Left ventricular (LV) volumes have important prognostic implications, but are commonly underestimated. We sought accuracy and reproducibility of LV volume measurement by live 3-dimensional (3D) echocardiography (3DE) and TI-201 single photon emission computed tomography (SPECT), compared with cardiac magnetic resonance imaging (MRI).

In all, 30 patients (age 62 ± 9 years, 23 men) underwent LV volume assessment with 3DE, SPECT, and cardiac MRI after myocardial infarction. LV volumes were measured using a semiautomated border detection algorithm for 3DE, gated SPECT software for SPECT, and a 3D display for MRI. Results of 3DE and SPECT volumes were compared with MRI as the standard of reference.

The 3DE volumes showed excellent correlation with cardiac MRI (end-diastolic volume [EDV], r = 0.90, P = .001; end-systolic volume [ESV], r = 0.94, P = .001), as did SPECT (EDV, r = 0.89, P = .001; ESV, r = 0.95, P = .001). However, both 3DE and SPECT underestimated LV volumes. The mean MRI EDV was 179 ± 56 mL compared with 3DE (mean difference, −10 ± 26 mL, P = .04) and SPECT (mean difference, −58 ± 28 mL, P < .001). There was a significant difference between SPECT EDV and 3DE (mean difference, −48 ± 31 mL, P < .001). The mean MRI ESV was 96 ± 54 mL and this was underestimated by SPECT (mean difference, −22 ± 19 mL, P < .001), but not by 3DE (mean difference, −0.9 ± 19 mL, P = not significant). ESV was also underestimated when SPECT was compared with 3DE (mean difference, −22 ± 27 mL, P < .001). The results of 3DE were reproducible with excellent intraobserver (ESV, r = 0.98, −2 ± 6 mL; EDV, r = 0.98, −1 ± 6 mL, P = .001) and interobserver (ESV, r = 0.97, −2 ± 6 mL; EDV, r = 0.95, −3 ± 10 mL, P = .001) correlation.

We have shown that 3DE is accurate and reproducible for the measurement of LV volumes for risk assessment in chronic ischemic heart disease and dilated cardiomyopathy. Furthermore, 3DE is more accurate than TI-201 SPECT with less underestimation of LV volumes.