PO-01-195 PHENOTYPING PERSISTENT AF WITH ELECTROGRAPHIC FLOW MAPPING

Abstract

There is broad consensus that differentiating AF into the broad categories of “paroxysmal AF” vs “persistent AF (PerAF)” vs “LS-PerAF” AF is a relatively crude means to differentiate AF types. Clearly, there is a pressing need for better strategies to phenotype AF. Electrographic Flow (EGF) mapping is designed to identify not only active sources of excitation, but also to estimate the consistency of observed atrial wavefront propagation. Electrographic flow consistency (EGFC) is computed from the Euclidean length of vector field estimates over time and may provide additional insight into an individual patient’s atrial fibrillation (AF) disease that in combination with EGF-identified sources may enable the phenotyping of PerAF patients. To 1) determine the relationship between EGFC and AF recurrence, and 2) propose a phenotyping framework for the treatment and prognosis based on functional mechanisms as described by EGF. In the randomized FLOW-AF trial, patients prospectively underwent PVI and subsequent EGF-guided source ablation in all cases when such sources were identified, and PVI-only when no sources were seen. In each patient, a series of 1-min EGF recordings were taken in multiple locations per atrium post-PVI but pre- any adjunctive or source ablation. Mean EGFC from these recordings were averaged for each basket position and these averages were averaged to determine an overall EGFC across all locations. A decision tree classifier was employed to determine the optimal splitting point below which EGFC scores were more commonly associated with recurrence within 12 months. Per FLOW-AF’s design, patients were divided into no sources (n=32) vs with sources above threshold randomized to EGF-guided ablation (n=21). The decision tree classifier identified 0.63 as EGFC above which recurrence is least common and below which recurrence is most common. Patients with high EGFC and no sources had 100% (10/10) freedom from AF (FFAF); patients with high EGFC and at least 1 ablated source had 88% (7/8) FFAF; patients with low EGFC and no sources had 50% (11/22) FFAF; and patients with low EGFC and at least 1 ablated source had 46% (6/13) FFAF. Using EGF mapping algorithms to detect the presence of functional mechanisms of AF, the clinically heterogeneous population of PerAF patients can be stratified into distinct EGF phenotypes that guide the minimum-required ablation strategy, and also provide post-ablation prognosis. A larger prognostication study is warranted.