Cardiac fibrosis is a major determinant of arrhythmogenesis, disrupting electrical conduction and creating substrates for reentrant circuits. While its influence is well known, most computational models still rely on oversimplified or idealized representations, failing to capture the true complexity of fibrotic microstructures.
In this study, we introduced a novel and versatile framework to simulate realistic fibrosis, capable of replicating the structural diversity seen in histological observations. Using an innovative Perlin noise-based generator, we synthetically created spatial patterns that reproduce four canonical fibrosis types — compact, diffuse, interstitial, and patchy — within fully fibrotic, ischemic two-dimensional domains. This generator enables the creation of customizable and physiologically plausible fibrotic geometries, marking a significant step forward in computational cardiac modeling.
We first estimated percolation thresholds for each pattern using cellular automata across multiple orientations and densities, identifying the critical point at which tissue conduction becomes fragmented. We then performed thousands of simulations using the GPU-accelerated MonoAlg3D solver, combining the monodomain model with a hypoxia-adapted Tusscher cellular model, to evaluate how each fibrosis type shapes the dynamics of excitation and reentry.
Our results showed that sustained reentries arise predominantly near the percolation threshold, with a strong dependence on the morphology and alignment of the fibrotic structures. Diffuse fibrosis, in particular, emerged as the most consistently arrhythmogenic, while interstitial and patchy fibrosis showed peak vulnerability when fibers were oriented perpendicularly to wavefront propagation. Compact fibrosis was the least conducive to reentry.
Two distinct reentry mechanisms were observed: circular reentrant loops in compact and diffuse fibrosis, and zig-zag conduction paths in interstitial and patchy fibrosis, shaped by anisotropic geometry.
This study paves the way for more accurate predictions of arrhythmic risk by incorporating realistic and customizable fibrosis into cardiac models. It also lays the foundation for personalized anti-arrhythmic strategies rooted in structural microdetail.