Statistical analysis of coronary blood flow monitoring data for hemodynamic assessment of the degree of coronary artery stenosis

Abstract

Statistical relationships between the pressure curves Pa(t), Pd(t) and blood flow velocity Va(t), recorded in vivo in the coronary arteries of patients before and after stenosis, as part of the standard clinical procedure for calculating dynamic indices FFR, HSR, CFR, and a number of other ones generally accepted in surgical practice are studied. It is shown that in the case of insignificant stenosis that does not require surgical intervention, there is a correlation between the curves, and their spectrum is represented by three main harmonics. In the case of significant stenosis requiring immediate stenting, the positive correlation between Pa(t) and Pd(t) is less pronounced, and there is a negative correlation with the Va(t) curve. The spectrum of the curves is much more complex and contains high-frequency harmonics. For patients from the so-called “gray zone”, an expert decision on the need for stenting can be made based on the appearance of additional harmonics in the spectrum and a negative correlation between the Pa(t), Pd(t) and Va(t) curves. The proposed approach can be used for automatic decision-making based on machine learning and the development of appropriate mathematical models.