Mesh splitting based model of cardiac fibrosis

Introduction

We introduce a numerical model of diffuse cardiac fibrosis based on splitting a volumetric mesh along a two dimensional surface embedded within the mesh. The surface has tunable geometric properties that model fibrosis characteristics. A passive border zone (PBZ) was also imposed by a systematic element selection based on the mesh split.

Methods

A tetrahedral mesh of a ventricular slab was used. A custom algorithm generated the embedded two dimensional surface. Three variants of PBZs were imposed on the slab. A split slab with no PBZ was also studied. A Purkinje network activated the endocardium of the slabs with either all Purkinje myocardial junctions (PMJs) active or with PMJs on the fibrotic region inactive. The Purkinje ionic activity was modelled using the Trovato model. A fully connected slab with all PMJs active was used as a control. The ionic activity of the myocardium was modelled by the Ten Tusscher Panifilov model. An endo- and epi-cardial electrode array was simulated to measure unipolar electrograms (EGMs).

Results

The split slabs with a PBZ had faster conduction velocity within the fibrosis than the equivalent continuous slabs.

When some PMJs were inactive, epicardial EGMs had increased QRS amplitude when there was no PBZ where as the QRS amplitude were decreased when there was PBZ. Endocardial QRS amplitudes were decreased with and without a PBZ.

When all PMJs were active and without a PBZ all characteristics of control endo- and epi-cardial EGMs were maintained in the fibrotic slabs. With a PBZ, endocardial QRS amplitudes were decreased and epicardial control QRS amplitudes were maintained in the fibrotic slabs.

EGMs were fractionated when a PBZ was present and epicardial EGMs had more significant fractionation than endocardial EGMs. In the absence of a PBZ endocardial EGMs were fractionated when some PMJs were inactive.

The pronounced fraction observed in epicardial EGMs reflected epicardial depolarisation patterns whereas endocardial EGM fractionation reflected the intramural tortuous conduction within the fibrosis. Bipolar endocardial EGMs had late and long duration potentials that reflected the intramural tortuous conduction within the fibrosis.

The effects of the fibrosis on EGMs were dependent on endocardial depolarisation patterns as the Purkinje network with inactive PMJs resulted in more significant fractionation of EGMs than a Purkinje network with all PMJs active.

Conclusion

The model of fibrosis we introduced replicated EGM features that have been observed in patients with structural heart diseases. Unipolar epicardial EGMs and bipolar endocardial EGMs were better at indicating the presence of fibrosis than unipolar endocardial EGMs. The faster conduction within the fibrosis of the split slabs is in accordance with observations and hypotheses of fast conducting fibrotic cardiac tissue.