Feasibility of left atrial strain assessment using cardiac computed tomography in patients with paroxysmal atrial fibrillation.

Abstract

To evaluate the feasibility of left atrial strain (LAS) assessment using cardiac computed tomography (CT) in patients with paroxysmal atrial fibrillation (PAF). This retrospective single-center study included 98 patients with PAF who underwent cardiac CT and echocardiography before the first catheter ablation. LAS was analyzed using cardiac CT (CT-LAS) and speckle-tracking echocardiography (STE; STE-LAS). LA reservoir (LASr), conduit (LASc), and pump strain (LASp) were calculated by averaging LAS measured in 4- and 2-chamber views. The results were compared using Pearson's correlation coefficients, paired t-tests, and Bland-Altman analysis. Intraclass correlation coefficients (ICCs) were used to evaluate reproducibility. CT-LAS could be analyzed in all patients, while STE-LAS could be analyzed in 53 (54%) patients. LASr, LASc, and LASp showed significant correlations between CT- and STE-LAS: LASr, r = 0.68, p < 0.001; LASc, r = 0.47, p < 0.001; LASp, r = 0.67, p < 0.001. LASr, LASc, and LASp of CT- and STE-LAS were 23.7 ± 6.0% and 22.1 ± 6.7%, 11.1 ± 3.6% and 11.1 ± 4.1%, and 12.6 ± 4.6% and 11.0 ± 4.1%, respectively. LASr and LASp were significantly higher in CT-LAS than that in STE-LAS (p = 0.023 for LASr and p = 0.001 for LASp). CT-LAS showed excellent reproducibility. The intra- and interobserver ICCs were 0.96 to 0.99 and 0.89 to 0.90, respectively. CT-LAS was successfully analyzed in more patients than STE-LAS and was highly reproducible. The findings suggest that CT-LAS is feasible for patients with PAF.