Left atrial volume: comparison of 2D and 3D transthoracic echocardiography with ECG-gated CT angiography.

RATIONALE AND OBJECTIVES Left atrial volume (LAV) measurement by conventional two-dimensional (2D) transthoracic echocardiography (TTE) may be limited by the geometric model, by suboptimal definition of left atrial endocardium, or by chamber foreshortening. Three-dimensional (3D) TTE is posited to eliminate chamber foreshortening, and LAV measurement by 3D TTE should be more reflective of true LAV. The aim of this study was to compare conventional 2D TTE and newer 3D TTE for measurements of LAV to multidetector computed tomographic (MDCT) measurements using automated chamber reconstruction (ACR). MATERIALS AND METHODS Twenty-two subjects consented to undergo 2D TTE and 3D TTE immediately prior to or following coronary computed tomographic angiography. LAV was calculated from 2D TTE using the area-length method (ALM) and from 3D TTE with the ALM as well as with a 3D model. Electrocardiographically gated coronary computed tomographic angiography was performed in helical mode. LAV was measured using the ALM as well as ACR. RESULTS LAV was significantly smaller by 2D TTE (80 ± 21 mL) and 3D-TTE (90 ± 24 mL with the ALM, 61 ± 16 mL with the 3D model) compared to MDCT ACR (120 ± 30 mL) (P < .01). Correlation between MDCT ALM and MDCT ACR was excellent (mean Δ = -1.4 ± 14 mL, r = 0.91). Correlation with MDCT ACR was no better for 3D TTE (r = 0.80) than for 2D TTE (r = 0.80). CONCLUSIONS LAV is underestimated by both 2D TTE and 3D TTE relative to coronary computed tomographic angiography. Excellent agreement between the ALM and ACR with MDCT imaging suggests that the geometric model plays a negligible role in the underestimation of LAV. Underestimation of LAV by echocardiography is likely related to suboptimal definition of left atrial contour.