Abstract
In contemporary medicine, the increasing proportion of emergency surgical interventions and elective operative procedures is directly proportional to the rising incidence of postoperative complications and mortality rates. Within the algorithm of surgical care delivery, the early identification of preoperative risk factors and potential complications, along with the appropriate planning of the intraoperative period and postoperative monitoring and care, holds substantial importance. Currently used methods for perioperative risk assessment in clinical practice, such as surgical scoring systems, the ASA classification, and clinical hypotheses, are often limited by their subjective nature and are insufficient for a comprehensive, multifactorial, and multi-spectral evaluation. Nevertheless, preoperative laboratory data, including routine hematological, biochemical, and immunological inflammatory markers, may be regarded as objectively significant indicators for perioperative patient assessment. In the current medical revolution, artificial intelligence models and machine learning methods are widely employed for sorting, analyzing, and processing data within databases. However, the structured investigation and practical validation of specifically trained AI models using machine learning for perioperative risk factor evaluation remains insufficiently explored. Accordingly, the development of AI models through specialized automated machine learning techniques, and the standardization of their performance for the accurate determination of risk types during the perioperative period, are currently regarded as relevant and promising strategies.