Abstract 2291: Myocardial Strain Measurements are not Equivalent across Techniques: a Validation of Two-Dimensional and Tissue Velocity Based Strain with Harmonic Phase Magnetic Resonance Imaging
Background: The measurement of myocardial deformation offers a means of quantifying regional LV function. Conventional tissue velocity (TVI) based strain suffers the problem of angle dependence, and 2D echocardiography based strain measurements have been developed to overcome this limitation. The aim of this study was to compare two such techniques (2DS1 & 2DS2) with TVI-strain, using harmonic phase magnetic resonance imaging (MRI, which offers excellent image quality) as the reference standard.
Methods: Thirty pts (mean age 62±11yrs) with known or suspected ischemic heart disease were included in the study. Peak-systolic longitudinal (ϵL) and radial strain (ϵR) were measured in 16 segments from the three conventional apical views using 2DS1, 2DS2, TVI (only ϵL) and MRI. Segments were also categorized as normal (n=247) or abnormal (n=102) based on gadolinium-contrast enhancement on MRI.
Results: Although there was significant correlation between the echo and MRI measures of ϵL, all three echo techniques underestimated ϵL compared to MRI (Table⇓). ϵL2DS2 was significantly higher than ϵL2DS1 (p=0.002) and had the best correlation with ϵLMRI. The correlation between ϵL2DS2 and ϵLMRI was better for abnormal than normal segments (p = <0.001). On ROC curve analysis, ϵL2DS2 was found to have the greatest AUC (0.84) compared to ϵL2DS1 (0.75; p= <0.001) or ϵLTVI (0.76; p= <0.001) for identification of dysfunctional segments. In case of radial strain, only ϵR2DS2 showed significant correlation with ϵRMRI.
Conclusions: Different 2D echocardiographic techniques provide significantly different measures of myocardial strain and the discrepancy is greater for radial strain than the longitudinal strain. These findings imply that strain measurement techniques cannot be used interchangeably.