Dissertations / Theses on the topic 'Spectral Photon Counting Computed Tomography'

Computed tomography (CT) is a widely used medical imaging modality. By rotating an x-ray tube and an x-ray detector around the patient, a CT scanner is able to measure the x-ray transmission from all directions and form an image of the patient’s interior. CT scanners in clinical use today all use energy-integrating detectors, which measure the total incident energy for each measurement interval. A photon-counting detector, on the other hand, counts the number of incoming photons and can in addition measure the energy of each photon by comparing it to a number of energy thresholds. Using photon- counting detectors in computed tomography could lead to improved signal-to-noise ratio, higher spatial resolution and improved spectral imaging which allows better visualization of contrast agents and more reliable quantitative measurements. In this Thesis, the feasibility of using a photon-counting silicon-strip detector for CT is investigated. In the first part of the Thesis, the necessary performance requirements on such a detector is investigated in two different areas: the detector element homogeneity and the capability of handling high photon fluence rates. A metric of inhomogeneity is proposed and used in a simulation study to evaluate different inhomogeneity compensation methods. Also, the photon fluence rate incident on the detector in a scanner in clinical use today is investigated for different patient sizes through dose rate measurements together with simulations of transmission through patient im- ages. In the second part, a prototype detector module is used to demonstrate new applications enabled by the energy resolution of the detector. The ability to generate material-specific images of contrast agents with iodine and gadolinium is demonstrated. Furthermore, it is shown theoretically and ex- perimentally that interfaces in the image can be visualized by imaging the so-called nonlinear partial volume effect. The results suggest that the studied silicon-strip detector is a promising candidate for photon-counting CT.

X-ray computed tomography (CT) has since its introduction in the early 1970s become one of the most important tools used for medical imaging. In CT, a large number of x-ray attenuation measurements are combined and reconstructed to form a three-dimensional image of the targeted area. In the recent years, a new type of detector called photon counting detector (PCD) has attracted considerable interest. This new type of detector acquires spectral information is associated with several benefits and has shown to be very valuable.  Furthermore, the use of deep learning to reconstruct images produced by CT has attracted significant attention in the last couple of years. However, the best way of incorporating deep learning into the reconstruction chain into the reconstruction chain is still incompletely understood. Additionally, the use of deep learning has mainly been investigated for the case of conventional CT and not for CT performed with PCDs. It these two points that this work aims to address.  Multiple deep learning architectures were implemented and evaluated on material images acquired by simulating a PCD. The deep-learning part of the reconstruction took the form of image-domain denoising after the material images had been obtained from the material sinograms through filtered back projection. Then, a comparison between the different deep learning architectures was made to find out which architecture is the most suited for denoising images produced by PCDs in the image domain.

Spectral computed tomography with energy-resolving detectors has a potential to improve the detectability of images and correspondingly reduce the radiation dose to patients by extracting and properly using the energy information in the broad x-ray spectrum. A silicon photon-counting detector has been developed for spectral CT and it has successfully solved the problem of high photon flux in clinical CT applications by adopting the segmented detector structure and operating the detector in edge-on geometry. The detector was evaluated by both the simulation and measurements. The effects of energy loss and charge sharing on the energy response of this segmented silicon strip detector with different pixel sizes were investigated by Monte Carlo simulation and a comparison to pixelated CdTe detectors is presented. The validity of spherical approximations of initial charge cloud shape in silicon detectors was evaluated and a more accurate statistical model has been proposed. A photon-counting energy-resolving application specific integrated circuit (ASIC) developed for spectral CT was characterized extensively by electrical pulses, pulsed laser and real x-ray photons from both the synchrotron and an x-ray tube. It has been demonstrated that the ASIC performs as designed. A noise level of 1.09 keV RMS has been measured and a threshold dispersion of 0.89 keV RMS has been determined. The count rate performance of the ASIC in terms of count loss and energy resolution was evaluated by real x-rays and promising results have been obtained. The segmented silicon strip detector was evaluated using synchrotron radiation. An energy resolution of 16.1% has been determined with 22 keV photons in the lowest flux limit, which deteriorates to 21.5% at an input count rate of 100 Mcps mm−2. The fraction of charge shared events has been estimated and found to be 11.1% for 22 keV and 15.3% for 30 keV. A lower fraction of charge shared events and an improved energy resolution can be expected by applying a higher bias voltage to the detector.

QC 20121123

During my time as a PhD student at the Physics of Medical Imaging group at KTH, I have taken part in the work of developing a photon counting spectrally resolved silicon detector for clinical computed tomography. This work has largely motivated the direction of my research, and is the main reason for my focus on certain issues. Early in the work, a need to quantify and optimize the performance of a spectrally resolved detector was identified. A large part of my work have thus consisted of reviewing conventional methods used for performance quantification and optimization in computed tomography, and identifying which are best suited for the characterization of a spectrally resolved system. In addition, my work has included comparisons of conventional systems with the detector we are developing. The collected result after a little more than four years of work are four publications and three conference papers. This compilation thesis consists of five introductory chapters and my four publications. The introductory chapters are not self-contained in the sense that the theory and results from all my published work are included. Rather, they are written with the purpose of being a context in which the papers should be read. The first two chapters treat the general purpose of the introductory chapters, and the theory of computed tomography including the distinction between conventional, non-spectral, computed tomography, and different practical implementations of spectral computed tomography. The second chapter consists of a review of the conventional methods developed for quantification and optimization of image quality in terms of detectability and signal-to-noise ratio, part of which are included in my published work. In addition, the theory on which the method of material basis decomposition is based on is presented, together with a condensed version of the results from my work on the comparison of two systems with fundamentally different practical solutions for material quantification. In the fourth chapter, previously unpublished measurements on the photon counting spectrally resolved detector we are developing are presented, and compared to Monte Carlo simulations. In the fifth and final chapter, a summary of the appended publications is included.

