The goal of conformal radiation techniques is to improve local tumour control through dose escalation to target volumes while at the same time sparing surrounding healthy tissue. Respiratory motion is known to be the largest intra-fractional organ motion and the most significant source of uncertainty in treatment planning for chest lesions. A method to account for the effects of respiratory motion is to use four-dimensional radiotherapy. While analytical models are useful, it is essential that the motion problem in radiotherapy is addressed by both modeling as well as experimentally studies so that different obstacles can be overcome before clinical implementation of a motion compensation method. Validation of techniques aimed at measuring and minimizing the effects of respiratory motion require a realistic dynamic deformable phantom for use as a gold standard. In this work we present the design, construction, performance and deformable image registration of a novel breathing, tissue equivalent phantom with a deformable lung that can reproducibly emulate 3D non-isotropic lung deformations according to any real lung-like breathing pattern. The phantom consists of a Lucite cylinder filled with water containing a latex balloon stuffed with dampened natural sponges. The balloon is attached to a piston that mimics the human diaphragm. Nylon wires and Lucite beads, emulating vascular and bronchial bifurcations, were glued at various locations, uniformly throughout the sponges. The phantom is capable of simulating programmed irregular breathing patterns with varying periods and amplitudes. A deformable, tissue equivalent tumour, suitable for holding radiochromic film for dose measurements was embedded in the sponge. Experiments for 3D motion assessment, motion reproducibility as well as deformable image registration and validation are presented using the deformable phantom.

Treatment delivery verification is an essential step of radiotherapy. The purpose of this thesis is to develop new methods to improve the verification of photon and electron beam radiotherapy treatments. This is achieved through developing and testing (1) a way to acquire portal images during electron beam treatments, (2) a method to reconstruct the dose delivered to patients during photon beam treatments and (3) a technique to improve image quality in kilovoltage (kV) cone beam computed tomography (CBCT) by correcting for scattered radiation. The portal images were acquired using the Varian CL21EX linac and the Varian aS500 electronic portal imaging device (EPID). The EGSnrc code was used to model fully the CL21EX, the aS500 and the kV CBCT system. / We demonstrate that portal images of electron beam treatments with adequate contrast and resolution can be produced using the bremsstrahlung photons portion of the electron beam. Monte Carlo (MC) calculations were used to characterize the bremsstrahlung photons and to obtain predicted images of various phantoms. The technique was applied on a head and neck patient. / An algorithm to reconstruct the dose given to patients during photon beam radiotherapy was developed and validated. The algorithm uses portal images and MC simulations. The primary fluence at the detector is back-projected through the patient. CT geometry to obtain a reconstructed phase space file. The reconstructed phase space file is used to calculate the reconstructed dose to the patient using MC simulations. The reconstruction method was validated in homogeneous and heterogeneous phantoms for conventional and IMRT fields. / The scattered radiation present in kV CBCT images was evaluated using MC simulations. Simulated predictions of the scatter distribution were subtracted from CBCT projection images prior to the reconstruction to improve the reconstructed image quality. Reducing the scattered radiation was found to improve contrast and reduce shading artifacts. / MC simulations, in combination with experimental techniques, have been shown to be valuable tools in the development of treatment verification methods. The three novel methods presented in this thesis contribute to the improvement of radiotherapy treatment verification. They can potentially improve treatment outcome by ensuring a better target coverage.

