Figures and data
Human-based multi-scale modelling and simulation in acute and chronic myocardial infarction.
(A) Simulations using a population of ventricular models (n=17) to produce ECGs (light blue traces) and pressure-volume (PV) loops (grey traces) superimposed with the baseline ventricular model (ECG in blue and PV in black). Biomarkers are calculated from the baseline simulation of ECG and PV, as illustrated. (B) Ventricular electrophysiology is simulated using a fast endocardial activation layer to approximate Purkinje-myocardial junction, experimentally-informed transmural and apico-basal heterogeneities in action potential duration, and transmurally varying myocyte orientation. Mechanical pumping behaviour is modelled by coupling the intraventricular pressures with a two-element Windkessel model of arterial haemodynamics with a fixed basal plane. (C) An anterior 75% transmural infarction is modelled with acute and chronic ionic remodelling embedded in the border zone and remote zones. Standard 12-lead ECG was evaluated at standard body-surface locations. (D) Simulated action potentials using populations of human ventricular models in healthy (baseline) and acute and chronic post-MI conditions with different degrees of ionic remodelling. (E) Schematic representation of ionic fluxes, calcium dynamics and actin/myosin contraction mechanisms in the human ventricular electromechanically-coupled cellular model.
Agreement between simulated and clinical ECGs demonstrating variability in clinical phenotypes in acute and chronic post-myocardial infarction (post-MI).
(A) In acute MI, simulated ECGs show T-wave inversion (border zone model 1 (BZ1)), Brugada phenocopy (BZ2), and normal phenotypes (BZ3), in accordance with phenotypes found in clinical databases. (B) In chronic MI, simulated ECGs show prolonged QT and upright T-waves with a range of amplitude and duration (remote zone model 1 and 2 (RZ1, RZ2)) comparable to those observed in clinical databases. (C) ECG simulations of control, and acute and chronic post-MI considering ionic variability of the baseline ToR-ORd model. T wave morphologies for acute and chronic post-MI are mostly preserved across ionic variability.
Multiscale explanation of ST and T-wave phenotypes in acute MI.
(A) Activation time maps reveal conduction delay in acute border zone in T-wave inversion and normal ST-T phenotypes, and conduction block in Brugada phenocopy, as well as large repolarisation dispersion and altered transmural repolarisation gradient in T-wave inversion and Brugada phenocopy. Red in activation map show regions of no activation (NA), green in repolarisation map highlights regions of no repolarisation (NR). (B) Action potential duration (APD) prolongation is present in T-wave inversion and Brugada phenocopy cellular phenotypes, as well as slight shortening (normal ST-T) (red arrows), with decreased calcium amplitude in all phenotypes, and a corresponding decrease in active tension generation. (C) INa, ICaL, and IKr remodelling underpin reduced conduction, reduced calcium amplitude, and alterations in action potential duration, respectively, in all acute phenotypes.
Multiscale explanation of QT and T-wave phenotypes in chronic MI.
(A) Conduction delay in chronic border zone causes slight QT prolongation and large T-wave phenotypes, as well as large repolarisation dispersion in large T-wave. Red in activation map show regions of no activation (NA), green in repolarisation map show regions of no repolarisation (NR). (B) Varying degrees of action potential duration (ADP) prolongation in the remote zone (RZ) corresponding to extent of QT prolongation (blue arrows), with decreased calcium amplitude in remote and border zone of both phenotypes, and corresponding decrease in active tension generation. (C) As in acute MI, INa, ICaL, and IKr remodelling underpin reduced conduction, reduced calcium amplitude, and degree of prolongation in action potential duration, respectively, in both chronic phenotypes.
Reduced LVEF and heterogeneous systolic deformation caused by BZ ionic remodeling in both acute and chronic post-myocardial infarction.
Pressure-volume loops are shown in black (control) or red (post-MI) traces for the baseline model, and in gray (control) or pink (post-MI) traces for the population of models. (A) Reduced LVEF in all acute phenotypes due to reduced active tension amplitude in the border zone (BZ1∼3). Brugada phenocopy shows the lowest LVEF due to activation block and loss of contractile function in the border zone in addition to reduced active tension amplitude in the activated border zone due to ionic remodelling. Reduced contractile function in infarct and border zone results in infarct thinning and bulging in systole. Systolic cross section of control simulation shown in black (CTL) with post-MI cross-sections superimposed. (B) Reduced LVEF in both chronic phenotypes due to reduced active tension amplitude in remote zone (RZ) and border zones, independent of the extent of QT prolongation (RZ1, RZ2). Scar stiffening helped to reduce infarct bulging. Systolic cross-section of control simulation shown in black (CTL) with post-MI cross-sections superimposed.
T-wave alternans in simulations underpinned by APD and calcium alternans at fast pacing (120 bpm), albeit with preserved left ventricular ejection fraction (LVEF=49%) at rest (75 bpm) for the chronic MI phenotype
(A). (B) Simulated APD and calcium traces in midmyocardial population of models with remote zone 2 (RZ2) remodelling. (C) Large action potential and calcium transient alternans were caused by EADs in simulations at 120 bpm with midmyocardial cells affected by RZ2 ionic remodelling (green traces, representative example at 75 versus 120 bpm). A single cell model (in green) was selected from the population of models (in grey) for embedding into the remote region for ventricular simulations.
Prolonged QT and preserved LVEF at rest can manifest as severely abnormal ECG at fast heart rates in chronic MI with RZ2
(A), due to electrotonically-triggered EADs across the border zone (B). In (B), membrane potential changes for the first 1010 ms of fast pacing simulation, showing ectopic wave generation driven by electrotonic gradient at 710 ms (arrow from red dot to green dot in lateral view), and anticlockwise propagation of ectopic wave starting at 810 ms (anticlockwise arrow in basal view, and arrow from green dot to purple dot in lateral view). Ectopic wave propagates towards the right ventricle via the posterior side at 910 ms (arrow in basal view) and at 1010 ms (arrow in basal view). (C) Local action potential at anterior (red), lateral (green), posterior (purple), and right ventricular (yellow) sites. (D) A population of models demonstrating chronic remote zone 2 (RZ2) remodelling in promoting EADs. A representative example was selected from the population of models that showed EAD and was embedded in ventricular simulations.