QC 20150303

Desenvolvimento de um sistema de tomografia computorizada de contagem de fotão único usando MPGDs
The development of computed tomography systems with energy resolving detectors is a current challenge in medical physics and biomedical engineering. A computed tomography system of this kind allows getting complementary informations relatively to conventional systems, that can help the medical diagnosis, being of great interest in medicine. The work described in this thesis is related to the development of a computed tomography system using micropattern gaseous detectors, which allow storing, simultaneously, information about the interaction position and the energy of each single photon that interacts with the detector. This kind of detectors has other advantages concerning the cost and characteristics of operation when compared with solid state detectors. Tomographic acquisitions were performed using a MicroHole & Strip Plate based detector, which allowed reconstructing cross-sectional images using energy windows, applying the energy weighting technique and performing multi-slice and tri-dimensional reconstructions. The contrast-to-noise ratio was improved by 31% by applying the energy weighting technique, comparing with the corresponding image obtained with the current medical systems. A prototype of a computed tomography with flexibility to change the detector was developed, making it possible to apply different detectors based on Thick-COBRA. Several images acquired with these detectors are presented and demonstrate their applicability in X-ray imaging. When operating in NeCH4, the detector allowed a charge gain of 8 104, an energy resolution of 20% (full width at half maximum at 8 keV), a count rate of 1 106 Hz/mm2, a very stable operation (gain fluctuations below 5%) and a spacial resolution of 1.2 mm for an energy photon of 3.6 keV. Operating the detector in pure Kr allowed increasing the detection efficiency and achieving a charge gain of 2 104, an energy resolution of 32% (full width at half maximum at 22 keV), a count rate of 1 105 Hz/mm2, very stable operation and a spatial resolution of 500 m. The software already existing in the group was improved and tools to correct geometric misalignments of the system were also developed. The reconstructions obtained after geometrical correction are free of artefacts due to the referred misalignments.
O desenvolvimento de sistemas de tomografia computorizada que incorporem detetores com resolução em energia é um desafio atual em física médica e engenharia biomédica. Um sistema de tomografia computorizada espetral permite obter informações complementares comparativamente a um sistema convencional, que podem auxiliar no diagnóstico médico, sendo por isso de grande interesse em medicina. O trabalho exposto nesta tese prende-se com o desenvolvimento de um sistema de tomografia usando detetores gasosos microestruturados que permitem, simultaneamente, ter informação da posição de interacção e da energia de cada fotão que interage com o detetor. Este tipo de detetores possui ainda outras vantagens relativamente a custo ou características de funcionamento quando comparados com detetores de estado sólido. Foram realizadas aquisições tomográficas usando um detetor baseado numa MicroHole & Srip Plate que permitiu reconstruir imagens utilizando diferentes gamas de energia, aplicar técnicas de ponderação em energia e fazer pela primeira vez reconstrução multi-corte e obter imagens tri-dimensionais. Aplicando a técnica de ponderação em energia foi possível melhorar a relação contraste-ruído em 31% comparativamente à imagem correspondente aquela obtida nos actuais sistemas médicos. Posteriormente, foi desenvolvido um protótipo de um sistema de tomografia computorizada com flexibilidade para alterar o detetor, tornando possível utilizar vários detetores baseados na microestrutura Thick-COBRA. São apresentadas várias imagens adquiridas com estes detetores que evidenciam a sua aplicabilidade em imagiologia por raio-X. A operar no meio gasoso NeCH4 o detetor permitiu um ganho de 8 104, uma resolução em energia de 20% (largura a meia altura a 8 keV), uma taxa de contagem de 1 106 Hz/mm2, um funcionamento muito estável (variações de ganho inferiores a 5%) e uma resolução espacial de 1.2 mm para fotões de 3.6 keV. A operar em Kr puro foi possível aumentar a eficiência de deteção e alcançar um ganho de 2 104, uma resolução em energia de 32% (largura a meia altura a 22 keV), uma taxa de contagem de 1 105 Hz/mm2, um funcionamento também bastante estável e uma resolução espacial de 500 m. O software já existente no grupo para reconstrução de imagem foi melhorado e foram ainda desenvolvidas ferramentas para correcção de desalinhamentos geométricos do sistema. As reconstruções obtidas após correção geométrica surgem livres de artefactos originados pelos referidos desalinhamentos.

Multibin photon-counting x-ray detectors are promising candidates to be applied in next generation computed tomography (CT), whereby energy information from a broad x-ray spectrum can be extracted and properly used for improving image quality and correspondingly reducing radiation dose. A silicon-strip detector has been developed for spectral CT, which operates in photon-counting mode and allows pulse-height discrimination with 8 adjustable energy bins. Critical characteristics, energy resolution and count-rate performance, of the detector are evaluated. An absolute energy resolution (E) from 1.5 keV to 1.9 keV with increasing x-ray energy from 40 keV to 120 keV is found. Pulse pileup degrades the energy resolution by 0.4 keV when increasing the input count rate from zero to 100 Mcps mm−2, while charge sharing shows negligible effect. A near linear relationship between the input and output count rates is observed up to 90 Mcps mm−2 in a clinical CT environment. In addition, no saturation effect appears for the maximally achieved photon flux of 485 Mphotons s−1 mm−2 with a count rate loss of 30%. The detector is energy calibrated in terms of gain and offset with the aid of monoenergetic x rays. The gain variation among channels is below 4%, whereas the variation of offsets is on the order of 1 keV. In order to do the energy calibration in a routinely available way, a method that makes use of the broad x-ray spectrum instead of using monoenergetic x rays is proposed. It is based on a regression analysis that adjusts a modelled spectrum of deposited energies to a measured pulse-height spectrum. Application of this method shows high potential to be applied in an existing CT scanner with an uncertainty of a calibrated threshold between 0.1 and 0.2 keV. The energy-calibration method is further used in the development of a spectral response model of the detector. This model is used to accurately bin-wise predict the response of each detector channel, which is validated by two application examples. First, the model is used in combination with the inhomogeneity compensation method to eliminate ring artefacts in CT images. Second, the spectral response model is used as the basis of the maximum likelihood approach for projection-based material decomposition. The reconstructed basis images show a good separation between the calcium-like material and the contrast agents, iodine and gadolinium. Additionally, the contrast agent concentrations are reconstructed with more than 94% accuracy.

