In this thesis, the use of inhomogeneity corrections in intensity modulated radiotherapy (IMRT) inverse treatment planning is investigated. Firstly, the dosimetric consequences of CT beam hardening artifacts present on images used for treatment planning are estimated and found to be of little clinical significance (<1% dose difference). Secondly, experiments to evaluate the PEREGRINE Monte Carlo system (Nomos, Cranberry, PA) are undertaken for a 6 MV photon beam. The use of inhomogeneity corrections in clinical treatment planning is assessed for five clinical head and neck cancer cases. The cases are planned with the CORVUS optimization module and the dose distributions are then calculated with CORVUS and PEREGRINE in order to compare the two calculation techniques with emphasis on how each method handles tissue inhomogeneities. The plans are assessed in terms of dose, dose-volume distributions and the biological indices of TCP and NTCP. On average, PEREGRINE calculates a 1% lower mean dose to the GTV and a 2% lower mean dose to the CTV compared to the CORVUS calculations with EPL inhomogeneity corrections. In the last part of this work, quality assurance (QA) measurements are performed for a clinical case to investigate how the CORVUS and PEREGRINE calculations agree with the dose measurements on a QA phantom.

Dynamically delivered intensity modulated beams (IMBs) pose unique verification problems that may be addressed with the use of integrating continuous 3D dosimeters such as gel based Fricke dosimeters. Accurate knowledge of the ability of these dosimeters to measure adequately and precisely the delivered dose is a prerequisite for their clinical use. The magnetic properties of the ferrous and ferric ions present in the gel based Fricke dosimeter after its irradiation are the basis for the use of magnetic resonance imaging (MRI) in the measurement of dose. This thesis presents the investigation of a 3D gel based Fricke dosimetry system (Fricke-gel). A software system is developed and spin-lattice relaxation rate (R1) images are computed from MR images of irradiated Fricke-gel phantoms in order to quantify the dosimetric uncertainties resulting from the MR imaging system, from the gel itself, as well as from the external parameters. The sensitivity and the minimum detectable dose of the Fricke-gel dosimeter are determined. Validation of the dosimeter's capacity to measure dose distributions is made through measurement of percent depth dose curves (PDD's), and field profiles (open and wedged). An example of clinical utilisation of the Fricke-gel dosimeter is presented. Dose distributions are evaluated visually by 3D software tools and quantitatively analyzed by dose-volume histograms. Results show a good correlation between the Fricke-gel measured dose distributions and treatment planning software dose calculations.

Introduction. Most patients with early-stage endometrial carcinoma are treated with hysterectomy. The knowledge of the stage of disease could impact greatly on the eventual therapeutic decision. Numerous studies have investigated the use of endovaginal ultrasonography (US) and magnetic resonance (MR) imaging in the preoperative staging of endometrial carcinoma. Although a number of studies have demonstrated no significant difference in the accuracy of MR imaging and endovaginal US in assessing myometrial invasion, more recently, dynamic contrast-enhanced MR imaging has been found to be significantly superior to the non-enhanced T2-weighted MR imaging in the preoperative evaluation of endometrial carcinoma. / Purpose. (1) To evaluate prospectively the relative accuracy of endovaginal US (including Doppler) and MR imaging using phased array pelvic multicoil in the preoperative staging of endometrial carcinoma with histopathological correlation on the same patient population, (2) To determine whether the addition of contrast-enhanced MR sequences to high resolution fast spinecho (FSE) T2-weighted sequences, can improve the diagnostic accuracy of staging endometrial carcinoma. (Abstract shortened by UMI.)

