Artificial intelligence and neural networks in urology.

Abstract

Introduction. The review is devoted to the analysis of the world experience in the use of artificial intelligence in medicine and urology. Materials and methods. The review was conducted on the basis of data published in the PubMed, Scientific Electronic Library eLibrary.ru databases and on the professional medical associations websites. The article contains 43 scientific publications. Results. In this literature review, the types of neural networks and the features of their application are considered in detail. Special attention is paid to the use of artificial intelligence in medical imaging, particularly in urology. Effectiveness and accuracy of developing the connectionism method application in artificial neural network (ANN) in recent publications are presented. Advantages and disadvantages of artificial intelligence for a predictive model are outlined. We also considered the main mistakes displayed in publications regarding unreliable or poorly created neural networks and studies in the use of ANNs for clinical practice. Conclusion. The use of neural networks as a tool in the daily clinical practice of a specialist is still limited. The main problems are still the unreliability of the created models and lack of easy-to-use algorithms.