A patient specific image planning system (IPS) was developed that can be used to assist in kV imaging technique selection during localization for radiotherapy. The IPS algorithm performs a divergent ray-trace through a three dimensional computed tomography (CT) data set. Energy-specific attenuation through each voxel of the CT data set is calculated and imaging detector response is integrated into the algorithm to determine the absolute values of pixel intensity and image contrast. Phantom testing demonstrated that image contrast resulting from under exposure, over exposure as well as a contrast plateau can be predicted by use of a prospective image planning algorithm. Phantom data suggest the potential for reducing imaging dose by selecting a high kVp without loss of image contrast. In the clinic, image acquisition parameters can be predicted using the IPS that reduce patient dose without loss of useful image contrast.

Le développement de nouvelles modalités de traitement telles que la RCMI et la radiothérapie stéréotaxique s'accompagne d'une utilisation croissante de champs d'irradiation complexes obtenus par superposition de faisceaux de petite taille ayant de multiples angles d'incidence. Ces nouvelles conditions de traitement sont très différentes des conditions de référence sur lesquelles se basent les protocoles dosimétriques internationaux. Ces travaux de thèse se proposent de réaliser des références dosimétriques pour des champs d'irradiation de dimensions inférieures à 10 x 10 cm², à savoir 4 x 4 et 2 x 2 cm². Il s'agit, dans la pratique, de comparer les coefficients d'étalonnage d'une chambre d'ionisation en termes de dose absorbée dans l'eau, pour les faisceaux de photons de 6 MV (avec et sans cône égalisateur) et de 12 MV de l'accélérateur linéaire médical du LNHB. Les références ont été déterminées à partir d'une mesure par calorimétrie graphite. Pour les mesures en champ 2 x 2 cm², un calorimètre disposant d'un absorbeur de petites dimensions a été construit. Par ailleurs, une chambre d'ionisation adaptée à cette taille de champ a été recherchée. Nous avons montré que, pour les faisceaux étudiés, le coefficient d'étalonnage de la chambre d'ionisation de référence est indépendant de la dimension du champ d'irradiation entre 10 x 10 et 2 x 2 cm², aux incertitudes près (environ 0,4 % à un écart-type).

La tomographie par émission de positrons est une technique d'imagerie nucléaire utilisant des noyaux radioactifs. Elle est utilisée dans le domaine clinique et préclinique. Cette dernière nécessite l'utilisation de petits animaux, comme la souris. Comme en imagerie clinique, l'objectif est d'obtenir le meilleur signal avec la meilleure précision spatiale possible. Cependant, un rapport d'échelle homme-souris suggère une résolution inférieure à 1 mm3. Un imageur conventionnel est constitué de modules de détection entourant le patient, orientés radialement. Cette approche lie efficacité et résolution spatiale. Ce travail concerne l'étude de la géométrie axiale. Les éléments de détection sont ici orientés parallèlement à l'objet. Ceci limite la corrélation entre efficacité de détection et résolution spatiale, et ainsi permet d'obtenir une haute résolution et haute sensibilité. La simulation de prototypes a permis d'envisager une résolution spatiale moyenne inférieure au millimètre et une efficacité de 15 ou 40% selon l'extension axiale. Ces résultats permettent de présager de bonnes perspectives en imageries préclinique et clinique.

