TRENDS IN THE USE OF ARTIFICIAL INTELLIGENCE IN ENDOSCOPY

This study provides a comprehensive overview of current trends in the application of artificial intelligence (AI) in endoscopic examinations of the gastrointestinal tract (GIT). AI significantly enhances medical diagnostic capabilities, particularly in detecting early signs of oncological and other diseases of the GIT. The analysis highlights the advantages and disadvantages of AI as an auxiliary tool for gastroenterologists, capable of improving examination 4 accuracy and reducing the risk of missing pathologies that are difficult to recognize during the procedure. Particular attention is given to the use of convolutional neural networks (CNN), which demonstrate significant results in image analysis and enhancement. The research discusses two primary approaches in endoscopy: classical and capsule endoscopy. Capsule endoscopy allows for thorough video analysis and is an effective tool for detecting lesions in hard-to-reach areas of the GIT, although its efficacy may depend on the quality of video, type of endoscope, camera, and algorithms used for data processing. Classical endoscopy, on the other hand, remains a widely used method that provides the opportunity for direct examination of the mucosal surface and allows specialists to utilize various image enhancement techniques to identify subtle changes. Studies indicate that the application of AI in classical endoscopy can achieve accuracy rates of up to 98% in identifying potentially hazardous areas, including the specific differentiation of lesion depths. Additionally, this work focuses on the challenges facing AI in endoscopy, particularly image artifacts that may negatively impact analysis results, as well as ethical considerations and data privacy compliance in medical data handling. Despite existing challenges, AI demonstrates significant potential as an assistant for both novice and experienced specialists, enabling the minimization of error risks and improving early disease detection. Overall, the findings indicate that artificial intelligence holds great promise in medicine and can significantly enhance the quality of endoscopic procedures, especially for early-stage diagnosis.