To improve the accuracy of measurements of left ventricular volume in the presence of an aneurysm, we used three-dimensional echocardiographic imaging to analyze the shape of left ventricles in 23 asymmetric model hearts with eccentric aneurysms of different sizes, shapes, and localizations. A standard 3.75 MHz ultrasound probe with a rotation motor device was used to obtain a three-dimensional data set. By rotating the probe stepwise 1 degree, 180 radial ultrasound pictures were digitized. On the basis of the three-dimensional data set, the following parameters were determined and compared with the dimensions of the model hearts obtained by direct measurement: total left ventricular volume (LVV), aneurysm volume, area of the aneurysm's base, the longest aneurysm long diameter, and the longest aneurysm cross diameter. In addition, quantification of LVV by three-dimensional echocardiography was compared with biplane two-dimensional echocardiographic measurement according to the disk method. Good agreements were found for LVV measured by both techniques, three-dimensional echocardiographic and direct measurement (mean of differences = 0.91 ml; SD of differences = +/- 6.23 ml; line of regression y = 1.07 x - 14.24 ml; r = 0.968; standard error of the estimate [SEE] = +/- 6.17 ml), aneurysm volume (mean of differences = 0.43 ml; SD of differences = +/- 2.14 ml; line of regression y = 1.05 x - 0.81 ml; r = 0.996; SEE = +/- 1.96 ml), area of the aneurysm's base (mean of differences = 0.24 cm2; SD of differences = +/- 1.72 cm2; line of regression y = 1.02 x - 0.02 cm2; r = 0.981; SEE = +/- 1.75 cm2), the longest aneurysm long diameter (mean of differences = -0.26 mm; SD of differences = +/- 1.60 mm; line of regression y = 0.97 x + 1.34 mm; r = 0.996; SEE = +/- 1.54 mm), and the longest aneurysm cross diameter (mean of differences = 1.35 mm; SD of differences = +/- 3.94 mm; line of regression y = 0.95 x + 3.17 mm; r = 0.941; SEE = +/- 3.99 mm). In contrast, in these extremely asymmetric-shaped model hearts, agreement between biplane two-dimensional echocardiographic and both direct LVV measurement (mean of differences = 7.8 ml; SD of differences = +/- 20.8 ml; line of regression y = 1.48 x - 92.45 ml; r = 0.874; SEE = +/- 18.36 ml) and three-dimensional echocardiographic measurements (mean of differences = -7.6 ml; SD of difference = +/- 18.1 ml; line of regression y = 0.59 x + 80.98 ml; r = 0.908; SEE = +/- 10.36 ml) was poor. Thus tomographic three-dimensional echocardiography allowed accurate volume determination of asymmetric model hearts in the shape of left ventricles with eccentric aneurysms.