Harnessing the Power of AI to Enhance Endoscopy: Promises and Pitfalls

Slideshow below.

The availability of AI technologies to improve the overall quality of GI endoscopy has grown significantly over the last decade.¹ AI algorithms have shown promise in detecting malignant, infectious, and inflammatory diseases in both upper and lower GI endoscopy imaging.² Most AI research in endoscopy is currently focused on computer-aided detection (CADe) and diagnosis (CADx), but other computer vision tools are also in development, ranging from algorithms to measure IBD activity to dysplasia detection in Barrett's esophagus.^2,3

Improved lesion detection and classification with AI can support clinical decision-making and lead to better patient outcomes, cost savings, clinician time management, and other efficiencies within the health care system.^2-7 However, some substantial barriers must be overcome and current projections must be validated with clinical and real-world trials before we can fully rely on AI in these settings.^8-11

1