Maslii B. Local anatomical and systemic determinants of predicting the severity of the course and choosing the volume of invasive intervention in patients with acute coronary syndrome

The dissertation is devoted to increasing the effectiveness of treatment of patients with acute coronary syndrome (ACS). The leading anatomical and functional determinants of the successful implementation of IRS in ACS were determined and statistically substantiated, in particular, the location of the coronary lesion, the number of affected segments, the degree and extent of stenosis/occlusion, the severity of the lesion according to the Gensini scale, and the development of CBF according to Rentrop. Local anatomical and systemic impact determinants on the severity of the course, immediate prognosis (hospital mortality), 6-month mortality and 5-year survival of patients after MI were established. In accordance with anatomical and functional changes in CA, the tactics of invasive intervention in patients with ACS have been optimized. Scientific data on the causes of pre-hospital and in-hospital delays in invasive management strategies for patients with acute coronary syndrome have been supplemented. A significant direct correlation with the severity of coronary artery disease according to the Gensini score and the degree of collateral development according to Rentrop (r = 0.555, p <0.001) has been established. Scientific data on the direct dependence between the degree of CA obstruction and the presence of well-developed collaterals (r=0.316, p<0.001) has been added. Direct correlations were found between the development of collaterals and damage to segments 4, 9, 11 and 13 (r>0.300, p<0.020). It was established that in a relatively short period of time, the presence of a developed CBF (Rentrop 2-3) has a positive effect on the 6-month survival with IRS and does not affect the 5-year survival of patients, which is explained by the older age of these patients and significantly worse initial clinical and functional characteristics, which in general affected the endpoints. The insufficient informativeness of the GRACE scale, recommended by ESC and widely used in clinical practice, with optimal IRS in multivessel lesions was established, because the actual mortality in these patients was 2.48 times lower than predicted by GRACE within 30 days, and 1.4 times lower within 6 months. To optimize the prediction of PCI outcomes in ACS, in addition to the recommended GRACE 2.0 scale, a machine learning method was applied, which was used to build a decision tree by classification and regression using Rentrop's assessment of collateral development, Gensini's degree of severity, and 5 risk groups were identified.