For patients with thoracic and abdominal lesions, respiration-induced internal organ motion and deformations during radiation therapy are limiting factors for the administration of high radiation dose. In order to escalate the dose to the tumor and reduce the treatment margins, the tumor movement during treatment must be minimized. In our approach we have established a largely automated deep-breath-hold technique for treating lung cancer patients. We have used a Linear Position Transducer to monitor tumor movement through changes in the patient's abdominal cross-sectional area. The technique aims to reduce the amount of healthy lung tissue in high-dose regions. Normal tissue can be spared as a result of two distinct features of this method: deep inspiration, which reduces the lung density, and breath hold, which immobilizes the tumor. Due to reduced tumor motion, the Planning Target Volume margins can be tightened and a higher dose of radiation can be delivered to the tumor with the same risk of normal tissue complications.

The Enhanced Dynamic Wedge (EDW) is an intensity modulated radiation therapy technique found in certain linear accelerators used in radiation oncology. Invoking the EDW causes one collimating jaw to move across the field to produce a predefined wedged dose distribution. Measured dose profiles were compared with plans calculated by a treatment planning system for various fields. EDW fields were examined for various wedge angles currently available for clinical use. EDW fields for non-clinical angles, which are not currently available, were also produced by applying the ratio of tangents method and delivered with the use of the Dynamic Beam Delivery (DBD) utility. The measured dose profiles correspond to the generated treatment plans within acceptable tolerance limits. Wedge factors were measured for EDW angles (clinical and non-clinical) for asymmetric fields. The production of the non-clinical angles serves as a validation of the current methodology used to generate wedge profiles, and extends the capabilities ofthe current EDW system. Lastly, a comprehensive set of quality assurance tests for the EDW were developed and are presented in this thesis.

A clinical linear accelerator and a Cobalt-60 teletherapy unit were used in combination with five cylindrical cavity ionization chambers to explore the possibility of transferring calibration coefficients from a calibrated reference ionization chamber to a field ionization chamber using linac megavoltage x-ray beams. Five cross-calibration setup methods were explored in Solid Water™ and Lucite™ solid phantoms using 6 MV and 18 MV beams. The calibration coefficients obtained using Lucite™ phantoms were shown to have no statistical difference when compared to those obtained using Solid Water™ phantoms. The calibration coefficients obtained with the 6 MV beam were statistically different than those obtained with the 18 MV beam. The robustness of the five transfer methods in a linac beam was investigated and one method was shown to have superior fault tolerance, with respect to the other methods. / Un accélérateur linéaire clinique et une unité de téléthérapie contenant une capsule de Cobalt-60 furent utilisés en association avec cinq chambres d'ionisation à cavité cylindrique afin d'explorer la possibilité de transférer les coefficients d'étalonnage à partir d'une chambre d'ionisation de référence étalonnée à une autre chambre d'ionisation pour des faisceaux de photons de 6 MV et 18 MV. Cinq méthodes de transfert furent étudiées dans des fantômes en Lucite™ et Solid Water™. Les coefficients d'étalonnage obtenus en utilisant des fantômes de Lucite™ ne montraient aucune différence statistique en comparaison avec ceux en Solid Water™. Par contre, les coefficients d'étalonnage obtenus avec les faisceaux de 6 MV différaient statistiquement de ceux obtenus avec les faisceaux 18 MV. La robustesse des cinq méthodes de transfert fut étudiée et une des méthodes a démontré avoir une meilleure tolérance aux erreurs comparativement aux autres méthodes.

Previous studies of modulation transfer function (MTF), noise power spectrum (NPS), and detective quantum efficiency (DQE) of metal-plate/film portal detectors have been performed on limited combinations of front and back metal-plates. We report on these parameters for an extensive set of forty-nine front-back metal-plate combinations. The portal detector consists of a double emulsion RP (Kodak localization therapy) film placed between metal-plates: Al, Cu, brass and Pb of thicknesses varying from 0.30 to 4.80 mm. Radiation sources included a Theratron Co-60 unit, and a Varian Clinac-18 linear accelerator delivering a polyenergetic 10 MV X-ray spectrum. In terms of the absolute efficiency of the detectors, the best DQE is obtained with the detector consisting of a 1.75 mm Cu front plate and a 1.62 mm Al back plate for the Clinac-18, and with the detector consisting of a 0.95 mm Cu front plate and a 0.80 mm Cu or a 1.62 mm Al back plate for the Co-60 gamma ray source.

