A Novel Approach to Spectral MRI
Tiffany A. Fetzner
Abstract
The motivation behind this research was to develop and investigate
how effective a spatial-spatial-spectral algorithm, combining filtered-backprojection,
variable-bandwidth and spatial-spectral imaging techniques would be in
extracting spectral information from magnetic resonance images. The technique
employed in this research, which combines the fore mentioned imaging techniques
is believed to be original. It is known that the chemical shift artifact,
in a magnetic resonance image becomes large as bandwidth decreases.
A Ramlak filter and an inverse Radon transform was used to apply filtered-backprojection
and spatial-spectral research to magnetic resonance imaging. The
preliminary work in developing a spatial-spatial-spectral algorithm to
extract spectral information from variable-bandwidth magnetic resonance
images has been accomplished, but more work is needed to determine if this
technique can successfully resolve metabolites. The effectiveness
of this algorithm was found to depend on both the geometric complexity
of the object, and the amount of resolution lost when an inverse Radon
transform is applied to the image data. Both fat and water were found to
be resolvable with this technique. In addition the CH2 and CH3
peaks were also found to be resolvable. It has been determined that this
technique can resolve at least 25 Hz, but the highest possible resolution
has yet to be determined.
Table of Contents