Transesophageal 3-dimensional echocardiography: in vivo determination of left ventricular mass in comparison with magnetic resonance imaging.

The objective of this study was to assess the accuracy and reproducibility of transesophageal 3-dimensional echocardiography (3DE) in comparison with magnetic resonance imaging (MRI) for the in vivo calculation of left ventricular mass (LVM). In addition, mass values obtained by M-mode echocardiography were compared with those calculated by MRI. Three-dimensional reconstruction of the left ventricle was performed from a transesophageal and transgastric transducer position with a multiplane transducer in 20 patients. Left ventricular mass was calculated from both transducer positions by using slices of various thicknesses, ranging from 5 to 20 mm. Reproducibility was determined by 5 repeated measurements of mass in each of 5 randomly selected left ventricles. M-mode echocardiography was performed according to the method described by Devereux. For MRI, multiple short-axis views with 10-mm slice thickness were acquired in inspiration hold. Correlation was high for mass determined by 3DE and MRI (for 10-mm slice thickness: r = 0.99; y = 0.99 x - 0.7 g; standard error of estimate = 8.5 g; P <.001). There was no statistical bias, and the limits of agreement ranged from +/-16.4 g to +/-27.2 g, depending on the slice thickness. Variability was lowest for a slice thickness of 10 mm (SD +/- 8.2 g). The reproducibility of mass determination was excellent (mean width of the 95% CI 12.8 g). Left ventricular mass values calculated from the transgastric and transesophageal transducer position were not different from each other (mean bias 0.6 +/- 9.1 g; P = ns). M-mode-based LVM calculations showed systematic overestimation and large measurement variability (bias 23.7 g; 95% CI +/- 92.8 g). Compared with MRI, transesophageal 3DE is an accurate and reproducible method for the determination of LVM and clearly superior to M-mode echocardiography.