<p> The following work describes investigations of spatial dosimetry using laser-induced fluorescence of a radio-fluorogenic detector embedded within water-equivalent media. The chemical composition of a gelatin-based coumarin-3-carboxylic acid detector was investigated and dose response characterized. Violet diode (405nm) excitation sources were explored and laser-induced fluorescence (LIF) employed to obtain the pattern of fluorescent emission yielding images of the integrated spatial dose distribution. The design of a three-dimensional reader is proposed to provide a foundation for future work. </p><p> Radio-fluorogenic processes create fluorescent products in response to ionizing radiation. Water radiolysis produced by ionizing radiation yields hydroxyl free radicals that readily hydroxylate coumarin-3-carboxylic acid to 7-hydroxy-coumarin-3-carboxylic acid, a derivative of umbelliferone. Umbelliferone is a known fluorophore, exhibiting peak excitation in the UV to near UV range of 365-405nm with a visible 445nm blue emission. Coumarin-3-carboxlyic acid has been studied in an aqueous gelatin matrix. </p><p> The radio-fluorogenic coumarin-gelatin detector has been shown to respond to an absorbed dose of ionizing radiation in a measureable manner. The detector was studied with respect to concentration of gelatin and coumarin in the presence of pH buffers. Dose response of the detector was investigated with regard to ionizing radiation type, energy, and rate of irradiation. Results demonstrate a functional detector. </p><p> Patterns of energy deposition were formed in response to ionizing radiation produced by a sealed-source of radioactive Ir-192 embedded in the gelatin matrix of the detector. Spatial distributions of absorbed dose were recorded and analyzed as a function of fluorescent emission. The distribution of energy deposition was imaged with LIF excitation by a divergent beam of 405nm light and determined by analysis of digital image pixel intensity values displaying the 445nm fluorescent emission. Results demonstrate spatial dosimetry proof of principle. </p><p> A basic dedicated reader system was fabricated employing LIF. Images of fluorescent emission excitation profiles were obtained in multiple aqueous samples and processed to obtain a dose response. Design of an optical reader system for the radio-fluorogenic detector is explained and a three-dimensional dosimetry system proposed. Three-dimensional imaging principles with LIF have been illuminated.</p>

<p> The primary goal of my thesis was to address my hypothesis that: there is preferential perfusion of the hindbrain regions, controlling autonomic function. To test this hypothesis I developed a system for delivering hypoxic challenges to volunteers while they were in the MRI. I developed NIRS protocols that allowed monitoring of the cerebellum. And I developed MRI methods that allowed for PC MRI to be used to monitor flow to the forebrain and hindbrain. Finally I combined these elements to investigate how the brain would react to hypoxia. Ultimately neither NIRS nor MRI detected systematic differences between the forebrain and hindbrain response to hypoxia but the developed methods are available for future studies that aim to explore the hemodynamic response in the developing brain or in adults with pathological conditions. </p>

<p> Various types of vascular diseases such as carotid stenosis, aneurysms, Arterio-venous Malformations (AVM) and Subarachnoid Hemorrhage (SAH) caused by the rupture of an aneurysm are some of the common causes of stroke. The diagnosis and management of such vascular conditions presents a challenge. In this dissertation we present a vascular analysis framework for Computer Aided Diagnosis (CAD) of the neuro-vasculature. We develop methods for 3D vascular decomposition, vascular skeleton extraction and identification of vascular structures such as aneurysms. </p><p> Owing to the complex and highly tortuous nature of the vasculature, analysis is often only attempted on a subset of the vessel network. In our framework we first, compute the decomposition of the vascular tree into meaningful sub-components. A novel spectral segmentation approach is presented that focuses on the eigenfunctions of the Laplace-Beltrami operator (LBO), FEM discretization. In this approach, we attain a set of vessel segmentations induced by the nodal sets of the LBO. This discretization produces a family of real valued functions, which provide interesting insights in the structure and morphology of the vasculature. Next, a novel Weighted Approximate Convex Decomposition (WACD) strategy is proposed to understand the nature of complex vessel structures. We start by addressing this problem of vascular decomposition as a cluster optimization problem and introduce a methodology for compact geometric decomposition. These decomposed vessel structures are then grouped into anatomically relevant sections using a novel vessel skeleton extraction methodology that utilizes a Laplace based operator. Vascular analysis is performed by obtaining a surface mapping between decomposed vessel sections. A non-rigid correspondence between vessel surfaces are achieved using Thin Plate Splines (TPS), and changes between corresponding surface morphologies are detected using Gaussian curvature maps and mean curvature maps. Finally, characteristic vascular structures such as vessel bifurcations and aneurysms are identified using a Support Vector Machine (SVM) on the most relevant eigenvalues, obtained through feature selection. </p><p> The proposed CAD framework was validated using pre-segmented sections of vasculature archived for 98 aneurysms in 112 patients. We first test our methodologies for vascular segmentation and next for detection. Our vascular segmentation approaches produced promising results, 81% of the vessel sections correctly segmented. For vascular classification, Recursive Feature Elimination (RFE) was performed to find the most compact and informative set of features. We showed that the selected sub-set of eigenvalues produces minimum error and improved classifier precision. This analysis framework was also tested on longitudinal cases of patients having internal cerebral aneurysms. Volumetric and surface area comparisons were made by establishing a correspondence between segmented vascular sections. Our results suggest that the CAD framework was able to decompose, classify and detect changes in aneurysm volumes and surface areas close to that segmented by an expert.</p>

