3-DIMENSIONAL RECONSTRUCTION OF CARDIAC V ALVES

The reconstruction of three-dimensional data sets from two-dimensional echocardiographic images offers several fundamental advantages: 1. mo re complete data than present in the few standard 2D-views; 2. off-lin e generation of any desired plane, cut, or perspective after the data set has been acquired; 3. access to quantitative parameters Like surfa ce areas (e. g., of valve leaflets or portions of leaflets), volumes, and others, without geometric assumptions. The mitral valve has been t he focus of several studies using various techniques of reconstruction of transthoracic or transesophageal images. These studies have shown the mitral annulus to be a non-planar, ''saddle-shaped'' structure, wi th an average distance of highest to lowest points of 14 mm in normals . This recognition of mitral annular non-planarity has led to a more s tringent echocardiographic definition of mitral valve prolapse. Furthe r studies have shown systolic shrinkage of mitral annular area by abou t 30% and systolic apico-basal translation of the annulus by approxima tely 1 cm in normals. In patients with dilated cardiomyopathy, the ann ulus is flattened, and both cyclic change in annular area and apico-ba sal translation are significantly reduced. 3D-studies of the left vent ricular outflow tract in hypertrophic obstructive cardiomyopathy allow measurement of outflow tract and leaflet surface areas and dynamic sp atial visualization of systolic anterior motion of the anterior mitral leaflet. Automated techniques to reconstruct the full grey value data set from a high number of parallel or rotational transesophageal plan es allow impressive visualization of normal and diseased mitral and ao rtic valves or valve prostheses, with special emphasis on generating ' 'surgical'' views and perspectives, which cannot be obtained by conven tional tomographic imaging. Current problems include loss of spatial r esolution in the 3D-reconstruction, lack of robust image segmentation, labor-intensiveness if manual contour tracing is used, and, conversel y, difficult access to quantitative data if automated reconstruction i s employed. A promising, very recent area of 3D-research in valvular h eart disease is the reconstruction of color Doppler maps into three-di mensional flow velocity fields, which may aid in the quantitation of c ardiovascular flows.