A high-fidelity geometric multiscale hemodynamic model for predicting myocardial ischemia

Jincheng Liu; Bao Li; Yanping Zhang; Liyuan Zhang; Suqin Huang; Hao Sun; Jian Liu; Xi Zhao; Mingzi Zhang; Wenxin Wang; Youjun Liu

doi:10.1016/j.cmpb.2023.107476

A high-fidelity geometric multiscale hemodynamic model for predicting myocardial ischemia

Comput Methods Programs Biomed. 2023 May:233:107476. doi: 10.1016/j.cmpb.2023.107476. Epub 2023 Mar 12.

Authors

Jincheng Liu¹, Bao Li¹, Yanping Zhang¹, Liyuan Zhang¹, Suqin Huang¹, Hao Sun¹, Jian Liu², Xi Zhao³, Mingzi Zhang⁴, Wenxin Wang⁵, Youjun Liu⁶

Affiliations

¹ Department of Biomedical Engineering, Faculty of Environment and Life, Beijing University of Technology, Beijing, China.
² Department of Cardiology, Peking University People's Hospital, Beijing, China.
³ Central Research Institute, United Imaging Healthcare, Shanghai, China.
⁴ Macquarie Medical School, Faculty of Medicine, Health, and Human Sciences, Macquarie University, Sydney, Australia.
⁵ Philips Healthcare China, Shanghai, China.
⁶ Department of Biomedical Engineering, Faculty of Environment and Life, Beijing University of Technology, Beijing, China. Electronic address: lyjlma@bjut.edu.cn.

PMID: 36933317
DOI: 10.1016/j.cmpb.2023.107476

Abstract

Background and objectives: Coronary computed tomography angiography (CCTA) derived fractional flow reserve (CT-FFR) requires a maximal hyperemic state to be modeled by assuming the total coronary resistance decreased to a constant 0.24 of that under the resting state. However, this assumption neglects the vasodilator capacity of individual patients. Herein, we proposed a high-fidelity geometric multiscale model (HFMM) to characterize coronary pressure and flow under the resting state, seeking to better predict myocardial ischemia by using CCTA-derived instantaneous wave-free ratio (CT-iFR).

Methods: Fifty-seven patients (62 lesions) who had undergone CCTA and were then referred to invasive FFR were prospectively enrolled. The coronary microcirculation resistance hemodynamic model (RHM) under the resting condition was established on a patient-specific basis. Coupled with a closed-loop geometric multiscale model (CGM) of their individual coronary circulations, the HFMM model was established to non-invasively derive the CT-iFR from CCTA images.

Results: With the invasive FFR being the reference standard, accuracy of the obtained CT-iFR in identifying myocardial ischemia was greater than those of the CCTA and non-invasively derived CT-FFR (90.32% vs. 79.03% vs. 84.3%). The overall computational time of CT-iFR was 61 ± 6 min, faster than that of the CT-FFR (8 h). The sensitivity, specificity, positive predictive value, and negative predictive value of the CT-iFR in discriminating an invasive FFR > 0.8 were 78% (95% CI: 40-97%), 92% (95% CI: 82-98%), 64% (95% CI: 39-83%), and 96% (95% CI:88-99%), respectively.

Conclusions: A high-fidelity geometric multiscale hemodynamic model was developed for rapid and accurate estimation of CT-iFR. Compared with CT-FFR, CT-iFR is of less computational cost and enables assessment of tandem lesions.

Keywords: Computational fluid dynamics (CFD); Coronary artery disease; Coronary pre-arterioles compensation; Instantaneous wave-free ratio (iFR).

MeSH terms

Coronary Angiography / methods
Coronary Artery Disease* / diagnostic imaging
Coronary Stenosis*
Fractional Flow Reserve, Myocardial*
Hemodynamics
Humans
Myocardial Ischemia* / diagnostic imaging
Predictive Value of Tests
Severity of Illness Index