This work establishes a framework for Monte Carlo simulations of complex, modulated electron fields produced by Varian's TrueBeam medical linear accelerator for investigations into modulated electron radiation therapy (MERT) and combined modulated photon and electron radiation therapy (MPERT). Both MERT and MPERT have shown potential for reduced low dose to normal tissue without compromising target coverage in the external beam radiation therapy of some breast, chest wall, head and neck, and scalp cancers. This reduction in low dose could translate into the reduction of immediate radiation side effects as well as long term morbidities and incidence of secondary cancers. Monte Carlo dose calculations are widely accepted as the gold standard for complex radiation therapy dose modelling, and are used almost exclusively for modelling the complex electron fields involved in MERT and MPERT. The introduction of Varian's newest linear accelerator, the TrueBeam, necessitated the development of new Monte Carlo models in order to further research into the potential role of MERT and MPERT in radiation therapy. This was complicated by the fact that the field independent internal schematics of TrueBeam were kept proprietary, unlike in previous generations of Varian accelerators. Two approaches are presented for performing Monte Carlo simulations of complex electron fields produced by TrueBeam. In the first approach, the dosimetric characteristics of electron fields produced by the TrueBeam were first compared with those produced by an older Varian accelerator, the Clinac 21EX. Differences in depth and profile characteristics of fields produced by the TrueBeam and those produced by the Clianc 21EX were found to be within 3%/3 mm. Given this information, complete accelerator models of the Clinac 21EX, based on its known internal geometry, were then successfully modified in order to simulate 12 and 20 MeV electron fields produced by the TrueBeam to within 2%/2 mm of measured depth and profile curves and to within 3.7% of measured relative output. While the 6 MeV TrueBeam model agreed with measured depth and profile data to within 3%/3 mm, the modified Clinac 21EX model was unable to reproduce trends in relative output as a function of field size with acceptable accuracy. The second approach to modelling TrueBeam electron fields used phase-space source files provided by Varian that were scored below the field-independent portions of the accelerator head geometry. These phase-spaces were first validated for use in MERT and MPERT applications, in which simulations using the phase-space source files were shown to model depth dose curves that agreed with measurement within 2%/2 mm and profile curves that agreed with measurement within 3%/3 mm. Simulated changes in output as a function of field size fell within 2.7%, for the most part. In order to inform the positioning of jaws in MLC-shaped electron field delivery, the change in output as a function of jaw position for fixed MLC-apertures was investigated using the phase-space source files. In order to achieve maximum output and minimize treatment time, a jaw setting between 5 and 10 cm beyond the MLC- field setting is recommended at 6 MeV, while 5 cm or closer is recommended for 12 and 20 MeV with the caveat that output is most sensitive to jaw position when the jaws are very close to the MLC-field periphery. Additionally, output was found to be highly sensitive to jaw model. A change in divergence of the jaw faces from a point on the source plane to a 3x3 mm^2 square in the source plane changed the shape of the output curve dramatically. Finally, electron backscatter from the jaws into the monitor ionization chamber of the TrueBeam was measured and simulated to enable accurate absolute dose calculations. Two approaches were presented for measuring backscatter into the monitor ionization chamber without specialized electronics by turning o the dose and pulse forming network servos. Next, a technique was applied for simulating backscatter factors for the TrueBeam phase-space source models without the exact specifications of the monitor ionization chamber. By using measured backscatter factors, the forward dose component in a virtual chamber was determined and then used to calculate backscatter factors for arbitrary fields to within 0.21%. Backscatter from the jaws was found to contribute up to 2.6% of the overall monitor chamber signal. The measurement techniques employed were not sensitive enough to quantify backscatter from the MLC, however, Monte Carlo simulations predicted this contribution to be 0.3%, at most, verifying that this component can be neglected. / Graduate / 0756 / lloyd.samantha@gmail.com

A fundamental challenge in the treatment planning process of multi-fractional external-beam radiation therapy (EBRT) is the tradeoff between tumor control and normal tissue sparing in the presence of geometric uncertainties (GUs). To accommodate GUs, the conventional way is to use an empirical planning treatment volume (PTV) margin on the treatment target. However, it is difficult to determine a near-optimal PTV margin to ensure specified target coverage with as much normal tissue protection as achievable. Coverage optimized planning (COP) avoids this problem by optimizing dose in possible virtual treatment courses with GU models directly incorporated. A near-optimal dosimetric margin generated by COP was reported to savvily accommodate setup errors of target and normal tissues for prostate cancer treatment. This work further develops COP to account for (1) deformable organ motion and (2) delineation uncertainties for high-risk prostate cancer patients. The clinical value of COP is investigated by comparing with two margin-based planning techniques: (i) optimized margin (OM) technique that iteratively modifies PTV margins according to the evaluated target coverage probability and (ii) fixed margin (FM) technique that uses empirically selected constant PTV margins. Without patient-specific coverage probability estimation, FM plans are always less immune to the degraded effect of the modeled GUs than the COP plans or the OM plans. Empirical PTV margins face more risks of undesirable target coverage probability and/or excessive dose to surrounding OAR. The value of COP relative to OM varies with different GUs. As implemented for deformable organ motions, COP has limited clinical benefit. Due to optimization tradeoffs, COP often results in target coverage probability below the prescribed value while OM achieves better target coverage with comparable normal tissue dose. For delineation uncertainties, the clinical value of COP is potentially significant. Compared to OM, COP successfully maintains acceptable target coverage probability by exploiting the slack of normal tissue dose in low dose regions and maximally limiting high dose to normal tissue within tolerance.

