3D electroanatomical mapping is less sensitive to atrial remodeling in estimation of true left atrial volume than echocardiography

Abstract

Background: Left atrial (LA) enlargement has been identified as a predictor of worse clinical outcome after catheter ablation for atrial fibrillation (AF). We investigated the correspondence of LA size parameters assessed by echocardiography, CT and 3D electroanatomical mapping in patients with AF treated by catheter ablation. Methods: We analyzed echocardiographic LA volume measurements by disc summation method (LAVDISC), computed tomography (LAVCT) and 3D electroanatomical mapping (LAVCARTO) in 100 pts. (71% males; aged 63±8years; paroxysmal AF in 55% of patients). Results: Mean LAVDISC was 83±25ml (median: 115; IQR: 98-140ml), mean LAVCT was 120±34ml (median: 115; IQR: 98-140ml) and mean LAVCARTO was 123±36ml (median: 118; IQR: 99-132ml). Pearson's correlation coefficient between LAVDISC a LAVCT was 0.6 (p<0.0001) and between LAVCARTO and LAVCT was 0.79 (p<0.0001). There was a significant difference between the two correlation coefficients (p<0.004). The absolute difference between LAVCARTO and LAVCT (3.5 (95% CI -42 - 43) ml) was significantly lower (p<0.0001) as compared to LAVDISC and LAVCT (-39 (95% CI -102 - 24) ml). In opposite to LAVDISC, the bias between LAV obtained by CT and CARTO did not differentiate according to presence of spherical remodeling (1.7±28 vs. vs. 5.1±31ml). Only presence of sinus rhythm was significant and independent covariate of the difference between CARTO and CT-derived LAVs by multivariate regression analysis. Conclusions: Even though LA volumes evaluated by 3D-electroanatomical mapping have quite good accuracy, the precision is low. For volumes estimated by echocardiography, both precision and accuracy are low.