We focus primarily on the bulk response to external illumination of conservatively scattering thick inhomogeneous media (or simply "clouds") which are exactly or statistically scale invariant; these radiative properties are compared to those of homogeneous media with the same shapes and masses. Also considered are the ensemble-average responses of multifractal distributions of optical thicknesses and the closely related spatially averaged responses obtained within the "independent pixel" approximation to inhomogeneous transfer. In all cases, the nonlinearity of the radiation/density field coupling induces systematic and specific variability effects. Generally speaking, the details of the scattering process and of the boundary shape affect only prefactors whereas "anomalous" scaling exponents are found for extreme forms of internal variability which, moreover, are different for different physical transport models (e.g., kinetic versus diffusion approaches). Finally, detailed numerical computations of radiation flows inside a log-normal multifractal illustrate the basic inhomogeneous transport mechanism of "channeling."

Recently, the field of external beam radiation therapy has seen an increase in the use of non-standard fields; those composed of many small fields or any field where charged particle disequilibrium conditions exist. Since current clinical reference dosimetry protocols do not cover the conditions of these non-standard fields a new formalism for reference dosimetry of small and non-standard fields has recently been proposed, introducing a plan-class specific reference field (pcsr ). This pcsr would represent a specific class of non-standard delivery and provide an intermediate step between the conventional reference field and a clinical delivery. The purpose of this study was to create a pcsr for an ear-nose-throat IMRT delivery as well as to establish a method of determining the pcsr quality correction factor used to convert calibration coefficients in reference conditions to those under pcsr conditions. The method created included the use of ionization chamber and GafChromic film measurements to determine the factor as well as the use of MLC dynalog files to be used with EGSnrc Monte Carlo simulation to calculate the factor. The method was found to work well in determining the pcsr quality correction factor for the full delivery of the pcsr plan but failed under a beam-by-beam examination of the factor. / Récemment, le domaine de la radiothérapie par faisceaux externes a vu une augmentation de l’utilisation de champs non-standards; ceux-ci composés de plusieurs petits champs ou bien de champs où les conditions de déséquilibre des particules chargées existent. Puisque les protocoles actuels de dosimétrie de référence ne couvrent pas les conditions de ces champs non-standards, un nouveau formalisme pour la dosimétrie de référence des petits et non-standards champs a récemment été proposé, introduisant un champ de référence spécifique pour chaque classe (localisation) de plan de traitement. Ce champ de référence spécifique représenterait une classe spécifique de traitement non-standard et fournirait une étape intermédiaire entre le champ de référence et un traitement clinique. Le but de cette étude était de créer un champ de référence spécifique pour un traitement IMRT ORL et d’établir une méthode déterminant le facteur de correction de qualité du champ de référence spécifique. Ce facteur de correction est utilisé pour passer des conditions de référence aux conditions du champ de référence spécifique. La méthode créée incluait l’utilisation d’une chambre d’ionisation, les mesures d’un film Gafchromic ainsi que l’utilisation des fichiers MLC dynalog, utilisés avec les simulations Monte Carlo EGSnrc, pour déterminer le facteur. La méthode était trouvée efficace lorsque le facteur de correction de qualité du champ de référence spécifique était déterminé pour toute la réalisation du plan mais échouait sous l’examen faisceau par faisceau du facteur.

