Three-dimensional full automated software in the evaluation of the left ventricle function: from theory to clinical practice

Abstract

Left ventricular systolic function evaluation is an essential part of all transthoracic echocardiographic (TTE) studies. 3D echocardiography (3DE) is superior to 2D and is recommended as the method of choice. However, since it is time consuming and requires training, it is rarely performed. Different automatic analysis software tries to overcome these limitations but they need to be accurate and reproducible before they can be used clinically. The aim of this study was to test the accuracy and reproducibility of new 3D automatic quantitative software in everyday clinical practice. 69 patients referred to our Echo Lab for a clinically indicated echocardiographic examination were included. All patients underwent a full TTE with 3D image acquisition. Left ventricular volumes and ejection fraction (LVEF) were obtained using Heart Model software, and compared with conventional 3D volumetric data. Automated analysis was performed using three different sliders setting, with or without regional editing if necessary. 20 patients underwent a cardiac magnetic resonance (CMR) study the same day of the echo and automated measurements were also compared with a CMR reference. Intra- and inter-technique comparisons including linear regression with Pearson correlation coefficients and Bland–Altman analyses were calculated. Mean age of the patients was 59 years, with 49.3% male. The automated 3DE model demonstrated excellent correlation with the conventional 3DE measurements of LVEF, using three different sliders settings (r = 0.906; r = 0.898 and r = 0.940). Correlations with CMR values were very good as well (r = 0.888; r = 0.869; r = 0.913). Similarly, no significant differences were noted between the values of EDV and ESV, measured with the automated model or CMR, with excellent correlation (EDV: r = 0.892, r = 0.842, 0.910; ESV: r = 0.925, r = 0.860, r = 0.907). Finally, volumes calculated with the automated software were significantly greater than those obtained manually, but they showed a very good correlation (EDV: r = 0.875, r = 0.856, r = 0.891; ESV: r = 0.929, r = 0.879, r = 949). 3D automatic software for LV quantification is feasible and shows excellent correlations with both CMR and 3D echocardiography, considered the gold standard. No clinically relevant differences were noted when applying different border settings. This technique holds promise to facilitate the integration of 3D TTE into clinical practice.