Clinical application of reconstructed three dimensional magnetic resonance imaging in cardiovascular diseases.

点击翻译按钮获取中文摘要

To assess the cardiovascular anatomy, we have developed a low cost three dimensional (3D) reconstruction system using magnetic resonance (MR) imaging data on a NeXT workstation. The former version of this system had several problems. It needed many slices to reconstruct the whole heart and great vessels because we used multisectional and multiphasic EGG-gated MR images by the spin-echo technique in the only axial view, It was not able to reconstruct naturally the curved or meandered structures such as the aortic arch and the pulmonary artery. The structures behind other structures were not seen. To conqure these problem, we used axial, coronal and sagittal images to reduced the number of slices in each directional view, improved the moleling technique to reconstruct the curved or meandered structures naturally and applied the partly transparent technique to evaluate the overlapped structures. Determination of spacial location in the data of three axis views were done by entering the number of slices, slice thickness, gap thickness between the slices and the locational data of the first slice in each directional view. In order to improve the modeling, geometric centers of the coutours were connected by a 3D space curve, and the central axis of each cardiac structure was determined. The contours are projected on the perpendicular plane to the central axis and semiautomatically processed until the contours of one pixel are obtained. Then the surface rendering with transparency was performed and the each structures were superimposed. We tried these modified technique in patients with annuloaortic ectasia and corrected transposition of the greatarteries. With these improvement the curved or meandered structures were reconstructed naturally and superimposed structures were evaluated well.As we can reconstruct 3D images well with less number of slices in each directional view, it becomes possible to reconstruct 3D images with using the data of a routine MR examination.