Human biventricular electromechanical simulations on the progression of electrocardiographic and mechanical abnormalities in post-myocardial infarction

Abstract

Aims: Develop, calibrate and evaluate with clinical data a human electromechanical modelling and simulation framework for multiscale, mechanistic investigations in healthy and post-myocardial infarction (MI) conditions, from ionic to clinical biomarkers.

Methods and results: Human healthy and post-MI electromechanical simulations were conducted with a novel biventricular model, calibrated and evaluated with experimental and clinical data, including torso/biventricular anatomy from clinical magnetic resonance, state-of-the-art human-based membrane kinetics, excitation-contraction and active tension models, and orthotropic electromechanical coupling. Electromechanical remodelling of the infarct/ischaemic region and the border zone were simulated for ischaemic, acute, and chronic states in a fully transmural anterior infarct and a subendocardial anterior infarct. The results were compared with clinical electrocardiogram and left ventricular ejection fraction (LVEF) data at similar states. Healthy model simulations show LVEF 63%, with 11% peak systolic wall thickening, QRS duration and QT interval of 100 ms and 330 ms. LVEF in ischaemic, acute, and chronic post-MI states were 56%, 51%, and 52%, respectively. In linking the three post-MI simulations, it was apparent that elevated resting potential due to hyperkalaemia in the infarcted region led to ST-segment elevation, while a large repolarization gradient corresponded to T-wave inversion. Mechanically, the chronic stiffening of the infarct region had the benefit of improving systolic function by reducing infarct bulging at the expense of reducing diastolic function by inhibiting inflation.

Conclusion: Our human-based multiscale modelling and simulation framework enables mechanistic investigations into patho-physiological electrophysiological and mechanical behaviour and can serve as testbed to guide the optimization of pharmacological and electrical therapies.

Keywords: Computer modelling; Ejection fraction; Electrocardiogram; Electromechanical simulations; Myocardial infarction.

Figure 1

Human biventricular electromechanical model properties in healthy and post-MI conditions. (A) Fibre and sheet-normal field. (B) Colour map for local times of activation in the endocardial fast activation layer in milliseconds. (C) ECG electrode placement in 3D space around the biventricular mesh. (D) Electrophysiological transmural heterogeneity, with blue—endocardial cell type, white—mid-myocardial cell type, and red—epicardial cell type. (E) Electrophysiological apex-to-base heterogeneity of the slowly activating delayed rectifier potassium channel current (IKs), with colour map showing conductance for the slow rectifier potassium current. (F) Infarct (red), border zone (white), and remote zone (transparent blue) for an anterior endocardial infarction. (G) Infarct (red), border zone (white), and remote zone (transparent blue) for a transmural left anterior descending artery infarction. (H) Epicardial action potentials and calcium transients for the baseline, ischaemic, acute, and chronic infarct regions. (I) Epicardial action potentials and calcium transient for the baseline, acute, and chronic borderzone regions.

Figure 2

Healthy electromechanical simulations. (A) Pressure–volume loop for both left ventricle (blue) and right ventricle (green) (left), pressure transient (right top), volume transient (right middle), and cardiac phase change (right bottom). (B) Simulated ECG in pre-cordial leads. (C) Mechanical deformation at four time points in the cardiac cycle 0 s (initial), 0.1 s (end of diastolic filling), 0.3 s (peak pressure), 0.4 s (end of systolic ejection). (D) Mid-ventricular slice showing radial strain at diastasis (DS), end-diastole (ED), and end-systole (ES). (E) Activation and repolarisation time maps with pre-cordial ECG locations shown as green spheres.

Figure 3

Effect of fully transmural myocardial infarction (MI) (A) on electromechanical function for three chronologically-ordered stages: (B) ischaemic, (C) acute post-MI, and (D) chronic post-MI. For each stage is shown: biventricular pressure–volume (PV) loop and pre-cordial ECG characteristics. Membrane potential plot at three consecutive time points in the cardiac cycle as labelled. Mid-ventricular short-axis slice (see A for slice position) showing deformation and radial strain at end-diastole (ED) and end-systole (ES). Negative strain (extension) in blue and positive strain (contraction) in red.

Figure 4

Clinical post-MI ECGs from the PTB Diagnostic ECG database (see reference in text) showing similar ST-segment and T-wave abnormalities at three chronological states post-MI as simulation. The patient had acute anterior MI on 18 October 1990 and was admitted in hospital on 19 October 1990. The catheterization date was 26 October 1990, and the three ECGs were obtained on 24 October 1990 (A), 29 October 1990 (B), and 03 December 1990 (C).