Use of an automatic application for wall motion classification based on longitudinal strain: is it affected by...

Assessing the quality of wall motion (WM) on echocardiograms remains a challenge. Previously, we validated an automated application used by experienced echocardiographers for WM classification based on longitudinal two-dimensional (2D) strain. The aim of this study was to show that the use of this automatic application was independent of the user's experience.

We compared the WM classifications obtained by the application when used by 12 highly experienced readers (Exp-R) vs. 11 inexperienced readers (InExp-R). Both classifications were compared with expert consensus classifications using the standard visual method. Digitized clips of cardiac cycles from three apical views in 105 patients were used for these analyses. Reproducibility of both groups was high (overall intra-class correlation coefficient: InExp-R = 0.89, Exp-R = 0.83); the lowest was noted for hypokinetic segments (InExp-R = 0.79, Exp-R = 0.72). InExp-R scores were concordant with Exp-R mode scores in 88.8% of segments; they were overestimated in 5.8% and underestimated in 3.2%. The sensitivity, specificity, and accuracy of InExp-R vs. Exp-R for classifying segments as normal/abnormal were identical (87, 85, and 86%, respectively).

Classification of WM from apical views with an automatic application based on longitudinal 2D strain by InExp-R vs. Exp-R was similar to visual classification by Exp-R. This application may be useful for inexperienced echocardiographers/technicians and may serve as an automated 'second opinion' for experienced echocardiographers.