Intensity modulated radiation therapy (IMRT) and volumetric modulated arc therapy (VMAT) have been introduced in radiation therapy to achieve highly conformal dose distributions around the tumour while minimizing dose to surrounding normal tissues. These techniques have increased the need for comprehensive quality assurance tests, to verify that customized patient treatment plans are accurately delivered during treatment. In vivo dose verification, performed during treatment delivery, confirms that the actual dose delivered is the same as the prescribed dose, helping to reduce treatment delivery errors. In vivo measurements may be accomplished using entrance or exit detectors. The objective of this project is to investigate a novel entrance detector designed for in vivo dose verification. This thesis is separated into three main investigations, focusing on a prototype entrance transmission detector (TRD) developed by IBA Dosimetry, Germany. First contaminant electrons generated by the TRD in a 6 MV photon beam were investigated using Monte Carlo (MC) simulation. This study demonstrates that modification of the contaminant electron model in the treatment planning system is required for accurate patient dose calculation in buildup regions when using the device. Second, the ability of the TRD to accurately measure dose from IMRT and VMAT was investigated by characterising the spatial resolution of the device. This was accomplished by measuring the point spread function with further validation provided by MC simulation. Comparisons of measured and calculated doses show that the spatial resolution of the TRD allows for measurement of clinical IMRT fields within acceptable tolerance. Finally, a new general research tool was developed to perform MC simulations for VMAT and IMRT treatments, simultaneously tracking dose deposition in both the patient CT geometry and an arbitrary planar detector system, generalized to handle either entrance or exit orientations. It was demonstrated that the tool accurately simulates dose to the patient CT and planar detector geometries. The tool has been made freely available to the medical physics research community to help advance the development of in vivo planar detectors. In conclusion, this thesis presents several investigations that improve the understanding of a novel entrance detector designed for patient in vivo dosimetry.

After a radiation dispersion device (RDD) event, there may be internally and/or externally contaminated victims. After the RDD event, victims may require immediate medical assistance prior to decontamination. The dose rates to which a healthcare provider is exposed due to the internal and external contamination of the victim were computed using Monte Carlo simulations and five anthropomorphic phantoms. The dose rates to which the victim is exposed due to his/her own external contamination were also computed. For the external contamination modeling, the contamination is assumed to be distributed over the entire exterior of the victimâ s body. The geometrical models of the human body were based on the MIRD stylized phantom. The specific isotopes considered were 60Co, 137Cs, 131I, 192Ir, and 241Am. The surface contamination was generated by creating a 2-mm thick layer adjacent to the outside of the skin of the victim and uniformly sampling the emissions of the radioactive sources throughout this volume. The attending healthcare provider was assumed to be standing 20 cm from mid-torso of the victim. The organ absorbed doses in both the contaminated individual and a healthcare professional were computed. The effective dose to the victim and the attending healthcare professional were computed using the tissue weighting factors in ICRP Publication 60. For example, the dose rate to a reference male healthcare provider from the victim six hours after the inhalation of one ALI by an adipose male victim will be 0.277 mSv/hr. In addition, the air kerma was computed at different distances from the surfaces of the victim phantom and ratios were generated for the air kerma and the effective dose due to the victim from the surface contamination on the victim.

Assaying internal contamination due to inhalation is a primary concern in developing emergency procedures related to Radioactive Dispersal Devices (RDD). One method of determining internal contamination makes use of a common medical instrument, a Thyroid Uptake System (TUS). The TUS used in this research has two collimators a thyroid uptake collimator and a bioassay collimator. Both collimators were considered and modeled in MCNP to be used in conjunction with six MIRD-type (Medical Internal Radiation Dose) phantoms. The collimators were placed in four positions on the phantoms the front right lung, the back right lung, the neck, and the thigh. Unit sources of Cs-137, Co-60, I-131, Ir-192, Am-241, and Sr/Y-90 were placed in the organs of the phantoms. MCNP particle tallies were performed over the detector crystal volume to determine the count-rate contributions from the unit source in each organ. Biokinetic modeling was performed using DCAL (Dose and Risk Calculation System) to generate coefficients to describe activity as a function of time in various organs. By folding the count-rate results with the organ concentrations, the detector response as a function of time after intake has been determined. This work was performed under funding provided by the Radiation Studies Branch of the Centers for Disease Control and Prevention.

