Architecture of a software modeling complex for predicting key indicators of patient condition using medical digital twins

In clinical practice, the patient's medical digital counterparts are already demonstrating their potential in pilot projects and research. They have found their practical application in many branches of medicine, such as oncology, cardiology, neurology, etc. Continuing the series of works devoted to the development of medical digital patient doubles, we consider it important to recall that they are based on various algorithmically interpreted biomedical data on the human condition. Their integration is carried out using artificial intelligence, which makes it possible to develop a personalized prognosis and prepare recommendations for further action by doctors. Often, the quality and accuracy of the forecast, as well as recommendations for action, depend on the significance of a lot of data that will later be used by potential digital counterparts. However, processing and interpreting large amounts of information in a relatively short time requires the development of a special software modeling complex. The latter should be able to determine the values of the leading indicators of the condition at certain points in time, set by expert doctors, of some patients, by modeling them according to the field critical parameters of others. The current absence of such a complex makes the task of creating it urgent, which is what our efforts were aimed at.