Reproducibility and beat-to-beat variability of digital echocardiograms

To test the hypothesis that digital echocardiography (DE) is a reliable substitute for video tape (VT) for the interpretation of routine studies, we examined 51 pts with known or suspected cardiac disease. Images of the 4 standard views with and without color Doppler were recorded on VT and 4 separate digital studies were created: 3 sets of single cardiac cycle cine loops (SCCL) and 1 triple cardiac cycle cine loop (TCCL). All 5 recordings were analyzed separately by a single blinded observer for: 1) estimated LVEF, 2) wall motion score index (WMSI), 3) MR severity, 4) presence or absence of regional wall motion abnormalities (5 regions: ant, post, inf, lat, apical). The following were assessed: 1) variability among the 3 SCCL, 2) reproducibility of each SCCL compared to VT, 3) comparison between SCCL and TCCL (to test the advantages of longer loops). SCCL and TCCL correlated highly with VT for both WMSI (SCCL 1,2,3 vs VT r=.99; TCCL vs VT r=.99) and LVEF (SCCL1 vs VT r=.94; TCCL vs VT r=.93). Correlations among the 3 SCCL for WMSI were similarly high (r=.99 for all 3 pairs). Using VT as the standard, SCCL had an overall sensitivity of 97% and specificity 90% for the detection of abnormal wall motion. For individual regions (n=256), sensitivity ranged from 90% to 96% and specificity 91% to 100%. MR severity by DE was identical to VT in 36/51 pts, underestimated in 8 and over-estimated in 7 (in each case by 1 grade). In conclusion, 1) DE provides diagnoses that are highly correlated with standard VT, 2) beat-to-beat variability of SCCL is very low, 3) longer cine loops (3 cycles) offer no advantage over single cycle loops.

Duke Scholars

Author Joseph Andrew Kisslo Medicine, Cardiology