QC 20160908

Le scanner CT spectral à comptage de photons (sCT) est apparu récemment comme une nouvelle technique d'imagerie présentant des avantages fondamentaux par rapport au scanner CT classique et au scanner CT à double énergie. Cependant, en raison du nombre réduit de photons dans chaque bande d'énergie du scanner sCT et des artéfacts divers, la reconstruction des images devient particulièrement difficile. Cette thèse se concentre sur la reconstruction d'images multi-énergie en sCT. Tout d'abord, nous proposons d'étudier la capacité du scanner sCT à réaliser simultanément une imagerie anatomique (aCT) et fonctionnelle (fCT) en une seule acquisition par reconstruction et décomposition des matériaux. La fonction aCT du scanner sCT est étudiée dans la même configuration que celle du scanner CT classique, et la fonction fCT du scanner sCT est étudiée en appliquant des algorithmes de décomposition de matériaux aux mêmes données multi-énergie. Ensuite, comme le bruit est un problème particulièrement aigu en raison du nombre largement réduit de photons dans chaque bande d'énergie du scanner sCT, nous introduisons un mécanisme de débruitage dans la reconstruction de l'image pour effectuer simultanément un débruitage et une reconstruction. Enfin, pour améliorer la reconstruction de l'image, nous proposons de reconstruire l'image à une bande d'énergie donnée en exploitant les informations dans toutes les autres bandes d'énergie. La stratégie clé de cette approche consiste à regrouper les pixels similaires issus de la reconstruction de toutes les bandes d'énergie en une seule classe, à les ajuster dans la même classe, à projeter les résultats de l'ajustement dans chaque bande d'énergie, et à débruiter les informations projetées. Elle est utilisée à la fois comme une opération post-débruitage pour démontrer son efficacité et comme un terme de régularisation ou un terme de régularisations combinées pour la réalisation simultanée du débruitage et de la reconstruction. Toutes les méthodes ci-dessus sont évaluées sur des données de simulation et des données réelles provenant d'un scanner sCT préclinique
Spectral photon-counting CT (sCT) appeared recently as a new imaging technique presenting fundamental advantages with respect to conventional CT and duel-energy CT. However, due to the reduced number of photons in each energy bin of sCT and various artifacts, image reconstruction becomes particularly difficult. This thesis focuses on the reconstruction of multi-energy images in sCT. First, we propose to consider the ability of sCT to achieve simultaneously both anatomical (aCT) and functional imaging (fCT) in one single acquisition through reconstruction and material decomposition. aCT function of sCT is studied under the same configuration as that of conventional CT, and fCT function of sCT is investigated by applying material decomposition algorithms to the same acquired multi-energy data. Then, since noise is a particularly acute problem due to the largely reduced number of photons in each energy bin of sCT, we introduce denoising mechanism in the image reconstruction to perform simultaneous reconstruction and denoising. Finally, to improve image reconstruction, we propose to reconstruct the image at a given energy bin by exploiting information in all other energy bins. The key strategy in such approach consists of grouping the similar pixels from the reconstruction of all the energy bins into the same class, fitting within each class, mapping the fitting results into each energy bin, and denoising with the mapped information. It is used both as a post-denoising operation to demonstrate its effectiveness and as a regularization term or a combined regularization term for simultaneous reconstruction and denoising. All the above methods are evaluated on both simulation and real data from a pre-clinical sCT system

Le rayonnement diffusé est une cause majeure de biais, de baisse de contraste et d'artéfacts en tomographie par rayons x. De nombreuses méthode de correction ont été proposées pour la tomographie conventionnelle (utilisant des détecteurs à intégration en énergie) mais le sujet reste ouvert dans le cadre de l'imagerie spectrale, une nouvelle modalité d'imagerie basée sur l'utilisation de détecteurs à comptage des photons résolus en énergie. L'objectif principal de ce travail de thèse a été l'étude de techniques de correction du diffusé adaptées à l'imagerie spectrale. La solution choisie améliore une méthode de correction développée pour la tomographie conventionnelle qui utilise un masque modulateur de primaire semi-transparent. L'atténuation du masque est d'abord compensée avec une matrice de correction qui bénéficie de l'information spectrale. Les autres contributions sont un modèle de diffusé basé sur des B-splines et qui permet de représenter précisément les images de diffusé à l'aide d'un très faible nombre de paramètres, et une fonction de coût qui prend en compte les structures du masque et de l'objet. Les performances de la matrice de correction, du modèle de diffusé, et de l'ensemble de la méthode de correction proposée ont été évaluées sur des données simulées, en considérant des détecteurs à comptage de photons avec différents nombres de canaux d'énergie. De plus, des acquisitions ont été réalisées sur un système tomographique à faisCEAux en éventail parallèles comportant un détecteur commercial résolu en énergie. La méthode a été validée avec succès dans les cas de deux fantômes dédiés à des mesures de qualité d'images médicales, avec une amélioration remarquable
Scattered radiation is a major cause of bias, loss of contrast and artifacts in x-ray computed tomography (CT). Many correction methods have been proposed for conventional CT (using energy-integrating detectors) but it is still an open research topic in the field of spectral CT, a novel imaging technique based on the use of energy-selective photon counting detectors. The main objective of the present thesis was to investigate scatter correction techniques adapted to spectral CT. The chosen solution refines a scatter correction method developed for integration-mode CT which uses a semi-transparent primary modulator mask. The attenuation of the primary modulator mask is first compensated for with a correction matrix which takes advantage of the spectral information. The other contributions are a scatter model based on B-splines allowing an accurate representation of scatter maps with the aid of a very few parameters and a cost function which takes into account the structures of the mask and the object. The accuracy of the correction matrix, the scatter model and the whole proposed scatter correction process were tested on simulated data considering photon counting detectors with various numbers of energy bins and have shown a significant bias reduction, contrast enhancement and artifact removal. In addition, physical experiments were performed using a parallel fan-beam set-up with a commercial energy-resolved detector. The method was successfully validated in the case of two phantoms dedicated to medical image quality measurements, with a remarkable improvement

