Biomedical Engineering Reference
In-Depth Information
FIGURE 11.10
18
One slice of a PET [
F]-DG scan of the brain is shown. The image on the left was recon-
structed with a conventional iterative reconstruction algorithm and the image on the right
was reconstructed using anatomical priors (gray matter, white matter, skull, CSF, and sub-
rachnoid space segmentation). The improvement in resolution can be seen in the “guided”
reconstruction. (Image courtesy of Dr Babek Ardekani, University of Technology, Sydney).
it against the measured projection data. This procedure is performed itera-
tively until a match between the measured and the estimated data is
achieved. The simplest example of such a guide is to use the outline of the
patient's body, which can be measured in a variety of ways (Compton scatter
window,
47
transmission scan, CT scan, etc.). By ''encouraging'' the solution to
match in some prescribed way the known anatomical boundaries of the
object, higher resolution reconstructions can be achieved.
48,49
To do this it is
necessary to have coregistered, segmented data which have been classified
according to the expectation of their functional uptake of the radiotracer.
50
An
example of the improvement achievable using anatomically guided recon-
struction is shown in Figure 11.10.
11.5
Spatial Normalization
Spatial normalization refers to a type of intersubject registration involving the
transformation of images into a common, predefined standard space, which
often implies a distortion of the original organ. The purpose of doing this is to
be able to compare data from one subject with those from the same subject
studied at a different time, or against other subjects. An important applica-
tion for this is to compare a patient scan against a database of normals having
the same scan. For this application, the data need to be transformed into a
standard representation.
One of the first uses of spatial normalization applied to nuclear medicine
images was in the analysis and display of myocardial perfusion scans. This is
one of the most frequently performed diagnostic tests in clinical nuclear med-
icine. The aim of the test is to assess areas of compromised blood flow in the
