ECG analysis of ventricular fibrillation dynamics reflects ischaemic progression subject to variability in patient anatomy and electrode location

Figures

• 

  Figure 1. Human-based modelling and simulation framework for ischaemia-induced ventricular fibrillation. (A) Human biventricular model obtained from clinical magnetic resonance images, including rule-based fibre orientation (left) and ischaemia-induced electrophysiological alterations (right). (B) 21 scenarios were considered including control versus ischaemic regions in three locations (rows), 2 degrees of transmurality (columns), and 3 degrees of ischaemic severity in remote myocardium (colour-coded). (C) Degrees of ischaemic severity considered in the simulations; the top row represents healthy values for reference. Regional ischaemia is considered “moderate”. (D) Simulation protocols: (I) healthy remote myocardium (0–12 s); (II) healthy remote myocardium (0–6 s) becoming mildly ischaemic (6–12 s); (III) mildly ischaemic remote myocardium (0–12 s) becoming moderately ischaemic (12–18 s) leading to global ischaemia. (E) Electrode locations used for ECG simulation.

• 

  Figure 2. Model validation based on ECG and action potential (AP). (A) Comparison of simulated ECGs under healthy conditions with clinical ECGs from the healthy subjects. Left: Healthy simulation using the ToR-ORd model (solid line) vs. ECG record from the patient whose cardiac magnetic resonance images were used to construct the anatomical model (dashed line). Right: healthy simulation using the ToR-ORdVF model (solid line) vs. a clinical ECG from a healthy subject presenting similar T wave morphology (dashed line, patient from the UK Biobank). (B) Comparison of the electrocardiographic alterations induced by acute ischaemia in simulated ECGs and representative clinical ECGs from the STAFF III database. In our simulations, ST elevation is obtained using the baseline ToR-ORd model (left panel), whereas the tombstone pattern was reproduced using the ToR-ORdVF model (right panel). (C) Simulated AP with the ToR-ORdVF (solid lines) and ToR-ORd baseline (dashed lines) models in healthy (black), mild ischaemia (orange), and moderate ischaemia (red) conditions. AP durations (APD90) are in agreement with the ranges of experimental data in human ventricular cells reported in Sutton et al. (26) in…

• 

  Figure 3. Comparison of clinical and simulated ECGs in ventricular fibrillation (VF). (A) Clinical ECG recording during VF from PhysioNet database. Dominant frequencies (DF) were annotated 3 and 28 seconds post-VF onset. (B) Simulated ECGs during VF in acute regional ischaemia following LAD coronary artery occlusion (right panel, darker region depicts the ischaemic core zone) during simulation protocols (i), (ii), and (iii) (top, middle, and bottom panels, respectively). ECG markers and wavefronts were quantified in windows of 2,045 ms. (C) Typical ECG signature, wavefront analysis, and phase singularities in the different VF stages in the simulations: panel 0 represents sinus rhythm, whereas panels 1, 2, and 3 correlate to Wiggers' VF stages I, II, and III, respectively, as reported in their seminal work (48). For each phase, the range of typical DF values is provided.

• 

  Figure 4. Ischaemic remote myocardium promoted ventricular tachycardia/fibrillation (VT/VF) inducibility and persistence in regional ischaemia simulations. For each regional ischaemia scenario (rows), the presence of VT/VF was quantified with healthy (left), healthy progressing to ischaemic after VT/VF induction (centre), and ischaemic (right) remote myocardium. White cells represent no arrhythmia, whereas light and dark-coloured cells indicate transient and persistent VT/VF, respectively; LAD, left anterior descending artery; LCX, left circumflex artery; RCA, right coronary artery.

• 

  Figure 5. Quantification of ECG markers (amplitude spectrum area (AMSA); median slope (MS); dominant frequency (DF)) vs. number of phase singularities (PS) detected, considering multiple bipolar electrode configurations typically used in resuscitation protocols. (A) Apex-anterior, (A2) apex-anterior, normalised signal (max amplitude of 1 mV during sinus rhythm), (B) anterior-lateral, (C) apex-posterior, and (D) anterior-posterior. Measurements were obtained from ventricular tachycardia/fibrillation (VT/VF) time windows under variable ischaemic conditions, represented by the symbol colour: white, regional ischaemia + healthy remote myocardium (RM); grey, regional ischaemia + mild ischaemia in RM; and black, global moderate ischaemia. Cases considering the VT/VF onset (orange border) were excluded in the clustering according to ischaemic severity (black rectangles). Dashed red lines represent the minimal values associated with favourable prognosis in the literature (10, 49).

• 

  Figure 6. Progression of ventricular fibrillation (VF) wave dynamics and ECG signature in transmural regional ischaemia (LAD occlusion) at increasing ischaemic severity in the remote myocardium (RM): (A) VT/VF onset, (B) early VF, and (C) late VF. ECG and spectrogram of the signal at different stages with annotations for amplitude spectrum area (AMSA), median slope (MS), dominant frequency (DF), and the number of phase singularities (PS) (left panel). Membrane potential maps (right panel) for selected time intervals of the ECG (red rectangle) to illustrate representative wave dynamics at that VF stage. Emerging subendocardial wavefronts are identified with a star, while wavefronts in the epicardial surfaces show direction annotations shaped as arrows. ECG marker values are based on apex-anterior electrode configuration.