Filtering strategies of electrocardiographic imaging signals for stratification of atrial fibrillation patients

Abstract

Background and objectiveElectrocardiographic imaging (ECGI) has been used for guiding atrial fibrillation (AF) ablation, identifying reentrant activity by phase analysis with promising results. The objective of this study is to identify the best post-processing configuration for reentrant activity detection that better differentiates AF patients with different prognoses after catheter ablation. MethodsECGI signals of 24 AF patients before pulmonary vein isolation (PVI) were recorded. Patients were classified based on recurrence 6 months after PVI. Reentrant metrics were compared using 3 types of post-processing: none, sinusoidal recomposition (SRC), and narrow band-pass filtering centered at the highest dominant frequency (NB HDF). Different thresholds for rotor duration were also compared (0.5, 1, and 1.5 turns). ResultsThe use of raw ECGI signals with a threshold of 1 turn presented the optimal processing to identify PVI-positive responders (p < 0.05). NB HDF showed a better ability to find statistical differences between patients than SRC. ConclusionAggressive filtering of AF ECGI signals does not improve rotor identification to predict PVI outcome. Restrictive rotor duration thresholds diminish patient stratification. This definition of a post-processing strategy that allows patient stratification can be used for the improvement of the standard of care for finding the best candidates for PVI.