Background: Cardiac fibrillation refers to rapid, irregular, and uncoordinated excitation of myocytes and can occur in both atria and ventricles. During fibrillation, machine learning algorithms can identify regions of spatiotemporal dispersion (stD), which are believed to be key drivers of arrhythmias. A first randomized trial targeting such regions by ablation has demonstrated improved rhythm control in atrial fibrillation (AF). However, the clinical benefit in broader, all-comer populations remains uncertain.
Methods: We retrospectively analysed patients with atrial (AF) and ventricular fibrillation (VF) who underwent high-density electroanatomical voltage, activation, and stD mapping of the left atrium or left ventricle, respectively. In a subset of AF patients, arrhythmia was re-induced by burst pacing, and stD remapping was performed. In all patients, targeted dispersion ablation was performed in conjunction with standard treatments, including pulmonary vein isolation (PVI) and scar homogenization. Clinical data, biomarkers, echocardiographic findings, and intraprocedural parameters were collected as indicators of remodeling.
Results: Our study included 9 patients with paroxysmal and 18 with persistent AF (cf. Fig. 1A), of whom 81% exhibited low-voltage areas (LVA) and 93% stD regions. LVA and stD burden correlated strongly (τ=0.412, p=0.007; cf. Fig. 1B). stD was confined to LVA in 56% of patients, while in 44%, it extended into regions of normal voltage. Patients with persistent AF exhibited a significantly higher number of dispersion sites (21±3) compared to those with paroxysmal AF (6±2). At follow-up, AF recurred in 33% of patients with paroxysmal AF and 60% of those with persistent AF.
In six persistent AF patients (age 66±9 years, BMI 30±3 kg/m², LAVI 41±8 mL/m², LVA burden 28±19 %), stD remapping was performed (c.f. Fig. 1C). The number of stD annotations decreased from 24±19 to 14±10 upon AF reinduction (2 patients showed an increase). The overall distribution of stD remained consistent. Despite similar coverage and mapping durations (14-22 min per map), remapping revealed new stD areas in patients, independent of mean LA pressure, LVA burden, or PVI.
During VF ablation (c.f. Fig. 2), stD mapping was generally poorly tolerated, even though both patients were supported by a left ventricular assist device. In the first patient (DCM post-myocarditis), no LVA was identified; however, stD mapping revealed a single proarrhythmogenic region. Targeted ablation of this site resulted in complete arrhythmia suppression, with no recurrence observed at the six-month follow-up. The second patient, who had ischemic cardiomyopathy with a large inferior LVA and ventricular aneurysm, showed stD both within the scar and in the pericannular region. Despite extensive scar homogenization and successful silencing of the stD regions, VF recurred after six months. The recurrence involved multiple complex tachycardias, all originating from the extensive inferior LVA.
Conclusion: Dispersion mapping reliably identifies proarrhythmogenic substrates in both atrial and ventricular fibrillation. While targeting these regions may contribute to improved rhythm control, it appears to be insufficient when used as a stand-alone mapping and ablation strategy.