This dissertation describes the application of two principled variance reduction strategies to increase the efficiency for two applications within medical physics. The first, called correlated Monte Carlo (CMC) applies to patient-specific, permanent-seed brachytherapy (PSB) dose calculations. The second, called adjoint-biased forward Monte Carlo (ABFMC), is used to compute cone-beam computed tomography (CBCT) scatter projections. CMC was applied for two PSB cases: a clinical post-implant prostate, and a breast with a simulated lumpectomy cavity. CMC computes the dose difference between the highly correlated dose computing homogeneous and heterogeneous geometries. The particle transport in the heterogeneous geometry assumed a purely homogeneous environment, and altered particle weights accounted for bias. Average gains of 37 to 60 are reported from using CMC, relative to un-correlated Monte Carlo (UMC) calculations, for the prostate and breast CTV’s, respectively. To further increase the efficiency up to 1500 fold above UMC, an approximation called interpolated correlated Monte Carlo (ICMC) was applied. ICMC computes using CMC on a low-resolution (LR) spatial grid followed by interpolation to a high-resolution (HR) voxel grid followed. The interpolated, HR is then summed with a HR, pre-computed, homogeneous dose map. ICMC computes an approximate, but accurate, HR heterogeneous dose distribution from LR MC calculations achieving an average 2% standard deviation within the prostate and breast CTV’s in 1.1 sec and 0.39 sec, respectively. Accuracy for 80% of the voxels using ICMC is within 3% for anatomically realistic geometries. Second, for CBCT scatter projections, ABFMC was implemented via weight windowing using a solution to the adjoint Boltzmann transport equation computed either via the discrete ordinates method (DOM), or a MC implemented forward-adjoint importance generator (FAIG). ABFMC, implemented via DOM or FAIG, was tested for a single elliptical water cylinder using a primary point source (PPS) and a phase-space source (PSS). The best gains were found by using the PSS yielding average efficiency gains of 250 relative to non-weight windowed MC utilizing the PPS. Furthermore, computing 360 projections on a 40 by 30 pixel grid requires only 48 min on a single CPU core allowing clinical use via parallel processing techniques.

To improve relatively poor outcomes for locally-advanced lung cancer patients, many current efforts are dedicated to minimizing uncertainties in radiotherapy. This enables the isotoxic delivery of escalated tumor doses, leading to better local tumor control. The current dissertation specifically addresses inter-fractional uncertainties resulting from patient setup variability. An automatic block-matching registration (BMR) algorithm is implemented and evaluated for the purpose of directly localizing advanced-stage lung tumors during image-guided radiation therapy. In this algorithm, small image sub-volumes, termed “blocks”, are automatically identified on the tumor surface in an initial planning computed tomography (CT) image. Each block is independently and automatically registered to daily images acquired immediately prior to each treatment fraction. To improve the accuracy and robustness of BMR, this algorithm incorporates multi-resolution pyramid registration, regularization with a median filter, and a new multiple-candidate-registrations technique. The result of block-matching is a sparse displacement vector field that models local tissue deformations near the tumor surface. The distribution of displacement vectors is aggregated to obtain the final tumor registration, corresponding to the treatment couch shift for patient setup correction. Compared to existing rigid and deformable registration algorithms, the final BMR algorithm significantly improves the overlap between target volumes from the planning CT and registered daily images. Furthermore, BMR results in the smallest treatment margins for the given study population. However, despite these improvements, large residual target localization errors were noted, indicating that purely rigid couch shifts cannot correct for all sources of inter-fractional variability. Further reductions in treatment uncertainties may require the combination of high-quality target localization and adaptive radiotherapy.