La décomposition de matériaux est un problème fondamental et primordial dans la tomographie spectrale (sCT—spectral computed tomography) par rayons X basée sur des détecteurs à comptage de photons (PCD—photon counting detector). La présente thèse porte sur le développement de méthodes de décomposition de matériaux en utilisant des informations spectrale et morphologique encodées dans des images sCT multi-énergie. Dans ce cadre, trois méthodes ont été développées. Pour la première méthode, en utilisant la densité de masse limitée, la parcimonie conjointe locale, et le faible rang structurel (DSR) dans le domaine de l'image, nous obtenons une décomposition très précise de matériaux tels que le gadolinium, l'iode et le fer. Les résultats sur les données numériques et physiques du fantôme ont démontré que la méthode DSR proposée conduit à une décomposition plus précise que la méthode pseudo-inverse habituelle avec décomposition en valeur singulière (SVD—singular value decomposition) et la méthode de régularisation parcimonieuse courante avec contrainte de norme L1 (lasso). La deuxième méthode opère par région. Elle consiste à optimiser les matériaux de base en se basant sur la segmentation spatio-énergétique des régions d'intérêt (ROI—regions-of-interests) dans les images sCT, à réduire le bruit en faisant le moyennage des images spatiales multi-énergie, et à effectuer une décomposition fine des matériaux impliquant une matrice de décomposition optimisée, une régularisation du débruitage et une régularisation parcimonieuse. Les résultats sur des données numériques et physiques ont montré que la méthode proposée de décomposition des matériaux ROI par ROI (ROI-wise—region-of-interests-wise) présente une fiabilité et une précision nettement supérieures à celles des méthodes de décomposition courantes fondées sur la régularisation de la variation totale (TV) ou de la norme L1. Dans la troisième méthode, nous proposons la notion d'imagerie sCT à super-résolution énergétique (SER—super-energy-resolution), qui est réalisée en établissant la relation entre la simulation et les fantômes physiques au moyen d'un apprentissage par dictionnaire couplé, de manière pixel par pixel. L'efficacité de ces méthodes proposées a été validée sur des données de fantômes numériques, de fantômes physiques et in vivo. Les résultats montrent que, pour la même méthode de décomposition de matériaux utilisant la régularisation par lasso, l'imagerie à super-résolution énergétique proposée présente une précision de décomposition et un pouvoir de détection beaucoup plus élevé que ce que peut fournir la machine sCT actuelle
Material decomposition is a fundamental and primordial problem in spectral photon-counting X-ray CT (sCT). The present thesis focuses on the development of material decomposition methods using spectral and morphological information embedded in multi-energy sCT images. In this framework, three methods were developed. For the first method, by using bounded mass density, local joint sparsity and structural low-rank (DSR) in image domain, we achieve highly accurate decomposition of materials such as gadolinium, iodine and iron. The results on both numerical phantom and physical data demonstrated that the proposed DSR method leads to more accurate decomposition than usual pseudo-inverse method with singular value decomposition (SVD) and current popular sparse regularization method with L1-norm constraint. The second method works in a region-wise manner. It consists in optimizing basis materials based on spatio-energy segmentation of regions-of-interests (ROIs) in sCT images, reducing noise by averaging multi-energy spatial images, and performing a fine material decomposition involving optimized decomposition matrix, denoising regularization and sparsity regularization. The results on both digital and physical data showed that the proposed ROI-wise material decomposition method presents clearly higher reliability and accuracy compared to common decomposition methods based on total variation (TV) or L1-norm (lasso) regularization. In the third method, we propose the notion of super-energy-resolution (SER) sCT imaging, which is realized through establishing the relationship between simulation and physical phantoms by means of coupled dictionary learning in a pixel-wise way. The effectiveness of the proposed methods was validated on digital phantom, physical phantoms and in vivo data. The results showed that for the same decomposition method using lasso regularization, the proposed super-energy-resolution imaging presents much higher decomposition accuracy and detection ability compared to what can be provided by current sCT machine

The goal in medical x-ray imaging is to obtain the image quality requiredfor a given detection task, while ensuring that the patient dose is kept as lowas reasonably achievable. The two most common strategies for dose reductionare: optimizing incident x-ray beams and utilizing energy informationof transmitted beams with new detector techniques (spectral imaging). Inthis thesis, dose optimization schemes were investigated in two x-ray imagingsystems: digital mammography and computed tomography (CT). In digital mammography, the usefulness of anti-scatter grids was investigatedas a function of breast thickness with varying geometries and experimentalconditions. The general conclusion is that keeping the grid is optimalfor breasts thicker than 5 cm, whereas the dose can be reduced without a gridfor thinner breasts. A photon-counting silicon-strip detector developed for spectral mammographywas characterized using synchrotron radiation. Energy resolution, ΔE/Ein, was measured to vary between 0.11-0.23 in the energy range 15-40 keV, which is better than the energy resolution of 0.12-0.35 measured inthe state-of-the-art photon-counting mammography system. Pulse pileup hasshown little effect on energy resolution. In CT, the performance of a segmented silicon-strip detector developedfor spectral CT was evaluated and a theoretical comparison was made withthe state-of-the-art CT detector for some clinically relevant imaging tasks.The results indicate that the proposed photon-counting silicon CT detector issuperior to the state-of-the-art CT detector, especially for high-contrast andhigh-resolution imaging tasks. The beam quality was optimized for the proposed photon-counting spectralCT detector in two head imaging cases: non-enhanced imaging and Kedgeimaging. For non-enhanced imaging, a 120-kVp spectrum filtered by 2half value layer (HVL) copper (Z = 29) provides the best performance. Wheniodine is used in K-edge imaging, the optimal filter is 2 HVL iodine (Z = 53)and the optimal kVps are 60-75 kVp. In the case of gadolinium imaging, theradiation dose can be minimized at 120 kVp filtered by 2 HVL thulium (Z =69).

QC 20160401

Spectral imaging is an emerging area of X-ray imaging that includes all the techniques which exploit the energy-dependent absorption properties of the matter to either provide quantitative information about the scanned object or to discriminate different materials. Though the potentials of spectral imaging are well known for over many years, the interest on spectral techniques is rapidly increasing only in recent years due to the development and widespread of dedicated acquisition systems able to acquire simultaneous (or nearly simultaneous) X-ray images at different energies. In this context, the development of energy resolving X-ray photon counting detectors (XPCDs) and the increasing availability of synchrotron radiation can provide interesting solutions for spectral imaging applications such as e.g. the K-edge technique. This thesis presents the implementation and the optimization of two state-of-the art acquisition systems suitable for spectral imaging applications with particular interest on the K-edge technique. The first system is an acquisition setup for spectral Computed Tomography (CT) made by coupling a polychromatic source with an energy resolving XPCD implementing two energy thresholds. The work carried out with this setup concerned the thorough characterization of an innovative XPCD featuring a sharp energy resolution, the design of a dedicated image processing procedure to achieve high quality CT images and the development of an automated procedure to produce accurate 3D maps of a K-edge element within a sample. Moreover, to allow further optimization studies, a simulator able to reproduce X-ray images with energy resolving XPCDs has been modeled, developed and validated. The last part of this thesis presents the development and the implementation of an acquisition setup for spectral imaging at synchrotron sources based on bent-Laue crystal optics. The results achieved with a prototype setup optimized for energies around 20 keV are presented, the future developments of the technique are discussed.

