Computer-aided diagnosis via model-based shape analysis: Automated classification of wall motion abnormalities in echocardiograms

Rationale and Objective. Shape analysis of endocardial contour sequences from echocardiograms can provide classification of wall motion abnormalities (WMA). Materials and Methods. We previously reported on active appearance motion models (AAMM) for automated detection of endocardial contours in sequences of echocardiograms. The shape analysis of AAMM renders eigenvariations of shape/motion, including typical normal and pathologic endocardial contraction patterns. A set of stress echocardiograms (single-beat four-chamber and two-chamber sequences with expert-verified endocardial contours) of 129 infarct patients was split randomly into training (n = 65) and testing (n = 64) sets. AAMMs were generated from the training set and AAMM shape coefficients (ASCs) were extracted for all sequences and statistically related to regional/global visual wall motion scoring (VWMS) and volumetric parameters. Results. Linear regression showed clear correlations between ASCs and VWMS. Discriminant analysis showed good prediction by ASCs of both segmental (74% correctness) and global WMA (90% correctness). Volumetric parameters correlated poorly to regional VWMS. Conclusion. 1) ASCs show promising accuracy for automated WMA classification. 2) VWMS and endocardial border motion are closely related; with accurate automated border detection, automated WMA classification should be feasible. 3) ASC shape analysis allows contour set evaluation by direct comparison to clinical parameters.