Open Access

ARTICLE

Using 2D In Vivo IVUS-Based Models for Human Coronary Plaque Progression Analysis and Comparison with 3D Fluid-Structure Interaction Models: A Multi-Patient Study

by

* School of Biological Science and Medical Engineering, Southeast University, Nanjing, 210096, China.
† Mallinckrodt Institute of Radiology, Washington University, St. Louis, MO, USA.
‡ Mathematical Sciences Department, Worcester Polytechnic Institute, MA, USA.
§ The Cardiovascular Research Foundation, NY, NY, USA.
¶ Network Technology Research Institute, China United Network Communications Co., Ltd., Beijing, China.
II Cardiovascular Division, Washington University School of Medicine, Saint Louis, MO, USA.
** Department of Cardiology, Zhongda Hospital, Southeast University, Nanjing, China.
†† Corresponding author. Tel: 508-845-1575; E-mail: dtang@wpi.edu

Molecular & Cellular Biomechanics 2015, 12(2), 107-122. https://doi.org/10.3970/mcb.2015.012.107

Abstract

Computational modeling has been used extensively in cardiovascular and biological research, providing valuable information. However, 3D vulnerable plaque model construction with complex geometrical features and multicomponents is often very time consuming and not practical for clinical implementation. This paper investigated if 2D atherosclerotic plaque models could be used to replace 3D models to perform correlation analysis and achieve similar results. In vivo intravascular ultrasound (IVUS) coronary plaque data were acquired from a patient follow-up study to construct 2D structure-only and 3D FSI models to obtain plaque wall stress (PWS) and strain (PWSn) data. One hundred and twenty-seven (127) matched IVUS slices at baseline and follow up were obtained from 3 patients. Our results showed that 2D models overestimated stress and strain by 30% and 33%, respectively, compared to results from 3D FSI models. 2D/3D correlation comparison indicated that 116 out of 127 slices had a consistent correlation between plaque progression (WTI) and wall thickness; 103 out of 127 slices had a consistent correlation between WTI and PWS; and 99 out of 127 slices had a consistent correlation between WTI and PWSn. This leads to the potential that 2D models could be used in actual clinical implementation where quick analysis delivery time is essential.

---

Keywords

---