Left atrial volume assessment in atrial fibrillation using multimodality imaging: a comparison of echocardiography, invasive three-dimensional CARTO and cardiac magnetic resonance imaging.

Left atrial size in atrial fibrillation is a strong predictor of successful ablation and cardiovascular events. Cardiac magnetic resonance multislice method (CMR-MSM) is the current gold standard for left atrial volume (LAV) assessment but is time consuming. We investigated whether LAV with more rapid area-length method by echocardiography (Echo-AL) or cardiac magnetic resonance (CMR-AL) and invasive measurement by 3D-CARTO mapping during ablation correlate with the CMR-MSM. We studied 250 consecutive patients prior to atrial fibrillation ablation. CMR images were acquired on 3T scanner to measure LAV by MSM and biplane area-length method. Standard echocardiography views were used to calculate LAV by biplane area-length method. LAV during ablation was measured by 3D-CARTO mapping. LAV was compared using intra-class correlation (ICC), Pearson's correlation and Bland-Altman plots. CMR-MSM was used as the reference standard. Mean LAV using CMR-MSM was 112.7 ± 36.7 ml. CMR-AL method overestimated LAV by 13.3 ± 21.8 ml (11.2%, p < 0.005) whereas 3D-CARTO and Echo-AL underestimated LAV by 8.3 ± 22.6 and 24.0 ± 27.6 ml respectively (8.7% and 20.0% respectively, p < 0.005). There was no significant difference between paroxysmal and persistent atrial fibrillation. CMR-AL and 3D-CARTO correlated and agreed well with CMR-MSM (r = 0.87 and 0.74, ICC = 0.80 and 0.77 respectively). However, Echo-AL had poor correlation and agreement with CMR-MSM (r = 0.66 and ICC = 0.48). Bland-Altman plots confirmed these findings. CMR-AL method may be used as an alternative to CMR-MSM, as it is non-invasive, rapid, and correlates well with CMR-MSM. LAV by different modalities should not be used interchangeably.