How does computer-assisted digital wall motion analysis influence observer agreement and diagnostic accuracy during stress echocardiography?

Abstract

This study assessed interobserver and intraobserver variation and diagnostic accuracy during 25 dipyridamole stress echocardiography tests interpreted with different analysis systems: a) computer display of high frame rate digital cineloops (47 frames/s); b) computer display of lower frame rate digital cineloops (24 frames/s); and c) videotape recordings. The majority of the patients (84%) had documented coronary artery disease with baseline wall motion abnormalities due to previous myocardial infarctions and/or coronary bypass surgery, thus comprising a population with difficult interpretation of stress echocardiography. Diagnostic accuracy was assessed using coronary angiography as reference method. Interobserver and intraobserver agreement was highest when analysis was performed from computer-displayed cineloops, 96 and 92%, respectively, compared to 84 and 80% respectively, using videotape recordings. Sensitivity for identification of coronary artery stenosis was similar using digital cineloops with high frame rate or videotape recordings (67% to 80% for both systems), and tended to be lower using cineloops with lower frame rate for analysis (53%). Inter- and intraobserver differences for wall motion score index were not significantly influenced by the analysis system. We conclude that computer assisted analysis with high frame rate of the displayed cineloops provides optimal observer agreement and diagnostic accuracy in the same range as videotape analysis in patients undergoing stress echocardiography.