X-ray Computed Tomography (CT) imaging has been playing a central role in clinical practice since it was invented in 1972. However, the traditional x-ray CT technique fails to distinguish different materials with similar density, especially for biological tissues. The lack of a quantitative imaging representation has constrained the application of CT technique from a broadening application such as personal or precision medicine. Therefore, my major thesis statement is to develop novel material-specific CT imaging techniques for molecular imaging in biological bodies. To achieve the goal, comprehensive studies were conducted to investigate three different techniques: x-ray fluorescence molecular imaging, material identification (specification) from photon counting CT, and photon counting CT data distortion correction approach based on deep learning. X-ray fluorescence molecular imaging (XFMI) has shown great promise as a low-cost molecular imaging modality for clinical and pre-clinical applications with high sensitivity. In this study, the effects of excitation beam spectrum on the molecular sensitivity of XFMI were experimentally investigated, by quantitatively deriving minimum detectable concentration (MDC) under a fixed surface entrance dose of 200 mR at three different excitation beam spectra. The result shows that the MDC can be readily increased by a factor of 5.26 via excitation spectrum optimization. Furthermore, a numerical model was developed and validated by the experimental data (≥0.976). The numerical model can be used to optimize XFMI system configurations to further improve the molecular sensitivity. Findings from this investigation could find applications for in vivo pre-clinical small-animal XFMI in the future. PCCT is an emerging technique that has the ability to distinguish photon energy and generate much richer image data that contains x-ray spectral information compared to conventional CT. In this study, a physics model was developed based on x-ray matter interaction physics to calculate the effective atomic number () and effective electron density () from PCCT image data for material identification. As the validation of the physics model, the and were calculated under various energy conditions for many materials. The relative standard deviations are mostly less than 1% (161 out of 168) shows that the developed model obtains good accuracy and robustness to energy conditions. To study the feasibility of applying the model with PCCT image data for material identification, both PCCT system numerical simulation and physical experiment were conducted. The result shows different materials can be clearly identified in the − map (with relative error ≤8.8%). The model has the value to serve as a material identification scheme for PCCT system for practical use in the future. As PCCT appears to be a significant breakthrough in CT imaging field, there exists severe data distortion problem in PCCT, which greatly limits the application of PCCT in practice. Lately, deep learning (DL) neural network has demonstrated tremendous success in medical imaging field. In this study, a deep learning neural network based PCCT data distortion correction method was proposed. When applying the algorithm to process the test dataset data, the accuracy of the PCCT data can be greatly improved (RMSE improved 73.7%). Compared with traditional data correction approaches such as maximum likelihood, the deep learning approach demonstrate superiority in terms of RMSE, SSIM, PSNR, and most importantly, runtime (4053.21 sec vs. 1.98 sec). The proposed method has the potential to facilitate the PCCT studies and applications in practice.
Doctor of Philosophy
X-ray Computed Tomography (CT) has played a central role in clinical imaging since it was invented in 1972. It has distinguishing characteristics of being able to generate three dimensional images with comprehensive inner structural information in fast speed (less than one second). However, traditional CT imaging lacks of material-specific capability due to the mechanism of image formation, which makes it cannot be used for molecular imaging. Molecular imaging plays a central role in present and future biomedical research and clinical diagnosis and treatment. For example, imaging of biological processes and molecular markers can provide unprecedented rich information, which has huge potentials for individualized therapies, novel drug design, earlier diagnosis, and personalized medicine. Therefore there exists a pressing need to enable the traditional CT imaging technique with material-specific capability for molecular imaging purpose. This dissertation conducted comprehensive study to separately investigate three different techniques: x-ray fluorescence molecular imaging, material identification (specification) from photon counting CT, and photon counting CT data distortion correction approach based on deep learning. X-ray fluorescence molecular imaging utilizes fluorescence signal to achieve molecular imaging in CT; Material identification can be achieved based on the rich image data from PCCT; The deep learning based correction method is an efficient approach for PCCT data distortion correction, and furthermore can boost its performance on material identification. With those techniques, the material-specific capability of CT can be greatly enhanced and the molecular imaging can be approached in biological bodies.

Les anévrismes intracrâniens constituent une pathologie critique de santé publique liée à la dégradation de la paroi d’artères cérébrales. Il n’existe actuellement aucune méthode permettant d'estimer le risque de rupture d’un anévrisme qui prenne en compte les propriétés mécaniques in vivo de la paroi anévrismale, pourtant reconnues comme essentielles dans le phénomène de rupture. Ce travail de doctorat s'inscrit dans un projet de grande envergure visant à améliorer les critères d’intervention, actuellement disponibles pour les praticiens, en développant un outil d'aide à la décision non invasif se basant sur l’état mécanique du tissu pour en évaluer la probabilité de rupture. Cet outil reposera sur la définition d'une relation entre la forme de l'anévrisme observé par imagerie clinique et une base de données contenant un ensemble d’images cliniques issues d’études préalables, associées aux propriétés mécaniques in vivo de la paroi et à une caractérisation de sa rupture. Pour produire cette base de données, un dispositif de déformation de la paroi anévrismale est développé dans le cadre du projet global. Ce travail doctoral se focalise sur (1) la calibration, l'optimisation et les tests in vitro de ce dispositif sur artères fantômes et (2) l’application in vivo du dispositif sur un modèle animal d'anévrisme intracrânien. Pour ce faire, un modèle numérique de l'expérimentation in vitro a été implémenté et validé au regard des résultats expérimentaux, grâce au développement d’une méthode de validation originale. Ce modèle éléments finis d’interaction fluide-structure a permis d'appréhender les incertitudes d'utilisation du dispositif au sein de l'anévrisme et d’aider au dimensionnement des artères fantômes. Le meilleur compromis en termes d'épaisseur et de souplesse de la paroi des artères fantômes a ainsi été identifié compte tenu des limites des techniques de fabrication. De plus, une procédure d'analyse inverse a été développée de sorte à estimer les caractéristiques mécaniques de la paroi anévrismale in vivo. Son utilisation repose sur la quantification de la déformation engendrée par le dispositif et visualisée par scanner spectral à comptage photonique, technique d’imagerie médicale émergente dont les résolutions spatio-temporelles permettent une sollicitation contrôlée du tissu sans risque accru de rupture. Les propriétés mécaniques identifiées sont cohérentes avec celles issues des caractérisations ex vivo d'anévrismes similaires disponibles dans la littérature. Enfin, un premier critère de rupture patient-spécifique de la paroi anévrismale, prenant en compte l’état de contrainte in vivo dans le tissu, a été proposé
Intracranial aneurysms are a critical public health condition linked to the degradation of the cerebral artery wall. There is currently no method for estimating the risk of aneurysm rupture that takes into account the in vivo mechanical properties of the aneurysm wall, which are believed to be essential in the rupture phenomenon. This doctoral work is part of a large-scale project aimed at improving the intervention criteria currently available to practitioners by developing a non-invasive decision-support tool based on the mechanical state of the tissue to assess the probability of rupture. This tool will be based on the definition of a relationship between the shape of the aneurysm observed by clinical imaging and a database containing a set of clinical images from previous studies, associated with the in vivo mechanical properties of the wall and a characterisation of the rupture. To produce this database, an aneurysm wall deformation device was developed as part of the overall project. This doctoral work focuses on (1) the calibration, the optimisation and in vitro testing of this device on phantom arteries and (2) the in vivo application of the device on an animal model of intracranial aneurysm. To do this, a numerical model of the in vitro experiment was implemented and validated against the experimental results by developing an original validation method. This finite element model of fluid-structure interaction was used to understand the uncertainties involved in using the device within the aneurysm and to help for dimensioning the phantom arteries. The best compromise in terms of phantom artery wall thickness and flexibility was identified, taking into account the limitations of the fabrication techniques. In addition, an inverse analysis procedure was developed to estimate the mechanical characteristics of the aneurysm wall in vivo. Its use is based on quantifying the deformation generated by the device and visualised by spectral photon-counting computed tomography, an emerging medical imaging technique whose spatio-temporal resolutions allow controlled stressing of the tissue without increasing the risk of rupture. The mechanical properties identified were consistent with those derived from ex vivo characterisations of similar aneurysms available in the literature. Finally, a first patient-specific criterion for rupture of the aneurysm wall, taking into account the state of stress in vivo in the tissue, was proposed

