Validation of a novel modified wall motion score for estimation of left ventricular ejection fraction in ischemic and non-ischemic cardiomyopathy.

BACKGROUND

Visual determination of left ventricular ejection fraction (LVEF) by segmental scoring may be a practical alternative to volumetric analysis of cine magnetic resonance imaging (MRI). The accuracy and reproducibility of this approach for has not been described. The purpose of this study was to validate a novel segmental visual scoring method for LVEF estimation using cine MRI.

METHODS

362 patients with known or suspected cardiomyopathy were studied. A modified wall motion score (mWMS) was used to blindly score the wall motion of all cardiac segments from cine MRI imaging. The same datasets were subjected to blinded volumetric analysis using endocardial contour tracing. The population was then separated into a model cohort (N=181) and validation cohort (N=181), with the former used to derive a regression equation of mWMS versus true volumetric LVEF. The validation cohort was then used to test the accuracy of this regression model to estimate the true LVEF from a visually determined mWMS. Reproducibility testing of mWMS scoring was performed upon a randomly selected sample of 20 cases.

RESULTS

The regression equation relating mWMS to true LVEF in the model cohort was: LVEF=54.23-0.5761×mWMS. In the validation cohort this equation produced a strong correlation between mWMS-derived LVEF and true volumetric LVEF (r=0.89). Bland and Altman analysis showed no systematic bias in the LVEF estimated using the mWMS (-0.3231%, 95% limits of agreement -12.22% to 11.58%). Inter-observer and intra-observer reproducibility was excellent (r=0.93 and 0.97, respectively).

CONCLUSION

The mWMS is a practical tool for reporting regional wall motion and provides reproducible estimates of LVEF from cine MRI.