ARTIFICIAL INTELLIGENCE ON THE IDENTIFICATION OF DIABETES-RELATED OSTEOMETABOLIC DISORDERS

Abstract

Introduction: Complications of diabetes mellitus (DM) are of great medical and social importance, as they cause severe disability and premature death of patients with diabetes mellitus. Bone remodeling disorders occurring in diabetes increase the risk of fractures and move the problem of diabetic osteopathy beyond the narrow specialty, making it the subject of extensive scientific research [1-3]. However, osteopathy remains an underestimated complication and is not considered in most diabetes guidelines. The fact that diabetic osteopathy is often asymptomatic leads to the fact that diabetic patients turn their attention to this pathology late and turn to a specialist, as a rule, already having a high degree of progression of this complication. One of the important issues is the timely detection and prediction of bone changes in diabetes mellitus. The introduction of artificial intelligence technologies (AIT) into clinical practice is one of the main trends in world medicine [4]. AIT and Artificial Neural Networks (ANN) can fundamentally change the criteria for diagnosis and prognosis, which will contribute to the development of new therapeutic approaches, improve the efficiency of medical care and reduce costs [5]. The prospects for using ANN can potentially provide almost limitless technical possibilities. Considering the possibilities of using these technologies in clinical practice, we came to the conclusion that the development and implementation of forecasting systems based on the construction of a model of an intelligent decision support system based on the apparatus of artificial neural networks is able to analyze clinical and laboratory indicators of patients with diabetes mellitus (DM) in order to predict the values of qualitative and quantitative indicators assessing the state of bone tissue.