Computed tomography angiography (CTA) of the brain poses challenges on the imaging system; the contrast between blood vessels and surrounding soft tissue is very low, and to render small intricate vessel structures high spatial resolution is needed. Higher precision angiography would facilitate more accurate diagnosis of pathological conditions. The aim of this work was to analyze the factors which contribute to the image quality in cerebrovascular imaging contexts and make a comparison between state-of-the-art energy-integrating and photon counting CT systems. A geometrical model was devised to mimic the conditions of cerebral angiography. Different parameters and detectors were used to reconstruct images of a spherical head phantom. Compton noise was added to several image acquisitions after a Monte Carlo study was used to estimate the scatter to primary ratio (SPR) with a spherical phantom. The images were evaluated qualitatively and quantitatively. A real phantom was scanned with an experimental photon counting detector and compared with the simulated approach. The work resulted in qualitative reconstructed images, a decrease in SPR when introducing air gaps and improved resolution but worsened contrast as a result of smaller detector sizes. The SPR was shown to be higher in cone-beam geometry than fan-beam geometry. Electronic noise present with energy integrating detectors was shown to degrade image quality significantly in low dose imaging, reducing contrast when imaging vascular-like structures. Photon counting detectors without electronic noise could provide greater image quality and better diagnostic information.

La tomographie (CT) aux rayons X joue un rôle important dans l'imagerie non invasive depuis son introduction. Au cours des dernières années, de nombreuses avancées technologiques en tomographie par rayons X ont été observées, notamment la CT spectrale, qui utilise un détecteur à comptage de photons (PCD) pour discriminer les photons transmis correspondant à des bandes d'énergie sélectionnées afin d'obtenir une information spectrale. La CT spectrale permet de surmonter de nombreuses limitations des techniques précédentes et ouvre de nombreuses applications nouvelles, parmi lesquelles la décomposition quantitative des matériaux est le sujet le plus étudié. Un certain nombre de méthodes de décomposition des matériaux ont été rapportées et différents systèmes expérimentaux sont en cours de développement pour la CT spectrale. Selon le type de données sur lequel l'étape de décomposition fonctionne, nous avons les méthodes du domaine des projections (décomposition avant reconstruction) et les méthodes du domaine de l'image reconstruite (décomposition après reconstruction). La décomposition couramment utilisée est basée sur le critère des moindres carrés, nommée proj-LS et méthode ima-LS. Cependant, le problème inverse de la décomposition du matériau est généralement mal posé et les mesures du CT spectral aux rayons X souffrent de bruits de comptage de photons de Poisson. Le critère des moindres carrés peut conduire à un surajustement des données de mesure bruitées. Dans le présent travail, nous avons proposé un critère de moindre log-carré pour la méthode du domaine de projection afin de minimiser les erreurs sur le coefficient d'atténuation linéaire: méthode proj-LLS. De plus, pour réduire l'effet du bruit et lisser les images, nous avons proposé d'ajouter un terme de régularisation par patch pour pénaliser la somme des variations au carré dans chaque zone pour les décompositions des deux domaines, nommées proj-PR-LLS et ima -PR-LS méthode. Les performances des différentes méthodes ont été évaluées par des études de simulation avec des fantômes spécifiques pour différentes applications: (1) Application médicale: identification de l'iode et du calcium. Les résultats de la décomposition des méthodes proposées montrent que le calcium et l'iode peuvent être bien séparés et quantifiés par rapport aux tissus mous. (2) Application industrielle: tri des plastiques avec ou sans retardateur de flamme. Les résultats montrent que 3 types de matériaux ABS avec différents retardateurs de flamme peuvent être séparés lorsque l'épaisseur de l'échantillon est favorable. Enfin, nous avons simulé l'imagerie par CT spectrale avec un fantôme de PMMA rempli de solutions de Fe, Ca et K. Différents paramètres d'acquisition, c'est-à-dire le facteur d'exposition et le nombre de bandes d'énergie, ont été simulés pour étudier leur influence sur la performance de décomposition pour la détermination du fer
X-ray computed tomography (X-ray CT) plays an important part in non-invasive imaging since its introduction. During the past few years, numerous technological advances in X-ray CT have been observed, including spectral CT, which uses photon counting detectors (PCDs) to discriminate transmitted photons corresponding to selected energy bins in order to obtain spectral information with one single acquisition. Spectral CT enables us to overcome many limitations of the conventional CT techniques and opens up many new application possibilities, among which quantitative material decomposition is the hottest topic. A number of material decomposition methods have been reported and different experimental systems are under development for spectral CT. According to the type of data on which the decomposition step operates, we have projection domain method (decomposition before reconstruction) and image domain method (decomposition after reconstruction). The commonly used decomposition is based on least square criterion, named proj-LS and ima-LS method. However, the inverse problem of material decomposition is usually ill-posed and the X-ray spectral CT measurements suffer from Poisson photon counting noise. The standard LS criterion can lead to overfitting to the noisy measurement data. In the present work, we have proposed a least log-squares criterion for projection domain method to minimize the errors on linear attenuation coefficient: proj-LLS method. Furthermore, to reduce the effect of noise and enforce smoothness, we have proposed to add a patchwise regularization term to penalize the sum of the square variations within each patch for both projection domain and image domain decomposition, named proj-PR-LLS and ima-PR-LS method. The performances of the different methods were evaluated by spectral CT simulation studies with specific phantoms for different applications: (1) Medical application: iodine and calcium identification. The decomposition results of the proposed methods show that calcium and iodine can be well separated and quantified from soft tissues. (2) Industrial application: ABS-flame retardants (FR) plastic sorting. Results show that 3 kinds of ABS materials with different flame retardants can be separated when the sample thickness is favorable. Meanwhile, we simulated spectral CT imaging with a PMMA phantom filled with Fe, Ca and K solutions. Different acquisition parameters, i.e. exposure factor and number of energy bins were simulated to investigate their influence on the performance of the proposed methods for iron determination

