Agreement of two vendor-independent strain analysis software platforms in assessing left ventricular global...

The new version of EchoPAC platform was recently developed by General Electronics (GE) to provide 'vendor-independent' full-myocardial-layer left ventricular (LV) global longitudinal strain (LV-GLS). The agreement of the LV-GLS by two vendor-independent software platforms was investigated under diverse clinical situations.

Two-dimensional speckle-tracking LV-GLS was separately measured by two software platforms. LV-GLS values were compared as default setting of each software platform (GE full-myocardial-layer [GE-Full], and TomTec endocardial-layer [TomTec-Endo]). Agreements according to various conditions and type of echocardiography vendors were evaluated using Bland-Altman analysis and intraclass correlation coefficients (ICC). Inter-observer reproducibility of each software platform was assessed, and agreements were further evaluated in various subgroups.

One hundred five subjects were initial candidates for the current study (normal LV function without any cardiac pathology [n = 25], hypertrophic cardiomyopathy [n = 40], dilated cardiomyopathy [n = 25], or restrictive cardiomyopathy [n = 15]). After excluding seven subjects with inadequate tracking quality, 98 subjects were finally analyzed. The average LV-GLS was lower in GE-Full than in TomTec-Endo. Agreement between GE-Full and TomTec-Endo was excellent in general; while the greatest bias was observed in the hypertrophic cardiomyopathy group, with TomTec-Endo exhibiting greater LV-GLS values than GE-Full (bias -1.71, limits of agreement -6.02 to 2.59). Both platforms showed excellent inter-observer reproducibility (GE-Full, ICC 0.99; TomTec-Endo, ICC 0.91), and were in good agreements regardless of the echocardiography vendors or subgroups according to age, heart rate, myocardial wall thickness, or LV ejection fraction.

LV-GLS by GE-Full showed excellent agreement with that by TomTec-Endo under various cardiac conditions.