Comparison of two different speckle tracking software systems: does the method matter?

BACKGROUND

Echocardiographic speckle tracking strain has gained clinical importance. However, the comparability of measurements between different software systems is not well defined.

METHODS

In 47 healthy subjects left ventricular (LV) two-dimensional (2D) peak strain and time to peak strain (TTP) generated by EchoPAC (2DS) and velocity vector imaging (VVI) were compared. For each type of strain (longitudinal [LS], circumferential [CS], and radial strain [RS]) we compared global, anatomical level and segmental values.

RESULTS

When comparing 2DS to VVI, Pearson correlation coefficients (r) of global LS, CS, and RS were 0.68, 0.44, and 0.59, respectively (all P < 0.05). Correlation of global TTP was higher: 0.81(LS), 0.80 (CS), and 0.68 (RS), all P < 0.01. Segmental peak strain differed significantly between 2DS and VVI in 8/18 (LS), 17/18 (CS), and 15/18 (RS) LV segments (P < 0.05). However, segmental TTP significantly differed only in 5/18 (LS), 7/18 (CS), and 4/18 (RS) of LV segments. Similar strain gradients were found for both systems: apical strain was higher than basal and midventricular strain in LS and CS, with a reversed pattern for RS (P < 0.05).

CONCLUSION

TTP strain as well as strain gradients were comparable between VVI and 2DS, but most peak strain values were not. The software-dependency of peak strain values must be considered in clinical application. Further studies comparing the diagnostic and prognostic accuracy of strain values generated by different software systems are mandatory.