Dans le domaine de la tomographie par rayons X préclinique, la tomographie spectrale est une voie de plus en plus en plus explorée. Les objectifs de la tomographie spectrale sont la caractérisation et la quantification des tissus et agents de contraste que l'amélioration de contraste entre tissus mous. Cela passe par l'exploitation de l'information spectrale (ou énergétique) des photons X et non plus seulement par leur quantité détectée comme en tomographie standard par rayons X. L'intérêt de la tomographie spectrale se trouve renforcé par l'arrivée des caméras à pixels hybrides comme le XPAD qui ont la capacité de sélectionner les photons X en fonction de leur énergie. La caméra XPAD3, la troisième version du XPAD est construite pour fonctionner dans le micro-tomodensitomètre, PIXSCAN développé au CPPM.Dans ce contexte, cette thèse a deux buts~: participer au développement du PIXSCAN et effectuer une preuve de concept de la tomographie spectrale à l'aide du PIXSCAN. Le premier but est rempli grâce au développement de l'interface d'acquisition du PIXSCAN. Le second est accompli par l'implantation de la méthode par séparation de composantes dont le but est d'isoler les contributions photoélectrique, Compton et des agents de contraste. Ce travail débute par la caractérisation de la méthode et se termine par la preuve de concept sur données réelles acquises à l'aide du PIXSCAN
In the field of preclinical X-ray tomography, spectral tomography is actively explorated. The aims of spectral tomography are the caracterisation of tissues and contrast agentstogether with the quantification of the latter and the enhancement of contrast between soft tissues. This is achived by the exploitation of spectral information (i.e. energy) and not only the detected quantities of photons X. The interest in spectral tomography is enforced by the arrival of hybrid pixel cameras like XPAD, because of their ability to select photons according to their energy. The XPAD3 camera, third version of XPAD, is built to be used in the micro-CT demonstrator PIXCAN fully developped at CPPM.In this context, this thesis has two goals : a contribution to the developement of the PIXSCAN and a realisation of a proof of concept of spectral tomography in PIXSCAN. The first goal is done by developing the data acquisition system of PIXSCAN. To accomplish the second one, we perform spectral tomography by implementing component separation in order to isolate photoelectric, compton and contrast agents contribution. This work begins by the caracterisation of this method and ends by a proof of concept on real data acquired by PIXSCAN

Semiconductor detectors are attracting interest for use in photon-counting spectral computed tomography. In order to obtain a high spatial resolution, it is of interest to find the photon interaction position. In this work we investigate if machine learning can be used to obtain a sub-pixel spatial resolution in a photon-counting silicon strip detector with pixels of 10 µm. Simulated charge distributions from events in one, three, and seven positions in each of three pixels were investigated using the MATLAB® Classification Learner application to determine the correct interaction position. Different machine learning models were trained and tested in order to maximize performance. With pulses originating from one and seven positions within each pixel, the model was able to find the originating pixel with an accuracy of 100% and 88.9% respectively. Further, the correct position within a pixel was found with an accuracy of 54.0% and 29.4% using three and seven positions per pixel respectively. These results show the possibility of improving the spatial resolution with machine learning.
Halvledardetektorer är av stigande intresse inom forskning för användning i fotonräknande datortomografi med spektral upplösning. För att erhålla en hög spatiell upplösning är det av intresse att hitta fotonens ursprungliga interaktionsposition. I detta arbete undersöks om maskininlärning kan användas för att erhålla en spatiell upplösning på subpixelnivå i en fotonräknande kiselstrippdetektor med 10 µm pixlar. Laddningsfördelningen från simulerade interaktioner i en, tre, och sju positioner inom var och en av tre pixlar undersöktes med hjälp av applikationen Classification Learner i MATLAB® för att bestämma den korrekta interaktionspositionen. Olika maskininlärningsmodeller tränades och testades för att maximera prestandan. När pulser från en och sju positioner inom pixeln användes, kunde modellen hitta den korrekta pixeln med en noggrannhet på 100% respektive 88.9%. Vidare kunde den korrekta positionen inom en pixel bestämmas med en noggrannhet på 54.0% och 29.4% när tre respektive sju positioner inom varje pixel användes. Resultaten visar att det skulle vara möjligt att förbättra den spatiella upplösningen med hjälp av maskininlärning.

