Computational modeling identifies embolic stroke of undetermined source patients with potential arrhythmic substrate

Background: Late-gadolinium enhanced (LGE)-MRI has revealed atrial fibrotic remodeling in embolic stroke of undetermined source (ESUS) patients comparable to that observed in atrial fibrillation (AFib) patients. The absence of AFib in ESUS patients may be due to differences in the fibrotic substrate characteristics or the lack of triggers needed to initiate AFib. We used simulations in computational models reconstructed from LGE-MRI scans to study the role of atrial fibrosis as a pathophysiological link between AFib and ESUS. Methods: ESUS (per standard criteria) was verified by a neurologist. 45 ablation-naive AFib patients and 45 ESUS patients within three months of stroke underwent LGE-MRI for fibrosis assessment. Left atrial (LA) models were built from LGE-MRI scans. Fiber orientations were mapped into each LA model using universal atrial coordinates. Burst pacing from 15 known AFib trigger sites was used to test inducibility of arrhythmia sustained by reentry. Results: We observed sustained reentry in 23/45 (51%) ESUS and 28/45 (62%) AFib models. Overall, the fibrosis burden was significantly higher for patients in whom simulations showed inducibility (16.8 {+/-} 5.04% vs. 10.19 {+/-} 3.14%; P<0.0001); however, within the inducible and non-inducible sub-groups, there was no significant difference in fibrosis burden for ESUS vs. AFib patients (P=0.068 and P=0.58, respectively). This suggests that the presence of a pre-clinical substrate in ESUS is correlated with fibrosis burden, although exceptions to this supposition were not uncommon (i.e., inducible low-fibrosis and non-inducible high-fibrosis models). Conclusions: In this modeling study, pro-arrhythmic properties of fibrosis in ESUS and AFib are indistinguishable suggesting that some ESUS patients have a pre-clinical fibrotic substrate but do not have AFib due to a lack of suitable triggers.