Real-time three-dimensional (3D) echocardiography (RT-3DE) has emerged as a new technique in measuring left atrial and ventricular volume. However, the impact of cutting planes of RT-3DE on the accuracy of volume measurement in patients with a normal or enlarged heart is still unknown. We enrolled 30 normal subjects (control group) and 30 patients with heart failure (patient group). RT-3DE was performed to measure maximal volume of the left atrium (LAVmax) and left ventricular end-diastole volume (LVEDV) with 2-, 4-, 8- and 16-cutting planes, compared with cardiac magnetic resonance imaging (CMRI). In both groups, LAVmax by RT-3DE using 2- and 4-cutting planes was significantly underestimated (mean difference: -10.4 +/- 16.6 mL, p = 0.001 and -8.8 +/- 14.2 mL, p = 0.002 in the control group and -13.4 +/- 19.6 mL, p = 0.001 and -11.2 +/- 17.5 mL, p = 0.001 in the patient group, respectively). These differences became nonsignificant when 8- and 16-cutting planes were adopted (mean difference: -2.1 +/- 7.6 mL and -1.9 +/- 7.4 mL in the control group and -2.7 +/- 8.4 mL and -2.2 +/- 8.3 mL in the patient group, respectively). The agreement for LVEDV was acceptable when 4- or more cutting planes were used in the control group and when 8- or 16-cutting planes were used in the patient group. The time expense for data analysis of LAVmax with 8-image planes was only 7 +/- 4 min in the control group and 6 +/- 5 min in the patient group, almost halving that of the 16-image planes. Similarly, 4- and 8-cutting planes were required for an accurate measurement of LVEDV in the control and patient groups, respectively. In conclusion, RT-3DE with 8-cutting planes is both accurate and timesaving for measurement of LAVmax and LVEDV in patients with normal or enlarged left atria and ventricles.