Artificial Intelligence in Gastrointestinal Endoscopy: Current Advances, Clinical Integration, and Future Directions:...

Area of Science:

• Gastroenterology and Endoscopy
• Medical Artificial Intelligence
• Diagnostic Technology

Background:

• Artificial intelligence (AI) is revolutionizing gastrointestinal (GI) endoscopy by enhancing diagnostic accuracy and workflow.
• Multiple AI-assisted systems are entering clinical practice, necessitating a review of their performance and limitations.

Purpose of the Study:

• To synthesize evidence on the effectiveness, limitations, and implementation of AI in GI endoscopy.
• To identify translational gaps and future research priorities for AI in this field.

Main Methods:

• A structured narrative synthesis of literature from major databases (PubMed, Scopus, Web of Science) from 2005-2025.
• Thematic analysis of AI applications including CADe, CADx, upper GI neoplasia, and capsule endoscopy.
• Extraction of key lessons, barriers, and research priorities, with performance estimates from select RCTs and meta-analyses.

Main Results:

• Computer-aided detection (CADe) systems show a 15-20% relative increase in adenoma detection rate (ADR).
• Computer-aided diagnosis (CADx) algorithms exceed 90% accuracy in polyp differentiation.
• AI tools achieve 88-96% accuracy for early esophageal and gastric neoplasia detection, surpassing non-expert endoscopists.

Conclusions:

• AI demonstrates proven efficacy in improving detection and diagnostic precision in GI endoscopy.
• Overcoming barriers like dataset bias, generalizability, and regulatory issues is crucial.
• Multicenter validation, standardized datasets, ethical guidelines, and clinician training are essential for AI integration.