Electron beam therapy planning and custom electron bolus design were identified as areas in which improvements in equipment and techniques could lead to significant improvements in treatment delivery and patient outcomes. The electron pencil beam algorithms used in conventional Treatment Planning Systems do not accurately model the dose distribution in irregularly shaped objects, near oblique surfaces or in inhomogeneous media. For this reason, at Christchurch Oncology Centre the TPS is not relied on for planning electron beam treatments. This project is an initial study of ways to improve the design of custom electron bolus, the planning of electron beam therapy, and other radiation therapy simulation tasks, by developing a system for the accurate assessment of dose distributions under irregular contours in clinically relevant situations. A shaped water phantom system and a diode array have been developed and tested. The design and construction of this water phantom dosimetry system are described, and its capabilities and limitations discussed. An EGS/BEAM Monte Carlo simulation system has been installed, and models of the Christchurch Oncology Centre linacs in 6MeV and 9MeV electron beam modes have been built and commissioned. A test was run comparing the EGS/BEAM Monte Carlo system and the CMS Xio conventional treatment planning system with the experimental measurement technique using the water phantom and the diode array. This test was successful as a proof of the concept of the experimental technique. At the conclusion of this project, the main limitation of the diode array system was the lack of data processing software. The array produces a large volume of raw data, but not enough processed data was produced during this project to match the spatial resolution of the computer models. An automated data processing system will be needed for clinical use of the array. It has been confirmed that Monte Carlo and pencil-beam algorithms predict significantly different dose distributions for an irregularly shaped object irradiated with megavoltage electron beams. The results from the diode array were consistent with the theoretical models. This project was an initial investigation. At the time of writing, the diode array and the water phantom systems were still at an early stage of development. The work reported here was performed to build, test and commission the equipment. Additional work will be needed to produce an instrument for clinical use. Research into electron beam therapy could be continued, or the equipment used to expand research into new areas.

The microscopic distribution of microspheres and the resulting radiation dose deposition patterns in human liver following hepatic arterial infusion of 90Y labelled microspheres have been investigated. Tissue samples from normal liver, the tumour periphery and tumour centre were taken from a patient following infusion of 3 GBq of 32 pm diameter resin microspheres labelled with 90Y as treatment for an 80 millimetre diameter metastatic liver tumour. Microspheres were found to deposit inhomogeneously in tissues, preferentially lodging in a region approximately 6 mm wide around the periphery of the tumour. A relative concentration of microspheres of 50 to 70 times that of normal hepatic parenchyma and 65 to 94 times that in the tumour centre was measured in this region. The deposition of microspheres in the tumour periphery was not uniform, and cluster analysis showed that the spheres could be classified into clusters. The number of microspheres in a cluster was skewed towards low numbers and cluster sizes varied from 20 pm to 1500 pm. Microsphere deposition in normal liver was demonstrated to be non-uniform, there being significant variations in concentration over distances on the order of 3 to 4 millimetres. The observed microsphere distributions in three dimensions were used to calculate radiation dose patterns, and the results showed that heterogeneous doses were delivered to all tissues. Within the tumour periphery average doses ranged from 200 Gy to 600 Gy with minimum doses between 70 Gy and 190 Gy. The maximum and minimum doses for the tumour centre sample were 920 Gy and 3.7 Gy respectively, the median dose was 5.8 Gy. In the normal liver sample the median dose was 7.3 Gy with maximum and minimum doses of 753 Gy and 5 Gy respectively. Less than 1% of the normal liver tissue volume received more than 30 GY, the level above which complications have resulted for ++ / whole liver exposure using external beam radiotherapy. These calculations suggest that preferential deposition of microspheres in the well vascularised periphery of large tumours will lead to a high proportion of the tumour volume receiving a therapeutic dose, with most of the normal liver tissue being spared substantial damage.

Thesis (M.Sc.)--University of Adelaide, Dept. of Physics and Mathematical Physics, 2003. / "August 2003" Bibliography: leaves 108-117.

Thesis (M.Sc.)--University of Wollongong, 2005. / Typescript. Includes bibliographical references: leaf 225-238.

Thesis (Ph.D.)--University of Wollongong, 2006. / Typescript. Includes bibliographical references: p. 161-174.