With the development of modern linear accelerators, the dosimetry of complex electron beams technique became a challenge for physicists. Over the past few years, lots of efforts have been done on developing accurate and fast dose algorithms for electrons. Numerous Monte Carlo (MC) models of therapeutic electron beams are presented in the literature. However, beam models built solely with manufacturer specifications of the medical accelerator do not systematically provide acceptable agreements with measurements. Clinically accurate beam models are crucial to MC treatment planning as electron dose calculations found in commercial treatment planning system (TPS) are generally inaccurate for complex geometry or with heterogeneities. Therefore, there is a strong motivation to use highly accurate MC simulations as standard information for commissioning commercial TPS. The current research project consists in developing an improved accurate electron beam model based on detailed information of the linear accelerator and to incorporate it into an in-house TPS: The McGill Monte Carlo Treatment Planning (MMCTP). / Avec le développement d'accélérateurs linéaires, la dosimétrie de techniques complexes de faisceaux d'électrons devient un défi pour les physiciens. Ces dernières années, des efforts considérables ont été faits pour développer un algorithme de calcul de dose précis et rapide pour les faisceaux d'électrons. De nombreux modèles Monte Carlo (MC) pour des faisceaux thérapeutiques d'électrons sont connus dans la littérature. Néanmoins, des modèles de faisceaux construits seulement avec les spécifications des constructeurs d'accélérateurs linéaires ne fournissent pas systématiquement des résultats en concordance avec les mesures. Au niveau clinique, des modèles de faisceaux précis sont d'une importance capitales pour les calculs de dose par Monte Carlo, dans la mesure où les algorithmes de calcul de faisceaux d'électrons présents dans les systèmes de plannification de traitements (SPT) commerciaux sont générallement imprécis dans des cas comportant des geométries complexes ou des hétérogénéités. Par conséquent, il y a une grande motivation à utiliser des simulations de Monte Carlo précis, comme information standard pour la mise en service des SPT commerciaux. La recherche présentée dans ce manuscrit vise à développer un modèle amélioré et précis de faisceaux d'électrons basé sur des informations détaillées d'accélérateurs linéaires et à incorporer ce modèle dans un SPT: le « McGill Monte Carlo Treatment Planning (MMCTP) ».

A new thermoluminescence dosimetry system for environmental applications was tested, which used high sensitivity lithium fluoride (TLD-100H). The energy response of the bare thermoluminescence dosimeters (TLDs) was measured for photon beams with mean energies from 24 keV to 1.1 MeV, and the results were compared with standard lithium fluoride (TLD-100). The energy response was also measured for TLD-100H card-mounted dosimeters encapsulated in Teflon RTM, used as part of the Harshaw Type 8855 environmental dosimeter. The EGSnrc Monte Carlo system was used to calculate the dose to the TLDs in both the bare chip holder and the 8855 dosimeter, in order to calculate the thermoluminescence per unit of absorbed dose to the TLDs. The results were broadly consistent with existing data, with the response of both TLD materials correlating with the ionization density of the photon beams.

The electron pencil beam algorithm particular to the CadPlan treatment planning system (Varian) was evaluated for energies and field sizes spanning the clinical range. Calculated distributions were compared to measured data from an Elekta SL 25 linear accelerator. Excellent accordance was achieved between measured and calculated distributions for all energies and field sizes in cases of a uniform medium with a flat surface, perpendicular beam incidence and the standard source to surface distance (SSD) of 100 cm. CadPlan was also proficient at producing accurate distributions involving varied SSDs within the range typically employed in the clinic, 98--102 cm. However, when beam incidence was non-orthogonal to the surface, calculated distributions differed from measurement by up to 2.5 mm or 7%. The most severe discrepancies were observed for larger gantry angles. Calculated percent depth dose (PDD) curves in cases of simple slab heterogeneities were found to deviate from measured curves by at most 1.5 mm. Discontinuities appeared in all CadPlan-predicted curves, worsening in instances of non-standard conditions. Many errors stemming from such discontinuity problems could be eliminated by interpolating between values of fitting parameters tabulated in CadPlan and by offering more pencil beam and calculation grid sizes. As CadPlan has no means of accounting for scattered radiation specific to the jaws and applicator, it is unable to predict output factors with any accuracy.

Diss. (sammanfattning) Stockholm : Stockholms universitet, 2006. / Härtill 5 uppsatser.