Optimization of prediction of morphological disorders of skeletal muscles in experimental acute ischemia-reperfusion on the basis of combined changes in lipid peroxidation and antioxidant protection by neural network clustering

Abstract

The paper proposes a method for optimizing the prediction of the development of morphological disorders of skeletal muscle in experimental acute ischemia-reperfusion based on combined changes in lipid peroxidation and antioxidant protection. The approach is based on the use of neural network clustering. The purpose of the study – to propose a method for optimizing the prediction of the severity of morphological disorders in experimental acute ischemia-reperfusion on the basis of combined changes in lipid peroxidation and antioxidant protection by neural network clustering. The experimental model of ischemic-reperfusion lesion was represented by five groups of rats with reperfusion terms of 1 and 2 hours, 1 day, 7 and 14 days (18 animals each). Acute limb ischemia-reperfusion was simulated by applying SWAT rubber tourniquets on the hind right limb of animals for two hours under thiopental-sodium anesthesia. Histological examination was performed at the Department of Pathological Anatomy with a sectional course and forensic medicine of I. Horbachevsky Ternopil National Medical University of Ministry of Health of Ukraine according to generally accepted methods. Determination of lipid peroxidation and antioxidant protection in the blood serum of each experimental group of rats was performed by the Central Research Laboratory of I. Horbachevsky Ternopil National Medical University of Ministry of Health of Ukraine. For in-depth analysis and clustering of the indicators of the studied groups in order to optimize the prediction of the course of ischemic-reperfusion lesion, a neural network approach was used using the NeuroXL Classifier add-in for Microsoft Excel. The greatest prognostic value in relation to the severity of morphological disorders in the early reperfusion period according to the data of neural network clustering have combined changes in the level of TBA-active products and catalase.