Physiologically based nonlinear segmentation of LV borders

Left-ventricular wall motion is measured as the directional displacement of points spaced along the end-diastolic (ED) and end-diastolic (ES) borders. A mapping of uniformly spaced points along both borders does not reflect the actual motion that isolated myocytes in normal and abnormal segments undergo. In order to account for regional differences in segment shortening, the authors developed a method for assigning the spacing of ES points relative to equally spaced ED points. The method, which is based on local changes in ED and ES border curvatures, was used to calculate both regional wall motion and segment shortening. Tested on a database of 149 ventricles of differing states of normalcy and abnormalcy, both motion and shortening parameters proved statistically significant in separating patient groups. The authors conclude that changes in region ventricular shape can provide physiologically relevant information on regional wall motion and segmental shortening