La tomodensitométrie par rayons-X (CT) est une modalité d’imagerie produisant une carte tridimensionnelle du coefficient d’atténuation des rayons-X d’un objet. En radiothérapie, le CT fournit de l’information anatomique et quantitative sur le patient afin de permettre la planification du traitement et le calcul de la dose de radiation à livrer. Le CT a plusieurs problèmes, notamment (1) une limitation au niveau de l’exactitude des paramètres physiques quantitatifs extraits du patient, et (2) une sensibilité aux biais causés par des artéfacts de durcissement du faisceau. Enfin, (3) dans le cas où le CT est fait en présence d’un agent de contraste pour améliorer la planification du traitement, il est nécessaire d’effectuer un deuxième CT sans agent de contraste à des fins de calcul de dose, ce qui augmente la dose au patient. Ces trois problèmes limitent l’efficacité du CT pour certaines modalités de traitement qui sont plus sensibles aux incertitudes comme la protonthérapie. Le CT spectral regroupe un ensemble de méthodes pour produire plusieurs cartes d’atténuation des rayons-X moyennées sur différentes plages énergétiques. L’information supplémentaire, pondérée en énergie qui est obtenue permet une meilleure caractérisation des matériaux analysés. Le potentiel de l’une de ces modalités spectrales, le CT bi-énergie (DECT), est déjà bien démontré en radiothérapie, alors qu’une approche en plein essor, le CT spectral à comptage de photons (SPCCT), promet davantage d’information spectrale à l’aide de détecteurs discriminateurs en énergie. Par contre, le SPCCT souffre d’un bruit plus important et d’un conditionnement réduit. Cette thèse investigue la question suivante : y a-t-il un bénéfice à utiliser plus d’information résolue en énergie, mais de qualité réduite pour la radiothérapie ? La question est étudiée dans le contexte des trois problèmes ci-haut. Tout d’abord, un estimateur maximum a posteriori (MAP) est introduit au niveau de la caractérisation des tissus post-reconstruction afin de débruiter les données du CT spectral. L’approche est validée expérimentalement sur un DECT. Le niveau de bruit du pouvoir d’arrêt des protons diminue en moyenne d’un facteur 3.2 à l’aide de l’estimateur MAP. Celui-ci permet également de conserver généralement le caractère quantitatif des paramètres physiques estimés, le pouvoir d’arrêt variant en moyenne de 0.9% par rapport à l’approche conventionnelle. Ensuite, l’estimateur MAP est adapté au contexte de l’imagerie avec agent de contraste. Les résultats numériques démontrent un bénéfice clair à utiliser le SPCCT pour l’imagerie virtuellement sans contraste par rapport au DECT, avec une réduction de l’erreur RMS sur le pouvoir d’arrêt des protons de 2.7 à 1.4%. Troisièmement, les outils développés ci-haut sont validés expérimentalement sur un micro-SPCCT de la compagnie MARS Bioimaging, dont le détecteur à comptage de photons est le Medipix 3, qui est utilisé pour le suivi de particules au CERN. De légers bénéfices au niveau de l’estimation des propriétés physiques à l’aide du SPCCT par rapport au DECT sont obtenus pour des matériaux substituts à des tissus humains. Finalement, une nouvelle paramétrisation du coefficient d’atténuation pour l’imagerie pré-reconstruction est proposée, dans le but ultime de corriger les artéfacts de durcissement du faisceau. La paramétrisation proposée élimine les biais au niveau de l’exactitude de la caractérisation des tissus humains par rapport aux paramétrisations existantes. Cependant, aucun avantage n’a été obtenu à l’aide du SPCCT par rapport au DECT, ce qui suggère qu’il est nécessaire d’incorporer l’estimation MAP dans l’imagerie pré-reconstruction via une approche de reconstruction itérative.
X-ray computed tomography (CT) is an imaging modality that produces a tridimensional map of the attenuation of X-rays by the scanned object. In radiation therapy, CT provides anatomical and quantitative information on the patient that is required for treatment planning. However, CT has some issues, notably (1) a limited accuracy in the estimation of quantitative physical parameters of the patient, and (2) a sensitivity to biases caused by beam hardening artifacts. Finally, (3) in the case where contrast-enhanced CT is performed to help treatment planning, a second scan with no contrast agent is required for dose calculation purposes, which increases the overall dose to the patient. Those 3 problems limit the efficiency of CT for some treatment modalities more sensitive to uncertainties, such as proton therapy. Spectral CT regroups a set of methods that allows the production of multiple X-ray attenuation maps evaluated over various energy windows. The additional energy-weighted information that is obtained allows better material characterization. The potential of one spectral CT modality, dual-energy CT (DECT), is already well demonstrated for radiation therapy, while an upcoming method, spectral photon counting CT (SPCCT), promises more spectral information with the help of energy discriminating detectors. Unfortunately, SPCCT suffers from increased noise and poor conditioning. This thesis thus investigates the following question: is there a benefit to using more, but lower quality energy-resolved information for radiotherapy? The question is studied in the context of the three problems discussed earlier. First, a maximum a posteriori (MAP) estimator is introduced for post-reconstruction tissue characterization for denoising purposes in spectral CT. The estimator is validated experimentally using a commercial DECT. The noise level on the proton stopping power is reduced, on average, by a factor of 3.2 with the MAP estimator. The estimator also generally con- serves the quantitative accuracy of estimated physical parameters. For instance, the stopping power varies on average by 0.9% with respect to the conventional approach. Then, the MAP estimation framework is adapted to the context of contrast-enhanced imaging. Numerical results show clear benefits when using SPCCT for virtual non-contrast imaging compared to DECT, with a reduction of the RMS error on the proton stopping power from 2.7 to 1.4%. Third, the developed tools are validated experimentally on a micro-SPCCT from MARS Bioimaging, which uses the Medipix 3 chip as a photon counting detector. Small benefits in the accuracy of physical parameters of tissue substitutes materials are obtained. Finally, a new parametrization of the attenuation coefficient for pre-reconstruction imaging is pro- posed, whose ultimate aim is to correct beam hardening artifacts. In a simulation study, the proposed parametrization eliminates all biases in the estimated physical parameters of human tissues, which is an improvement upon existing parametrizations. However, no ad- vantage has been obtained with SPCCT compared to DECT, which suggests the need to incorporate MAP estimation in the pre-reconstruction framework using an iterative reconstruction approach.

Contrast agents are often crucial in medical imaging for disease diagnosis. Novel contrast agents, such as gold nanoparticles (AuNPs) and lanthanides, are being ex- plored for a variety of clinical applications. Preclinical testing of these contrast agents is necessary before being approved for use in humans, which requires the use of small animal imaging techniques. Small animal imaging demands the detection of these contrast agents in trace amounts at acceptable imaging time and radiation dose. Two such imaging techniques include x-ray fluorescence computed tomography (XFCT) and photon-counting CT (PCCT). XFCT combines the principles of CT with x-ray fluorescence by detecting fluorescent x-rays from contrast agents at various projections to reconstruct contrast agent maps. XFCT can image trace amounts of AuNPs but is limited to small animal imaging due to fluorescent x-ray attenuation and scatter. PCCT uses photon-counting detectors that separate the CT data into energy bins. This enables contrast agent detection by recognizing the energy dependence of x-ray attenuation in different materials, independent of AuNP depth, and can provide anatomical information that XFCT cannot. To achieve the best of both worlds, we modeled and built a table-top x-ray imaging system capable of simultaneous XFCT and PCCT imaging. We used Monte Carlo simulation software for the following work in XFCT imaging of AuNPs. We simulated XFCT induced by x-ray, electron, and proton beams scanning a small animal-sized object (phantom) containing AuNPs with Monte Carlo techniques. XFCT induced by x-rays resulted in the best image quality of AuNPs, however high-energy electron and medium-energy proton XFCT may be feasible for on-board x-ray fluorescence techniques during radiation therapy. We then simulated a scan of a phantom containing AuNPs on a table-top system to optimize the detector arrangement, size, and data acquisition strategy based on the resulting XFCT image quality and available detector equipment. To enable faster XFCT data acquisition, we separately simulated another AuNP phantom and determined the best collimator geometry for Au fluorescent x-ray detection. We also performed experiments on our table-top x-ray imaging system in the lab. Phantoms containing multiples of three lanthanide contrast agents were scanned on our tabletop x-ray imaging system using a photon-counting detector capable of sustaining high x-ray fluxes that enabled PCCT. We used a novel subtraction algorithm for reconstructing separate contrast agent maps; all lanthanides were distinct at low concentrations including gadolinium and holmium that are close in atomic number. Finally, we performed the first simultaneous XFCT and PCCT scan of a phantom and mice containing both gadolinium and gold based on the optimized parameters from our simulations. This dissertation outlines the development of our tabletop x-ray imaging system and the optimization of the complex parameters necessary to obtain XFCT and PCCT images of multiple contrast agents at biologically-relevant concentrations.
Graduate