Year
2021
Degree Name
Doctor of Philosophy
Department
School of Physics
Abstract
Prostate cancer is quickly becoming the most common form of cancer across the globe, and is commonly treated with low dose rate brachytherapy due to its curative measures and highly conformal dose delivery. It is important to ensure there is a means of real time monitoring of the dose and seed placements when radioactive seeds are implanted in the prostate gland during a low dose rate brachytherapy treatment. The BrachyView system presents as a unique system that provides the capability of 3D seed reconstruction within an intraoperative setting.
In this thesis the BrachyView system is tested for its suitability, accuracy and the system is further developed so that its application in real-time intraoperative dosime-try can become a reality. The system was tested with a clinically relevant number of seeds, 98, where previously the system had only been tested with a maximum number of 30 seeds. The BrachyView system was able to reconstruct 91.8% of implanted seeds from the 98 seed dataset with an average overall discrepancy of 3.65 mm without the application of the baseline subtraction algorithm, however with its application to the data the detection efficiency was improved to 100% and an overall positional accuracy of 11.5%, correlating to a reduced overall discrepancy of 3.23 mm, was noted. It was found that with seed numbers of 30 or lower that the addition of a background subtrac-tion algorithm was not necessary, whereas for datasets containing a clinically relevant number of seeds the application of a background subtraction algorithm was paramount to reducing the noise, scatter and means for identification of newly implanted seeds that may be masked by those seed previously implanted.
Recommended Citation
Brennen, Taylah, 3D BrachyView System, Doctor of Philosophy thesis, School of Physics, University of Wollongong, 2021. https://ro.uow.edu.au/theses1/1352
FoR codes (2008)
0299 OTHER PHYSICAL SCIENCES, 0801 ARTIFICIAL INTELLIGENCE AND IMAGE PROCESSING
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.