Abstract 2341: Contrast Echocardiography improves Interobserver Agreement for Wall Motion Score Index and Correlation with Ejection Fraction
Background. Left ventricular ejection fraction (LV-EF) measurements are time-consuming and often a wall motion score index (WMSI) is provided as a surrogate marker for LV-EF. Unfortunately, there are only few data on the relation between the WMSI and LV-EF. In addition, poor non-contrast enhanced echo windows can make the WMSI (and LV-EF) unreliable. The value of contrast imaging for WMSI assessment was not investigated before.
Aim of the study: To compare segmental wall motion and WMSI interobserver agreement in patients who underwent both two-dimensional second harmonic (SH) and contrast left ventricular opacifi-cation (LVO) echocardiography.
Methods. The study comprised 100 consecutive patients (mean age 57 ± 13 years, 85% males) who underwent both SH and SonoVue LVO echocardiography for clinical evaluation of LV function. Two independent physicians assessed segmental quality and wall motion for both the SH and LVO studies according to a 17-segment model. Systolic wall motion was defined as (1) normokinesia, (2) hypokinesia (systolic inward endocardial motion <7mm), (3) akinesia, and (4) dyskinesia. LV-EF was assessed from the LVO images according to the biplane modified Simpson’s method by a third blinded physician.
Results: Of the 1,700 analyzed segments, 453 (26.6%) were poorly visualized with SH imaging, and 173 (10,2%) with LVO imaging (P <0.0001). The two independent observers agreed on segmental wall motion score in 1,299 of the 1,700 segments (agreement 76%, Kappa 0.60) with SH imaging and in 1,491 of the 1,700 segments (agreement 88%, Kappa 0.78) with LVO imaging. Interobserver correlation (R2) was 0.86 for the SH-imaged WMSI and 0.93 for the LVO-imaged WMSI. The limits of agreement for interobserver LVO-imaged WMSI (mean relative difference -1.0% ± 6.8%, agreement -14.6%, 12.6%) were lower than those for SH-imaged WMSI (mean relative difference -2.3% ±10.1%, agreement -22.5, 17.9). The LVO-imaged WMSI correlated well with LVO-imaged LV-EF (R2 = 0.71). LV-EF could be estimated according to the formula 1.01 - 0.32 x WMSI.
Conclusion: Echo contrast improves interobserver agreement for wall motion and WMSI scoring. The LVO-imaged WMSI correlates well with LVO-imaged LV-EF.