3D Stress echocardiography: development of novel visualization, registration and segmentation algorithms

Abstract

Stress echocardiography is a routinely used clinical procedure for diagnosing myocardial ischemia. In this procedure, images of the left ventricle are acquired both at rest and upon stress; these are then compared to detect the onset of new wall motion abnormalities post-stress or worsening of existing wall motion abnormalities at rest. Conventional two-dimensional ultrasound used as the imaging tool in this procedure makes image acquisition challenging and limits scanning the left ventricle thoroughly. It also leads to misaligned pre- and post-stress views. Furthermore, the diagnosis relies on visual estimates and remains subjective. Most of these problems can be overcome if the standard stress testing protocol is used together with the newly developed real-time three-dimensional (3D) ultrasound, a procedure we have termed 3D stress echocardiography. To realize the full potential of 3D stress echocardiography, we present development of new registration algorithms to align pre- and post-stress images, new visualization algorithms for interactive viewing of large 3D data sets, and new segmentation algorithms to allow making objective diagnoses. 3D Stress echocardiography, in conjunction with the advanced visualization and image analyses techniques we have proposed, has the potential to lead to improved diagnosis of myocardial ischemia and hence definitive treatment of coronary artery disease.