55Non-invasive ECG-imaging for identification of atrial arrhythmogenic low voltage substrate in patients with persistent atrial fibrillation

Abstract

Abstract Background/Introduction Left atrial (LA) fibrosis is associated with increased arrhythmia recurrence rates after pulmonary vein isolation (PVI) and increased stroke risk in patients with atrial fibrillation (AF). So far, detection and quantification of LA fibrosis is only feasible by invasive electrophysiological mapping of low-voltage-substrate (LVS) or delayed enhancement areas in MRI. Purpose The aim of this study was to assess the distribution and extent of atrial fibrosis by non-invasive ECG-Imaging (ECGI) in patients with persistent AF prior to PVI. Methods Thirty-seven consecutive patients (66 ± 9 years, 84% male) presenting for their first PVI were included. Patients with AF were cardioverted into sinus rhythm (SR). One day prior to AF ablation procedure, patients underwent ECGI in SR using the 252-electrode-array (CardioInsight) and a low-X-ray-dose, non-injected cardiac CT-scan to assess the relationship between ECGI-electrodes and cardiac epicardial structures. Prior to PVI, high-density biatrial voltage and activation maps were acquired in SR (CARTO-3). Localization and extent of atrial LVS (relevant fibrosis: LA-LVS: ≥5cm2 at <0.5mV threshold) and biatrial activation times depicted by CARTO were compared with atrial activation/conduction times assessed by non-invasive ECGI. Presence of LA-LVS was classified according to its extent into 3 stages and compared to the inter- and intraatrial conduction delay in ECGI. Results Relevant atrial fibrosis was found in 17/37(46%) patients. Presence of biatrial LVS resulted in a linear increase of the biatrial activation time in CARTO-SR-maps (146 ± 18ms in patients without LVS vs 184 ± 27ms in patients with LVS, p < 0.001) and in non-invasive ECGI (133 ± 11ms vs 170 ± 20ms, p < 0.001). Both the extent of biatrial LVS and invasively measured total activation time correlated well with non-invasive total atrial conduction time (TACT) in ECGI (r = 0.91 and r = 0.82, respectively, figure). Moreover, the extent of LA-LVS showed an excellent correlation to TACT in ECGI (r = 0.89). A combination of inter-atrial (RA-LA) conduction delay and TACT in ECGI allowed to quantify the extent of LA-LVS and to distinguish between three stages of LA-LVS: Stage 1 (minimal LA-LVS: 1 ± 2cm2): ECGI revealed rapid RA&LA activation with short TACT 132 ± 9ms; Stage 2 (moderate LA-LVS: 14 ± 8cm2 involving the anteroseptal LA) was associated with delayed LA activation and prolonged TACT measuring 161 ± 7ms; Stage 3 (extensive LA-LVS involving the anteroseptal and posterior LA: 26 ± 17cm2) was characterized by a significantly delayed LA activation with a TACT of 178 ± 24ms in ECGI. Conclusion Analysis of interatrial conduction delay and total atrial conduction time (TACT) in non-invasive ECGI allows accurate staging of patients with arrhythmogenic atrial LVS who present an increased risk for arrhythmia recurrences and stroke. Abstract Figure.