Radiotherapy and hybrid imaging

Ackerly, T 2011, Radiotherapy and hybrid imaging, Doctor of Philosophy (PhD), Medical Sciences, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Ackerly.pdf Thesis application/pdf 7.62MB
Title Radiotherapy and hybrid imaging
Author(s) Ackerly, T
Year 2011
Abstract Radiotherapy and Hybrid imaging deals with three coupled problems associated with the use of functional imaging in radiotherapy.
First is the tumour threshold of significance within a positron emission tomography image, abstracted to the threshold of motionless spheres of uniform activity. For two dimensional positron emission tomography images, a simple model of the imaging process as a Gaussian blurring is sufficient to explain the variation of the thresholds with sphere volume qualitatively, adding a simple perturbation for noise allows the threshold to be predicted to two significant figures from the spatial resolution and nois characteristics of the scanner.
Second is the problem of information transfer. No modern radiotherapy centre can function without information transfer mediated by the Digital Imaging and Communication in Medicine protocol. It is shown in detail how a hospital can develop the organic expertise to monitor and decode computer network communication between protocol endpoints to identify sources of communication error, implement a solution using local resources and/or resolve otherwise intractable problems.
Third is the problem of registration of functional imaging to the treatment isocentre. The particular problem treated is the 2D-3D registration problem for intracranial lesions, where the the 3D image set is a stereotactic CT scan, and the 2D images are paired non-orthogonal oblique kV images obtained in the treatment room. A method of inference based on the principle of rotational invariance is developed, tested and applied. Clinical data from more than a year of treatments, over 16,000 image pairs, was analysed. The standard error in the reported positions with the system is found to be 0.3-0.4 mm. Estimates of the probability of positioning the patient within tolerance are predicted for a wide range of tolerance parameters to allow determination of the expected efficiency. The variation of accuracy with tolerance parameters is also predicted and rendered as probability distribution functions around the isocentre. The variation of accuracy with the number of iterative re-positionings required to achieve tolerance is also investigated and rendered in the same way. No clinically useful variation of accuracy with distance moved is observed, and an increase in accuracy after the first two fractions is observed, which is considered due to faster treatment times. The standard error of determined rotations is found to be 0.2°, which agrees with values in the literature. It is expected that a perfect system for measuring the standard error of the reported patient position in the instant they are determined existed, it would measure standard error as 0.20 and 0.24 mm in the superior-inferior and left right axes respectively. The standard error in the anterior posterior axis would be 0.14 mm, excluding the room laser calibration error, which has been determined indirectly as 0.17 mm, which is within error of a direct measurement of 0.18 mm noted in the literature.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Medical Sciences
Keyword(s) Radiologic Physics
Positron-Emission Tomography
PACS (Radiology)
Radiotherapy Image-Guided
Version Filter Type
Access Statistics: 228 Abstract Views, 1343 File Downloads  -  Detailed Statistics
Created: Wed, 12 Sep 2012, 12:27:24 EST
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us