Quantitative assessment of left ventricular volume and ejection fraction using two-dimensional speckle tracking echocardiography.

AIMS Two-dimensional speckle tracking echocardiography (2DSTE) allows measurements of left ventricular (LV) volumes and LV ejection fraction (LVEF) without manual tracings. Our goal was to determine the accuracy of 2DSTE against real-time 3D echocardiography (RT3DE) and against cardiac magnetic resonance (CMR) imaging. METHODS AND RESULTS In Protocol 1, 2DSTE data in the apical four-chamber view (iE33, Philips) and CMR images (Philips 1.5T scanner) were obtained in 20 patients. The 2DSTE data were analysed using custom software, which automatically performed speckle tracking analysis throughout the cardiac cycle. LV volume curves were generated using the single-plane Simpson's formula, from which end-diastolic volume (LVEDV), end-systolic volume (LVESV), and LVEF were calculated. In Protocol 2, the 2DSTE and RT3DE data were acquired in 181 subjects. RT3DE data sets were acquired, and LV volumes and LVEF were measured using QLab software (Philips). In Protocol 1, excellent correlations were noted between the methods for LVEDV (r=0.95), ESV (r=0.95), and LVEF (r=0.88). In Protocol 2, LV volume waveforms suitable for analysis were obtained from 2DSTE images in all subjects. The time required for analysis was <2 min per patient. Excellent correlations were noted between the methods for LVEDV (r=0.95), ESV (r=0.97), and LVEF (r=0.92). However, 2DSTE significantly underestimated LVEDV, resulting in a mean of 8% underestimation in LVEF. Intra- and inter-observer variabilities of 2DSTE were 7 and 9% in LV volume and 6 and 8% in LVEF, respectively. CONCLUSIONS Two-dimensional speckle tracking echocardiography measurements resulted in a small but significant underestimation of LVEDV and EF compared with RT3DE. However, the accuracy, low intra- and inter-observer variabilities and speed of analysis make 2DSTE a potentially useful modality for LV functional assessment in the routine clinical setting.