Particle Filter Based Framework for the Prognosis of Atherosclerosis via Lumped Cardiovascular Modeling

Abstract

Atherosclerosis refers to the plaque deposition in the arteries that can eventually lead to any of the three cardiovascular diseases, namely, heart attack, stroke, or peripheral vascular disease, depending upon the site of the blockage in the human arterial network. This work attempts to prognose this pathological condition via lumped cardiovascular modeling while utilizing the radial artery blood pressure measurements. To achieve this, the cardiovascular system has been modeled as a third order non-linear system with explicit emphasis on the systemic circulation. The parameters of the model are estimated using non-linear least squares estimation technique by minimizing the error between the measured and the estimated arterial pressure waveforms. The arterial pressure is found to be sensitive to three of the model parameters, namely, arterial compliance, systemic vascular resistance, and the peak cardiac muscle elastance. Based on the analysis, a growth model of systolic blood pressure is developed as a function of the arterial blockage. A particle filter based mathematical framework is then utilized to predict the time it would take to reach the stage of critical arterial blockage that may cause heart attacks.