Atrial tissue characterization by cardiac magnetic resonance and high-density mapping in patients with atrial fibrillation

Abstract

Abstract Funding Acknowledgements Type of funding sources: None. BACKGROUND Left atrial fibrosis is a marker of atrial disease and it has an important role in the pathophysiology of atrial fibrillation (AF). Late gadolinium enhancement cardiac magnetic resonance (LGE-CMR) is an emerging tool to detect left atrial fibrosis. However, data on the correlation between LGE-CMR detected fibrosis and low voltage areas to define fibrotic tissue is scarce. PURPOSE To assess the correlation and degree of concordance between LGE-CMR and high-density bipolar voltage mapping for the identification of left atrial abnormal tissue. METHODS Seven patients scheduled for AF ablation (including first and redo procedures) underwent a preprocedural 1.5 Tesla LGE-CMR including left atrial 3D inversion-recovery steady-state free precession sequence (ECG and respiratory triggering) 20 minutes after the injection of 0.2 mmol/kg of gadobutrol. A high-density electroanatomical voltage mapping was acquired with a 16-electrode grid configuration mapping catheter during sinus rhythm. LGE-CMR studies were analyzed off-line with an advanced image post-processing tool (ADAS 3D). Atrial wall intensity was normalized to blood pool, obtaining an image intensity ratio (IIR) value for each CMR point of the 3D model. High-density bipolar voltage maps and LGE-CMR 3D left atrial reconstruction were merged (figure, panel A). Voltage points were projected to the LGE-CMR left atrial 3D model, allowing point-by-point correlation analysis between voltage (log transformed due to non-normal distribution) with IIR. Left atrial fibrosis area and percentage were quantified using the standard cut-off values (bipolar voltage <0.5mV and IIR >1.2). We assessed the degree of concordance for normal and abnormal (fibrosis) tissue classification between the two techniques using different cut-off values (< 0.5mV and <1mV for bipolar voltage and >0.9, >1, >1.1 and >1.2 for IIR). RESULTS The average fibrosis area detected with LGE-CMR was lower than that detected with high-density bipolar voltage, using standard cut-off values (18.6 ± 5.7 cm2 vs. 40.6 ± 12.5 cm2, p = 0.13 respectively). There was a poor global point-by-point correlation between log-transformed voltage and IIR was r=-0.093, p < 0.001 (figure, panel B). The best concordance was obtained when using bipolar voltage and IIR of 0.5mV and 1.2, respectively (64.7 %; Kappa 0.101). However, the highest kappa index (0.142) for concordance was achieved with cutoff values of bipolar voltage <1mV and IIR >1, with an agreement percentage of 54.6%. CONCLUSIONS Left atrial tissue characterization between LGE-CMR and high-density bipolar voltage mapping showed significant but poor point-by-point correlation. Although the highest concordance was obtained using standard cutoff values, the Kappa index was best when applying non-standard cutoffs (1mV for bipolar voltage and >1 for IIR). Abstract Figure.