The purpose of this work is to present the development and validation of a novel method for reconstructing time-dependent, or 4D, cone-beam CT (4DCBCT) images. 4DCBCT can have a variety of applications in the radiotherapy of moving targets, such as lung tumors, including treatment planning, dose verification, and real time treatment adaptation. However, in its current incarnation it suffers from poor reconstruction quality and limited temporal resolution that may restrict its efficacy. Our algorithm remedies these issues by deforming a previously acquired high quality reference fan-beam CT (FBCT) to match the projection data in the 4DCBCT data-set, essentially creating a 3D animation of the moving patient anatomy. This approach combines the high image quality of the FBCT with the fine temporal resolution of the raw 4DCBCT projection data-set. Deformation of the reference CT is accomplished via a patient specific motion model. The motion model is constrained spatially using eigenvectors generated by a principal component analysis (PCA) of patient motion data, and is regularized in time using parametric functions of a patient breathing surrogate recorded simultaneously with 4DCBCT acquisition. The parametric motion model is constrained using forward iterative projection matching (FIPM), a scheme which iteratively alters model parameters until digitally reconstructed radiographs (DRRs) cast through the deforming CT optimally match the projections in the raw 4DCBCT data-set. We term our method FIPM-PCA 4DCBCT. In developing our algorithm we proceed through three stages of development. In the first, we establish the mathematical groundwork for the algorithm and perform proof of concept testing on simulated data. In the second, we tune the algorithm for real world use; specifically we improve our DRR algorithm to achieve maximal realism by incorporating physical principles of image formation combined with empirical measurements of system properties. In the third stage we test our algorithm on actual patient data and evaluate its performance against gold standard and ground truth data-sets. In this phase we use our method to track the motion of an implanted fiducial marker and observe agreement with our gold standard data that is typically within a millimeter.

Purpose: This dissertation focuses on establishment of pre-clinical methods facilitating the use of PET imaging for selective sub-volume dose escalation. Specifically the problems addressed are 1.) The difficulties associated with comparing multiple PET images, 2.) The need for further validation of novel PET tracers before their implementation in dose escalation schema and 3.) The lack of concrete pre-clinical data supporting the use of PET images for guidance of selective sub-volume dose escalations. Methods and materials: In order to compare multiple PET images the confounding effects of mispositioning and anatomical change between imaging sessions needed to be alleviated. To mitigate the effects of these sources of error, deformable image registration was employed. A deformable registration algorithm was selected and the registration error was evaluated via the introduction of external fiducials to the tumor. Once a method for image registration was established, a procedure for validating the use of novel PET tracers with FDG was developed. Nude mice were used to perform in-vivo comparisons of the spatial distributions of two PET tracers, FDG and FLT. The spatial distributions were also compared across two separate tumor lines to determine the effects of tumor morphology on spatial distribution. Finally, the research establishes a method for acquiring pre-clinical data supporting the use of PET for image-guidance in selective dose escalation. Nude mice were imaged using only FDG PET/CT and the resulting images were used to plan PET-guided dose escalations to a 5 mm sub-volume within the tumor that contained the highest PET tracer uptake. These plans were then delivered using the Small Animal Radiation Research Platform (SARRP) and the efficacy of the PET-guided plans was observed. Results and Conclusions: The analysis of deformable registration algorithms revealed that the BRAINSFit B-spline deformable registration algorithm available in SLICER3D was capable of registering small animal PET/CT data sets in less than 5 minutes with an average registration error of .3 mm. The methods used in chapter 3 allowed for the comparison of the spatial distributions of multiple PET tracers imaged at different times. A comparison of FDG and FLT showed that both are positively correlated but that tumor morphology does significantly affect the correlation between the two tracers. An overlap analysis of the high intensity PET regions of FDG and FLT showed that FLT offers additional spatial information to that seen with FDG. In chapter 4 the SARRP allowed for the delivery of planned PET-guided selective dose escalations to a pre-clinical tumor model. This will facilitate future research validating the use of PET for clinical selective dose escalation.

