Computational simulations

Segmented images containing LGE were processed as individual short-axis slices and meshed (with computational geometry software Computational Geometry Algorithms Library (CGAL) (https://doc.cgal.org/) into triangular finite element models with maximum edge length 250 μm (9). A monodomain representation was used to simulate electrical activity with the Cardiac Arrhythmia Research Package (10). Conductivities were tuned to match experimentally observed conduction velocities (11), suitably modulated within scar regions (9). Patchy distributions of fibrosis within regions of LGE were represented using the percolation method (9,12, 13, 14).

Simulated programmed electrical stimulation was performed from an endocardial pacing location— consistent with respect to scar in each model (9)—to attempt to induce re-entry; 600 ms activity after the final extrastimulus was simulated, with 10 simulations performed per image slice, testing various scar microstructure combinations (9). For further details, see the Supplemental Appendix.

An additional simulated re-entries metric was derived using the computational simulations. This metric summed the total number of successfully induced re-entries that could be simulated from the 2-dimensional models that were derived from each patient’s LGE short-axis images. Finally, a mechanistic analysis of transmembrane potential dynamics was performed to probe the role of different LGE metrics defining the scar in unidirectional conduction block and re-entry initiation. For further details, see the Supplemental Appendix.