Introduction. In recent years, artificial intelligence (AI) has become an important tool in clinical medicine, including in urology. The main areas of AI use in medicine are diagnostics, forecasting, and natural language processing etс. The integration of AI into everyday clinical practice is hindered by ethical and technical constraints.
Objective. This review aims to systematize and critically analyze data on the application of AI technologies in urology, including visual diagnostics, predictive modeling, automation of medical documentation, as well as discussing the ethical and legal aspects of implementation.
Materials and methods. A literature search was conducted in the PubMed, Scopus, and eLibrary databases. The search depth was 5 years (2019–2024) to ensure data relevance in this rapidly evolving field. The following keywords were used: «artificial intelligence», «urology», «machine learning», «deep learning», «neural networks», «diagnostics», «prognosis», «ethical aspects», «legal regulation». The review included original studies, systematic reviews, and meta-analyses on the use of AI in urology. Exclusion criteria were conference abstracts, news articles, and studies without clinical validation of results.
Results. The summarized data demonstrate high diagnostic accuracy of AI algorithms in detecting urinary stones (over 90%), differential diagnosis of kidney tumors (93.9%), and detection of bladder neoplasms during cystoscopy (sensitivity 90.9%, specificity 98.6%). Machine learning-based prognostic models achieved 94.8% accuracy for predicting postoperative complications after percutaneous nephrolithotomy and 98% – for ten-year cancer-specific survival in prostate cancer. The implementation of AI scribes can save physicians up to 1 hour per day when completing medical documentation.
Conclusion. AI does not replace the urologist but can be a powerful tool for improving diagnostic accuracy, optimizing treatment, and automating routine processes. The implementation of these technologies requires the development of industry-specific validation standards, ethical protocols, and a regulatory framework. Based on the analysis, practical recommendations for integrating AI into clinical practice are formulated, and unresolved scientific problems defining future research directions are identified.
