Clement David-Olawade Aanuoluwapo, Aderinto Nicholas, Egbon Eghosasere, Olatunji Gbolahan Deji, Kokori Emmanuel, Olawade David B
Endoscopy Unit, Glenfield Hospital, University Hospitals of Leicester, NHS Trust, Leicester, United Kingdom.
Department of Medicine and Surgery, Ladoke Akintola University of Technology, Ogbomoso, Nigeria.
J Gastrointest Surg. 2025 Oct;29(10):102195. doi: 10.1016/j.gassur.2025.102195. Epub 2025 Aug 19.
Endoscopy remains the gold standard for gastrointestinal (GI) diagnostics, enabling direct visualization and intervention within the GI tract. However, diagnostic accuracy and procedural outcomes vary significantly depending on the endoscopist's skill and experience, leading to potential missed lesions and inconsistent patient care. The integration of artificial intelligence (AI) into endoscopic practice offers a promising solution to address these limitations and enhance diagnostic precision. This review explores the current applications of AI in endoscopy, focusing on image analysis, lesion detection, classification, and workflow optimization, while evaluating the impact on clinical practice and identifying implementation challenges.
A literature search was conducted using PubMed, Google Scholar, and IEEE Xplore databases for studies published between January 2010 and December 2024. Keywords included "Artificial Intelligence," "Endoscopy," "Gastrointestinal Diseases," "Image Analysis," and "Lesion Detection." Studies were selected based on their focus on AI applications in endoscopy with quantitative or qualitative data on performance and clinical impact.
AI demonstrates exceptional capabilities in polyp detection, achieving detection rates that often surpass those of human practitioners, with systems such as GI Genius showing high sensitivity and specificity. Convolutional neural networks excel in real-time lesion identification and classification, differentiating between benign and malignant growths with remarkable precision. AI also optimizes endoscopic workflows through automated reporting and advanced training tools.
Although AI integration shows promise for enhancing endoscopic diagnostic accuracy and procedural efficiency, successful implementation requires careful consideration of current limitations, including reliance on industry-sponsored studies, and addressing challenges in data quality, clinical workflow integration, and regulatory considerations. Future developments in advanced algorithms, personalized medicine, and telemedicine may further advance endoscopic practice and improve patient outcomes.
