Patient-Specific Vascular Modeling - Microsurgery: Advances, Simulations, and Applications - page 81

The range of the use of MRI is limited because X-ray CT with
high spatial resolution and time resolution are widespread, but
MRI is useful to acquire the shape of blood vessel and organization
information. The resolution of MRI is above 0.7 mm.

3.3.2 Three-Dimensional Vessel Shape Reconstruction

from CT

In this section, we describe the method for reconstructing the three-
dimensional shape of the vascular lumen structure from the cross-
sectional image of human head, which was taken by the above-
mentioned multi-slice CT modality. The fault image of the CT image
is given as two-dimensional scalar place where the density value is
assumed to be a scalar. The extraction of the blood vessel area from
each fault image is achieved by giving a threshold processing to this
density value. The extraction of the blood vessel area from each fault
image is achieved by giving a threshold processing to this density value.
This processing is useful for separation of a desired object with speciic
density value from the surroundings within a fault image. The density
value in the CT images is given using the Hounsield unit. In general, the
density value changes according to the tissue type and organization.
Because blood is principally composed of water, the CT value
is adjacent to the soft tissue. Therefore, in luoroscopic images the
distinction between a blood vessel area and the surrounding area is
dificult. Therefore, to improve the CT value of blood in luoroscopic
images, a contrast medium is administered from the cervix part in the
blood stream. The blood vessel image is taken by luoroscopy of the
mixture liquid of blood and the contrast medium lows into the brain
blood vessel. In this case, it is about 150-250 HU, though the density
value of the blood vessel depends on the kind of the contrast medium.
However, the CT value is variable because the density of the contrast
medium is not uniform on the entire brain blood vessels. Moreover,
when taking a CT image, the phenomenon that is called PVE appears,
making the edge of the image duller.
As a result, when the blood vessel is extracted by the threshold
processing, a thin blood vessel is recognized thinner. Partial volume
effect makes the density value to decrease and to become a value
almost similar to the blood vessel in an outline part of the bone tissue
that has a high CT value (+1000 HU). Especially, the separation
of both areas becomes dificult for vessels adjacent to the skull.