Biomedical Engineering Reference
In-Depth Information
operate in coincidence mode, so called gamma camera PET (GC-PET), are
more widely installed. [
18
F]-DG is an analogue of glucose which is potentially
useful for monitoring the response of tumors to anticancer therapy. The [
18
F]-
DG scans permit the assessment of any change in the pattern of uptake by the
tumors in response to the therapy. To do so in a quantitative manner, it is desir-
able to realign the patient's scans so that regions of interest placed over the
tumors on the initial study can be used to assess the change in [
18
F]-DG uptake
over time in subsequent studies. Typical time courses for these studies are to
scan the subject before treatment (baseline) and in the weeks and months after
therapy. [
18
F]-DG is particularly promising, as early changes in uptake may
indicate the efficacy (or not) of a particular therapy regimen. The alternatives
to [
18
F]-DG scans are x-ray CT or MRI, where change in tumor size is taken as
the measure of efficacy of treatment. Change in size, however, may take many
months to occur and therefore the nuclear medicine scan can give a much ear-
lier indication of the treatment's effectiveness.
In a collaborative study with the Oncology Imaging Group from the Royal
Free Hospital in London, image registration techniques have been employed
to half-body, attenuation-corrected [
18
F]-DG scans acquired with a gamma cam-
era PET system in patients with advanced cancers undergoing novel anticancer
therapy.
22
18
F]-DG uptake. The
patients are scanned at baseline and at 28 and 56 days after treatment. The first
step in the processing is to correct each data set for differences in the injected
amount of radioactivity and variations in normal physiological uptake. This is
done by normalizing the reconstructed data to a reference tissue which is taken
as being invariant over the different studies. As this is a study using antibody-
directed therapy, the liver has been chosen as the reference tissue against which
to normalize all [
The aim of the work is to assess changes in [
18
F]-DG uptake in the body. After this normalization the data
are realigned to the baseline scan using the mutual information algorithm of
Studholme et al.,
23
described in Chapter 3. This fully automated algorithm
operates on the volumetric half-body [
18
F]-DG scan and produces coregistered
data for further analysis. Regions of interest placed over the tumors on the
baseline scan are then transferred to the subsequent studies in the identical
locations to provide an objective assessment of the change in [
18
F]-DG uptake.
This removes any operator bias in defining regions of interest on individual
studies. The assessment will, of course, be complicated if the tumor has
changed in size over this time; however, in this approach an attempt has been
made to assess objectively the quantitative uptake by the tumors in the same
region. An example of one of these studies is shown in Figure 11.7.
11.4
Image Registration for Correction of Nuclear
Medicine Emission Data
Nuclear medicine images suffer from a number of degrading effects. These
include limited resolution, poor signal-to-noise ratio, and spatially varying
loss or corruption of signal due to photon interaction with matter. Anatomical
