Apparent growth tensor of left ventricular post myocardial infarction – In human first natural history study

• Published in: Computers in biology and medicine 2021-02, Vol.129, p.104168-104168, Article 104168
• Main Authors: Li, Wenguang, Gao, Hao, Mangion, Kenneth, Berry, Colin, Luo, Xiaoyu
• Format: Article
• Language: English
• Subjects: Biomarkers; Biomechanics; Cardiac magnetic resonance imaging; Cardiovascular diseases; Continuum mechanics; Coronary artery disease; Diastole; Gadolinium; Geometry; Growth and remodelling; Growth patterns; Growth tensor; Heart attacks; Heart diseases; Heart failure; Image enhancement; In vivo methods and tests; Left ventricle; Magnetic resonance imaging; Mathematical analysis; Medical imaging; Myocardial infarction; Patients; Regions; Remote regions; Reperfusion; Resonance; Statistical analysis; Tensors; Thoracic surgery; Ventricle
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2020.104168

An outstanding challenge in modelling biomechanics after myocardial infarction (MI) is to estimate the so-called growth tensor. Since it is impossible to track pure growth induced geometry change from in vivo magnetic resonance images alone, in this work, we propose a way of estimating a surrogate or apparent growth tensor of the human left ventricle using cine magnetic resonance (CMR) and late gadolinium enhanced (LGE) images of 16 patients following acute MI. The apparent growth tensor is evaluated at four time-points following myocardial reperfusion: 4–12 h (baseline), 3 days, 10 days and 7 months. We have identified three different growth patterns classified as the Dilation, No-Change and Shrinkage groups defined by the left ventricle end-diastole cavity volume change from baseline. We study the- trends in both the infarct and remote regions. Importantly, although the No-Change group has little change in the ventricular cavity volume, significant remodelling changes are seen within the myocardial wall, both in the infarct and remote regions. Through statistical analysis, we show that the growth tensor invariants can be used as effective biomarkers for adverse and favourable remodelling of the heart from 10 days onwards post-MI with statistically significant changes over time, in contrast to most of the routine clinical indices. We believe this is the first time that the apparent growth tensor has been estimated from in vivo CMR images post-MI. Our study not only provides much-needed information for understanding growth and remodelling in the human heart following acute MI, but also identifies novel biomarker for assessing heart disease progression. [Display omitted] •Growth tensors are estimated from left ventricle cine MRI and late gadolinium enhanced images post myocardial infarction.•Patients can be classified as Dilation, No-Change and Shrinkage groups; with wall remodelling in all groups.•The growth tensor invariants can be used as biomarkers for the growth and remodelling processes.