Artificial intelligence in gastrointestinal endoscopy: current evidence and future directions

Area of Science:

• Gastroenterology
• Medical Imaging
• Artificial Intelligence

Background:

• Gastrointestinal endoscopy is crucial for diagnosing GI disorders but faces limitations in human-dependent lesion detection.
• Artificial intelligence (AI) offers a promising technological solution to improve diagnostic precision and efficiency in endoscopic procedures.

Purpose of the Study:

• To review the impact of AI on gastrointestinal (GI) endoscopic procedures, focusing on lesion identification, classification, malignant polyp detection, and clinical interventions.
• To identify existing obstacles and future guidelines for AI implementation in clinical gastroenterology.

Main Methods:

• A comprehensive literature survey was performed using databases such as PubMed, Scopus, ScienceDirect, Elsevier, and Springer.
• Studies published after 2019 focusing on AI performance in GI endoscopic examinations were evaluated.

Main Results:

• AI integration in endoscopic visual examinations significantly improves diagnostic accuracy for esophagogastroduodenoscopy and colonoscopy lesions, surpassing traditional methods.
• AI enhances the prediction of malignant polyp status, aiding treatment decisions and reducing unnecessary biopsies.
• AI adoption boosts diagnostic outcomes, treatment efficiency, and clinical decision-making capabilities in GI endoscopy.

Conclusions:

• AI advancements in GI endoscopy improve diagnostic accuracy and treatment efficiency.
• Overcoming technical, data bias, regulatory, and clinical implementation challenges is crucial for AI's full integration into routine practice.
• Further research and technological improvements are needed to optimize AI's role in enhancing patient outcomes in GI endoscopy.