Abstract 17916: Effect of Spatial and Temporal Resolution on the Calculation of Global Longitudinal Strain by Three-Dimensional Echocardiography

Abstract

Background: Global longitudinal strain by three-dimensional speckle tracking echocardiography (3D ECHO) is a sensitive marker of left ventricular (LV) mechanical dysfunction and correlates with diminshed ejection fraction (EF) by cardiac magnetic resonance imaging (CMR). However, calculation of global longitudinal strain by 3D ECHO can be performed in several ways, some of which maximize the spatial information obtained from the 3D dataset and some of which correct for the temporal dispersion of data. The objective of this study was to explore the effects of spatial and temporal dispersion of 3D data on the calculation of 3D global longitudinal strain in a mixed population of normal and heart failure subjects.

Methods: 113 subjects (60% male) were prospectively studied. Of those, 69 were normal (Group 1), 16 had heart failure with preserved LV EF (Group 2) and 38 had heart failure with reduced LV EF (Group 3). Global longitudinal strain was derived by three methods: 1) by averaging the peak systolic values from 16 regional segments (Regional Average), 2) by computing the global minimum value by combining data from 16 regional segments and then deriving the longitudinal strain value (Global Minimum), thus increasing temporal resulution, and 3) by averaging the peak systolic values from 1200 voxels taken from the 3D data set and then deriving the global longitudinal strain (GLS), thus increasing spatial resolution at low temporal resolution. The three methods (Regional Average, Global Minimum and GLS) were correlated by Pearson's correlation coefficient. A normality test was performed to ensure a normal distribution of data. The groups were analyzed separately as well as in a combined fashion.

Results: All methods were highly correlated (p<0.0001) both separately and when the groups were combined. The r values for the combined data are r=0.88 for GLS vs. Regional Average, r=0.97 for Global Minimum vs. Regional Average, and r=0.90 for GLS vs. Global Minimum.

Conclusions: Low temporal resolution contributed less variability to measurement of 3D global longitudinal strain than high spatial resolution.