[Reconstruction techniques in spiral CT angiography]

Clinical applications of spiral CT angiography are mainly based on transverse CT scans which can be completed by several categories of additional reformations in order to provide the most accurate evaluation of the vessel(s)-of-interest. Two-dimensional reformations are the second most frequent category of reconstructions generated from a spiral data set, enabling the creation of planar (frontal, sagittal or oblique) or curved views of given vessel. Three-dimensional displays can also be generated from most post-processing consoles, namely surface shaded displays (SSD), maximum intensity projections (MIP) and images using the volume rendered technique (VRT). Shaded surface displays are based on a binary selection of voxels using a single or double thresholding whereas MIP images result from the selection of the voxels of highest attenuation value. Volume rendered images are generated by means of a probabilistic classification of voxels consisting of a systematic estimation of the amount of the tissue-of-interest in each voxel. In addition, a variable degree of opacity (or transparency) can be chosen for the vascular structures and/or the surrounding anatomical structures, thus avoiding interpretive difficulties due to superimposition. A fourth category of reconstruction, STS-MIP (i.e., sliding-thin-slab MIP) can be applied to the evaluation of peripheral intraparenchymal vasculatures.