Longitudinal In-silico Biomechanical Study of Patient-Specific Atherosclerotic Arteries Reveals Key Mechanically Mediated Plaque Remodeling Features

Introduction:: In 2020 the CDC reported that approximately one in five deaths were attributed to cardiovascular disease (CVD) from 1999-2020 [1], [2]. In particular, coronary artery disease (CAD), also called atherosclerosis, accounted for roughly 55% of CVD-attributed deaths, making it the leading cause of death across all CVD [2]. CAD is an inherently complex disease wherein a fatty plaque forms and accumulates within the artery wall. It is known that the growth and progression of plaque in coronary arteries is largely driven by inflammation caused by low or oscillatory wall shear stress on the lumen [3], increased leukocyte transendothelial migration [4], and low-density lipoprotein (LDL) infiltration [5]. What is less known is how solid mechanics within the lesion contribute to CAD. The work presented herein aimed to quantify this contribution of inter-plaque mechanics to overall plaque growth. To do so, we used patient-specific data and in-silico biomechanical modeling to correlate the interwall biomechanics of real-patient arteries with changes in plaque morphology over 6 months. The outcome of this work will further the current understanding of mechanically mediated plaque growth with implications in personalized treatment to minimize disease progression.


Materials and Methods:: Patients often undergo coronary angiography and virtual histology intravascular ultrasound (VH-IVUS) to diagnose plaque severity and uncover locational and compositional features. Recently, we developed an automated algorithm capable of using patient-specific VH-IVUS images and coronary angiography data to generate volumetric 3D meshes of coronary arteries for in-silico biomechanical modeling [6]. In collaboration with a medical center, Emory University, we have access to a 16-patient dataset of paired VH-IVUS and angiography images at baseline (BL) and 6-month follow-up (FU) which were used for this study.



Meshing and FEA. Volumetric meshes of BL arteries for 16 patients were generated using our automated algorithm. Each mesh comprised quadratic 10-node tetrahedral element shapes with edge lengths of approximately 0.15mm which was suitable to accurately represent semi-fine details present in the VH-IVUS images. Patient-specific blood pressure was applied to the luminal surface of each mesh. A nonlinear Neo-Hookean material model with literature-derived material properties for each plaque constituent was applied. Finite element analysis (FEA) simulations were performed using the open-source FEBio software. Mean elemental effective stresses and strains were calculated from the FEA results and co-registered to each image.



Image-derived Features. A separate automated script analyzed each pair of BL and FU VH-IVUS images and extracted image-derivable features. These features included surface area, radial coverage, radial depth, and radial thickness of each plaque constituent [Figure 1]. Differences between BL and FU values were then calculated for use in the following analysis.


Results, Conclusions, and Discussions:: RESULTS

Preliminary results using 13 of the 16 patient datasets (n = 842 images) suggest larger changes in plaque area between BL and FU occur at low strains whereas smaller changes in surface area occur at mid-high strains [Figure 2]. Similar patterns were seen when comparing radial coverage and radial thickness of the plaque as well. A noteworthy observation is how fibrotic and fibrofatty surface area decrease at smaller strains, whereas calcium and necrotic surface area increase at lower strains. This key finding suggests that low strain potentially causes fibrotic and fibrofatty to present a more necrotic or calcified morphology over time.



DISCUSSION

Our preliminary results corroborate the work by Samady et al. which found that low hemodynamical WSS correlated with increased plaque volume over a 6-month period [7]. Here, it is thought that endothelial cells lining the lumen must be in a constant state of motion to maintain healthy cell-cell connections between them and minimize transport of leukocytes or LDL into the wall. Similarly, we hypothesize that normal function of smooth muscle cells (SMCs) inside the artery wall, which are responsible for remodeling the extracellular matrix, can be impaired by lower strains. This pathological remodeling via SMCs has been replicated with in-vitro experiments. SMCs in low stretch conditions (0-10%) were found to proliferate less than those under large stretch conditions [8].



CONCLUSION

In conclusion, our preliminary results paired with the reported phenomenological observations suggest a mechanobiological remodeling paradigm largely influenced by the mechanics felt at the cellular level. Yet, we do see some instances where it is the stress distribution rather than the magnitude that drives volume changes. Further investigation is needed to quantify these relationships, but initial results suggest that plaque remodeling is a mechanically mediated process.


Acknowledgements (Optional): : Authors gratefully acknowledge funding for this work provided by the American Heart Association (17SDG33400239), the National Heart, Lung, And Blood Institute of the National Institutes of Health (NIH) under Award Number R01HL136776, and the Louis Stokes Alliances for Minority Participation program. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. Emory University (ClinicalTrials.gov; Identifier: NCT00576576) for the data set used in this study. Vascular Mechanobiology Lab at UTD for providing opportunity to work on this research through the guidance of Dr. Hayenga, the lab director.


References (Optional): : [1] CDC WONDER, Online Database, 2022.
[2] Tsao C. W. et al., Circulation, 145(8), 2022
[3] Sheikh S. et al., Blood, 102(8):2828–2834, 2003
[4] Bhui R. et al., PLoS Comput Biol, 13(5), 2017
[5] Dabagh M. et al., AJPHCP, 297(3):983–996, 2009
[6] Warren J. L. et al., Front Med Tech., 4(1), 2022
[7] Samady H. et al., Circulation, 124(7):779–788, 2011
[8] Mantella L.-E. et al., Vasc Cell, 7(1):7, 2015