PETSIM, a Monte Carlo simulation program of Positron Emission Tomography (PET) systems, was improved in terms of accuracy and efficiency. First, the accuracy, the speed and the ease of use of PETSIM were improved by using tabulated values of the Compton scattering and photoelectric absorption partial interaction attenuation coefficients for all common biological, collimator and detector materials. These were generated from chemical formula, or physical composition, and density of the absorbing medium. / Furthermore, simulations of PET systems waste considerable time generating events which will never be detected. For events in which the original photons are usually directed towards the detectors, the efficiency of the simulations was improved by giving the photons additional chances of being detected. For simulation programs which cascade the simulation process into source, collimation, and detection phases such as PETSIM, the additional detections resulted in an improvement in the simulation precision without requiring larger files of events from the source/phantom phase of the simulation. This also reduced the simulation time since fewer positron annihilations were needed to achieve a given statistical precision. This was shown to be a useful improvement over conventional Monte Carlo simulations of PET systems.

Dose predictions by Monte-Carlo (MC) techniques could alleviate the measurement load required in linac commissioning and clinical radiotherapy practice, where small or irregular electron fields are routinely encountered. In particular, this study focused on the MC calculation of cutout factors for clinical electron beams. A MC model for a Varian linac CL2300C/D was built and validated for all electron energies and applicators. A MC user code for simulation of irregular cutouts was then developed and validated. Supported by a home-developed graphical user interface, it determines in situ cutout factors and depth dose curves for arbitrarily shaped electron fields and collects phase space data. Overall, the agreement between simulations and measurements was excellent for fields larger than 2 cm. / The MC model was also used to calculate dose distributions with the fast MC code XVMC in CT images of phantoms of clinical interest. These dose distributions were compared to dose calculations performed by the pencil-beam algorithm-based treatment planning system CadPlan and verified against measurements. Good agreement between calculations and measurements was achieved with both systems for phantoms containing 1-dimensional heterogeneities, provided a minimal quality of the CT images. In phantoms with 3-dimensional heterogeneities however, CadPlan appeared unable to predict the dose accurately, whereas MC provided with a more satisfactory dose distribution, despite some local discrepancies.

Liquid ionization chambers have characteristics that can remedy some of the drawbacks of air-filled ionization chamber dosimetry: large sensitive volumes, fluence perturbations, and energy dependence. However, high ion recombination rates can be a significant problem in liquid chambers. In this work, we have investigated properties of a new liquid chamber, called the GLIC-03 (Guarded Liquid Ionization Chamber), including chamber stability, reproducibility, and establishing recombination corrections. The response varied by less than 1% over 10 hours, and was reproducible within 1.5% of the mean over different liquid fills. Recombination corrections were established, and were small for low dose rates and high voltages. The establishment of these characteristics allowed us to compare measurements of the GLIC-03 in a region of charged particle disequilibrium to those made with a diamond detector. Results show the GLIC-03's suitability as a high resolution detector.

This treatise examines the ICRP's new respiratory tract model and its implementation into G scENMOD, a program that facilitates internal dosimetric calculation for the body. The inclusion of the ICRP66 model into G scENMOD improves the radiation dose estimates to the respiratory tract. The accuracy of this implementation was confirmed by validating G scENMOD's output against results published in ICRP66.