IVUS-based Fluid-structure Interaction Models for Novel Plaque Vulnerability Indices: A Study in Patients with Coronary Artery Disease

It is believed that mechanical stresses play an important role in atherosclerotic plaque rupture process and may be used for better plaque vulnerability assessment and rupture risk predictions. IVUS data were acquired from 14 patients (11M, 3F, Mean age: 59,) for constructing 3D computational models combining fluid-structure interaction (FSI), cyclic bending due to cardiac contraction and patient-specific pressure loading to quantify mechanical conditions in the human coronary. The computational models were solved by a finite element package ADINA to obtain plaque wall stress (PWS), strain (PWSn) and flow shear stress (FSS) and investigate correlation between the mechanical conditions and morphological characteristics. For all 617 IVUS slices yielded from the 14 patients, plaque morphological features lipid percentage and min cap thickness were calculated for each slice, and three types of plaque morphology related indices: lipid index, cap index and morphological index (MPVI) were introduced as quantitative measures of plaque vulnerability. PWS, PWSn and FSS values at critical sites were denoted as critical plaque wall stress (CPWS), critical plaque wall strain (CPWSn) and critical flow shear stress (CFSS) for each slice, and a stress index was proposed based on the value of the CPWS. The conventional Pearson's correlation is used to analyze the correlation between each of the mechanical conditions and each plaque morphological feature indices. Our results suggest there is significant correlation between the CPWS and min cap thickness, cap index with the correlation coefficient r=-0.6570, r=0.8016 respectively, while the correlation between CPWS and lipid percentage and the lipid index are weaker (r=0.2209, r=0.2304) even though they are significantly correlated. The correlation results between CPWS and morphological index (r=0.7725, p-value<0.0001) showed there is a strong positive relationship between the mechanical stress and morphological features. For all 617 slices, the stress index has a 66.77% agreement with morphological index. More patient follow-up data and large-scale studies are needed to continue our investigations.