Abstract:
The clinical application of fluoroscopic imaging systems has undergone a drastic change from diagnostic procedures towards predominant use for real time control of interventions. Diagnostic imaging modalities, like CT and MR, have been developed from cross sectional imaging to volume imaging with isotropic resolution. Driving elements of this change have been multislice CT-technology and parallel acquisition techniques for MR. The correspondingly increased speed of data acquisition allowed the 3D-imaging of anatomical volumes with isotropic spatial resolution. As a necessary consequence also the rendering of the data, viewing and reading has shifted from slices to volumes. In view of those changes in the field of medical imaging the integration of three-dimensional, morphological as well as functional information into the interventional process has gained significant importance. This applies to the usage of three-dimensional data for planning the intervention, for guidance as well as for post-interventional control of the therapeutic outcome. Despite advances of other imaging modalities, fluoroscopy is the method of choice for real time control of interventional procedures and the deployment of interventional devices. Therefore, particularly from this application the need arises for registration of 2D-and 3D-images in combination with the localization of devices in space. For the general case of 2D-3D registration with data volumes from CT, MR or US corresponding methods did not yet penetrate into clinical use. However 3D-imaging of high contrast objects with rotating C-arms found a widespread application in neuroradiology. There it is used to image the vascular tree under selective, arterial injection to identify vascular abnormalities like stenoses, aneurysms or AVMs. Typically the 3D information is used to reliably quantify the extent of such lesions and to identify the feeding vessels for embolization. Currently the 3D data acquisition with rotating C-arms has several technical limitations , restricting the use to imaging of high contrast structures. Volume imaging with soft tissue resolution – e.g. tumour versus normal tissue – or functional information – e.g. myocardial function and viability – is the domain of CT, MR and nuclear imaging. The integration of corresponding 3D data into interventional or surgical procedures by 2D-3D registration bears a high potential to improve vascular and nonvascular
