Artificial intelligence technologies in digestive system endoscopy: the state of the problem and prospects (literature review)

The review presents a comprehensive analysis of the current state and prospects for the use of artificial intelligence (AI) technologies in endoscopy of the digestive system. The research covers the main areas of AI implementation in endoscopic practice, including CADe and CADx computer vision systems, machine learning methods and deep learning algorithms. The paper examines the features of endoscopic procedures that affect the effectiveness of AI technologies: patient preparation, imaging quality depending on the skills of the endoscopist, and the multimodality of modern endoscopic methods. The results demonstrate the active development of AI technologies in endoscopy, especially in the field of detecting pathological changes in the gastrointestinal tract. Key applications of AI include cancer detection, diagnosis of Helicobacter pylori, assessment of inflammatory diseases, and quality control of research. The analysis shows that despite significant advances in the development of AI systems for endoscopy, their implementation is limited by a number of factors, including dependence on the operator and the complexity of standardization. In the near future, new approaches will be introduced to train AI models, including recurrent neural networks and multimodal AI systems that combine visual data with other patient information.