The optimisation of image quality in medical imaging techniques is a significant
factor in favourable patient prognoses. The number of counts in a nuclear medicine image
is one factor in determining the diagnostic value of the image. The current study aims to
determine the variation in counts in whole-body bone scan images with patient height and
weight. Three separate studies were undertaken as part of the investigation.
First, 65 whole-body bone scans were analysed together with patient height, weight,
age and sex. Weight was found to the most important anatomy influence on image counts.
However, significant influences from patient sex and age meant that a useful relationship
between image counts and patient anatomy based solely on height and weight could not be
determined.
For the second study, a model of General Electric Millennium MG gamma camera
was created and validated within the SIMIND Monte Carlo software. The results indicate
that the model is an accurate representation of the gamma camera.
Third, the 4D NCAT whole-body patient phantom was modified to represent the
average male and female clinical study participants. The phantoms were used in
conjunction with the gamma camera model to simulate the whole-body bone scan
procedure. The counts in the simulated images were consistent with the average measured
counts of the clinical study indicating that it is feasible to use the NCAT phantom for
nuclear medicine bone imaging. However, the phantom’s method of activity distribution
should be refined to allow a more realistic distribution of activity throughout the skeleton.
Identifer | oai:union.ndltd.org:canterbury.ac.nz/oai:ir.canterbury.ac.nz:10092/1586 |
Date | January 2007 |
Creators | McGurk, Ross James |
Publisher | University of Canterbury. Physics and Astronomy |
Source Sets | University of Canterbury |
Language | English |
Detected Language | English |
Type | Electronic thesis or dissertation, Text |
Rights | Copyright Ross James McGurk, http://library.canterbury.ac.nz/thesis/etheses_copyright.shtml |
Page generated in 0.0018 seconds