314 Automated left atrial volume measurement by two-dimensional speckle-tracking echocardiograpy. Feasibility, accuracy, and reproducibility

Abstract

Abstract Aims A by-product of left atrial (LA) strain analysis is the automated measurement of LA maximal volume (LAVmax), which may decrease the time of echocardiography reporting, and increase the reproducibility of the LAVmax measurement. However, the automated measurement of LAVmax by two-dimensional speckle-tracking analysis (2DSTE) has never been validated. Accordingly, we sought to: (i) assess the feasibility of automated LAVmax measurement by 2DSTE; (ii) compare the automated LAVmax by 2DSTE with conventional two-dimensional (2DE) biplane and three-dimensional echocardiography (3DE) measurements; and (iii) evaluate the accuracy and reproducibility of the three echocardiography techniques. Methods and results LAVmax (34–197 ml) were obtained from 198/210 (feasibility 94%) consecutive patients with various cardiac diseases (median age 67 years, 126 men) by 2DSTE, 2DE, and 3DE. 2DE and 2DSTE measurements resulted in similar LAVmax values (bias = 1.5 ml, limits of agreement, LOA ± 7.5 ml), and slightly underestimated 3DE LAVmax (biases = −5 ml, LOA ± 17 ml, and −6 ml, LOA ± 16 ml, respectively). LAVmax by 2DSTE and 2DE were strongly correlated to those obtained by cardiac magnetic resonance (CMR) (r = 0.946, and r = 0.935, respectively; P < 0.001). However, LAVmax obtained by 2DSTE (bias = −9.5 ml, LOA ± 16 ml) and 2DE (bias = −8 ml, LOA ± 17 ml) were significantly smaller than those measured by CMR. Conversely, 3DE LAVmax were similar to CMR (bias = −2 ml, LOA ± 10 ml). Excellent intra- and inter-observer intraclass correlations were found for 3DE (0.995 and 0.995), 2DE (0.990 and 0.988), and 2DSTE (0.990 and 0.989). Conclusions Automated LAVmax measurement by 2DSTE is highly feasible, highly reproducible, and provided similar values to conventional 2DE calculations in consecutive patients with a wide range of LAVmax.