L’étude des effets biologiques des radiations ionisantes à l’échelle de la cellule individuelle et en particulier sur l’ADN du noyau cellulaire reste un enjeu majeur de la radiobiologie actuelle. L’objectif principal des recherches actuelles est de déterminer quels peuvent être les effets biologiques délétères des radiations ionisantes pour la santé humaine, en particulier dans le domaine des faibles doses de radiation. Afin d’étudier précisément la réponse des cellules aux radiations ionisantes, de nombreuses études expérimentales des effets des radiations ionisantes sur les cellules, tissus et organismes biologiques aux basses énergies ont accumulées de grandes quantités de données de qualité sur la réponse de cellules aux radiations. Il existe également de nombreux modèles semi-empiriques de survie cellulaire qui incorporent des paramètres biologiques et physiques. En parallèle, des stochastiques basées sur la technique « Monte Carlo » pour modéliser les processus élémentaires en physique, chimie et biologie sont en cours de développement. L’outil Geant4 développé dès 1993 (CERN et KEK) en utilisant des techniques informatiques de dernière génération (C++) permet à l’utilisateur de construire une simulation complète grâce à de nombreuses fonctionnalités : formes géométriques, matériaux, particules élémentaires, processus physiques électromagnétiques et hadroniques, visualisation, analyse de données, interactivité et extensibilité… Cependant, Geant4 présente un certain nombre de limitations pour la simulation des effets biologiques des radiations ionisants à l’échelle subcellulaire : les modèles standard ne prennent pas compte le technique « pas-à-pas », les modèles physique sont limités à basse énergie, il n’a pas des descriptions des cibles moléculaires et Geant4 n’est pas capable de simuler les étapes physico-chimique et chimique nécessaire pour déterminer l’oxydation des bases et les éventuelles cassures d’ADN.Dans ce contexte, le projet Geant4-DNA propose d’étendre Geant4 afin de modéliser les interactions des radiations ionisantes à l’échelle de la cellule biologique et la molécule d’ADN et aux basses énergies. Au cours du travail de thèse, j’ai tout d’abord validé les modèles physiques en comparant les résultats de simulation à une grande collection de données expérimentales disponibles dans la littérature. L’accord entre les valeurs de sections efficaces totales et différentielles et les mesures expérimentales a été quantifié à l’aide du test statistique Kolmogorov-Smirnov. J’ai par la suite amélioré les classes des processus de diffusion élastique des électrons et travailler sur les calculs théoriques du modèle de diffusion élastique des protons et des alphas dans l’eau liquide auparavant inexistant dans Geant4-DNA. J’ai effectué une combinaison des processus multi-échelles des modèles de Geant4-DNA (à l’échelle microscopique) avec les modèles électromagnétiques disponibles dans l’outil Geant4 (les processus d’interaction des photons et autres modèles de Geant4). A la fin de mon travail, j’ai participé à l’estimation des performances de Geant4-DNA pour la dosimétrie dans des géométries de petite taille (jusqu’à l’échelle du nanomètre) dans l’eau liquide à l’aide des distributions « Dose Point Kernel ». J’ai ensuite calculé les fréquences de dépôts d’énergie dans des petits cylindres de dimensions nanométriques correspondant à des cibles biologiques et des modèles de noyau cellulaire humain simplifié pour l’estimation des cassures directes simple et double. Mon travail de thèse a fournit les premiers résultats de Geant4-DNA pour la prédiction de cassure de brin d’ADN combinant physique et géométries à l’échelle de l’ADN. Enfin, nous avons développé des classes de processus et modèles basés sur l’approche CTMC-COB (Classical Trajetory Monte Carlo avec critère d’Over Barrier) spécifique aux bases de la molécule d’ADN et à l’eau liquide. / A large experimental and modeling activity is currently taking place, aimed at better understanding the biological effects of ionizing radiation at the molecular scales. Considerable amounts of experimental data have been accumulated over the past decades in order to measure quantities such as macroscopic cellular survival curves and DNA strand damages after irradiation. In parallel, computer codes have been proposed to use a stochastic approach based on Monte Carlo technique to model physical interaction in the irradiated medium. The Geant4 toolkit uses the object-oriented technology (C++) to describing particle-matter interactions, such as bio-medical physics and space physics, from sub-micrometer cells up to planetary scales. Geant4-DNA project is included in the Geant4 toolkit and benefits from the easy accessibility of the Geant4 code for the development of a computing platform allowing estimation effects of ionizing radiations. In my thesis, firstly, I have contributed in the project the validation of various models with the experimental data collections extracted from the recent literature. A good agreement between total and differential cross section values corresponding to each available Geant4-DNA model and experimental data is validated by Kolmogorov-Smirnov testing. Secondly, I have improved elastic scattering process and working on the calculation of the DDCS for proton elastic scattering in water in the Geant4-DNA. In addition, I have combined Geant4 electromagnetic processes with the Geant4-DNA. This combination brought additional Geant4 simulation capabilities in complement of the possibility to combine Geant4-DNA models with other Geant4 electromagnetic models at different sizes and energy scales in a single simulation application. Finally, we have presented the usage of Geant4-DNA physics processes in nanometer-size targets fully implemented in a single Geant4 application. The frequencies of the deposited energy and number of direct DNA single strand break and double strand break in the simplified nucleus model are compared with other codes results and with a collection of experimental data on direct DNA dimensions on plasmid DNA. Furthermore I have implemented in Geant4-DNA theoretical cross sections of physics processes based on a Classical Trajectory Monte Carlo (CTMC) approach for modeling the detailed transport of protons and neutral hydrogen atoms in liquid water and in DNA nucleobases.