Non-Invasive Evaluation of Heart Function with Four-Dimensional Echocardiography

Abstract

Background: The aim of this study is to assess the accuracy and feasibility of left ventricular systolic function determined by four-dimensional echocardiography (4DE).

Methods: Latex balloons were sewn into the left ventricle (LV) of 20 freshly harvested pig hearts which were then passively driven by a pulsatile pump apparatus. Global longitudinal strain (GLS), global circumferential strain (GCS), global area strain (GAS) and left ventricular ejection fraction (LVEF) derived from 4DEand two-dimensional echocardiography (2DE)-derived LVEF were quantified at different stroke volumes (SV) 30-70 ml and correlated with sonomicrometry data.

Results: In all comparisons, GLS, GCS, GAS, 2DE-LVEF, and 4DE-LVEF demonstrated strong correlations with sonomicrometry data (r = 0.77, r = 0.89, r = 0.79, r = 0.93, r = 0.96, all P <0.001). Bland-Altman analyses showed slight overestimations of echo-derived GLS, GCS, 2DE-LVEF and 3DE-LVEF over sonomicrometry values (bias = 2.88, bias = 3.99, bias = 3.37, bias = 2.78, respectively). Furthermore, there is better agreement between GCS, 4D LVEF and sonomicrometry values compared with GLS and 2D LVEF.

Conclusion: Four-dimensional echocardiography accurately assesses LV function. GCS derived by 4DE is a potential alternative parameter to quantify LV systolic function.

Fig 1. Strain Values derived by three-dimensional speckle tracking program: 3D datasets were displayed in 2 cross-sectional planes.

Panel A and B showed longitudinal apical four-chamber and two-chamber views displayed on the right side. 3 short-axis views were displayed on the left side, which were redefined by adjusting the lines in both apical longitudinal views at different short-axis levels (apex, middle, base) to get clear cardiac borders. The region of interest was defined by points at the right and left mitral annulus as well as the apex at end-diastole in apical four-chamber and two-chamber views; the program automatically generated endocardial and epicardial border lines which were manually adjusted in all long-axis planes and short-axis planes. 3D strain parameters could be directly acquired in an exported Excel spreadsheet.

Fig 2. Linear correlation and Bland-Altman analyses of 2D LVEF, 4D LVEF and EF derived by sonomicrometry volume data.

A. The correlation analysis of 2D LVEF and sonomicrometry values; B. The correlation analysis of 4D LVEF and sonomicrometry values; C. Bland-Altman analysis of 2D LVEF and sonomicrometry values; D. Bland-Altman analysis of 4D LVEF and sonomicrometry values.

Fig 3. The correlation analysis of strain values derived by 3D speckle tracking program and EF obtained by sonomicrometry data.

A. The correlation of 3D GLS and sonomicrometry data; B. The correlation of 3D GCS and sonomicrometry data; C. The correlation of 3D GAS and sonomicrometry data.

Fig 4. Bland-Altman analysis of strain values by 3D speckle tracking program and strain value derived by sonomicrometry data.

A. Bland-Altman analysis of 3D GLS and strain value derived by sonomicrometry data; B. Bland-Altman analysis of 3D GCS and strain value derived by sonomicrometry data.

Fig 5. Line regression analysis of strain values obtained by 3D speckle tracking program and SV.

A. Line regression analysis of 3D GCS and SV; B. Line regression analysis of 3D GAS and SV; C. Line regression analysis of 3D GLS and SV.

Fig 6. Correlation analysis of strain values by 2D, 4D speckle tracking program and sonomicrometry data.

A. Correlation analysis of 2D GLS and strain value derived by 4DE; B. Correlation analysis of 2D GLS and sonomicrometry data.