LV STRAIN ESTIMATION FROM 4D ECHOCARDIOGRAPHY

Accurate quantification of regional left ventricular (LV) deformation is crucial for detection of patients with ischemic heart disease. In this paper, we propose an integrated image analysis system that accurately quantifies regional LV strain from four dimensional (3 spatial dimensions and time) echocardiographic (4DE) images. Intramural displacement is estimated using a phase-sensitive-correlation-based speckle tracking algorithm, that derives information from radiofrequency (RF) signal acquired from an ultrasound array, giving access to beam-formed acoustic data. Displacement information at myocardial surfaces is derived from B-mode images using generalized robust point matching (G-RPM) algorithm that takes curvature as an important shape feature and finds the correspondence between points with close curvature values. Intramural and surface displacement information is combined using an integrated system based on a biomechanical model to produce comprehensive, 4D estimates of myocardial strain and strain rate. Preliminary experiments on 3D echocardiographic images with RF signals show the potential of our method