies combine AI with next-generation sequencing and biomarkers for comprehensive disease insights. Federated learning enables multi-institutional model development while preserving patient data privacy. Future prospects include robotic integration for semi-autonomous endoscopy and surgery, aiming for ‘self-driving’ endoscopes.<br /><br />Clinical decision-making benefits from AI-powered risk models in conditions like choledocholithiasis and post-ERCP pancreatitis, improving patient stratification and management. Workflow enhancements involve ambient AI for documentation, reducing non-endoscopy tasks by over 67%, leveraging speech-based endoscopy note systems (EndoScribe) that generate real-time, high-quality reports to improve efficiency.<br /><br />The American Society for Gastrointestinal Endoscopy (ASGE) supports innovation through commercialization guidance, workshops, and training to transition AI solutions from research to scalable clinical tools. The takeaway highlights AI’s transformative potential to personalize diagnostics and treatment, reduce administrative burden, and address unmet clinical needs—advocating for continued innovation and adoption within GI practice.

Keywords

Artificial Intelligence

Gastroenterology

Endoscopy

Colonoscopy

Polyp Detection

Computer-Aided Diagnostics

Endoscopic Ultrasound

Pancreatic Cancer Prediction

Federated Learning

Workflow Automation