Objectifs : Etablir des corrélations dose-volume effet entre les paramètres dosimétriques proposés par le GEC-ESTRO et la probabilité de survenue d’événements tels que le contrôle tumoral ou une toxicité radio-induite.Matériel et méthodes : Les données cliniques et dosimétriques de cohortes de patientes traitées à Gustave Roussy et dans différents centres ont été confrontées. Dans un premier temps les paramètres dosimétriques de la curiethérapie 3D ont été comparés à ceux de la curiethérapie classique. Dans un second temps, la topographie des zones les plus exposées des organes à risque, ainsi que l’impact des mouvements de la vessie, du rectum, et du colon sigmoïde sur l’évaluation de la dose délivrée, ont été étudiés. Enfin, des analyses dose-volume effets ont été réalisées.Résultats : Les valeurs des paramètres dosimétriques volumétriques (D2cm3) de la vessie et du rectum se sont révélées faiblement corrélées et significativement supérieures aux doses évaluées aux points de l’ICRU ou à un point vésical alternatif. Les zones les plus exposées de la vessie et du rectum sont apparues situées au-dessus des points de l’ICRU. Les mouvements des organes autour de l’implant pendant la délivrance du traitement sont apparus marginaux pour la vessie et sigmoïde, en dehors de variations individuelles. En revanche, la dose délivrée au rectum étaient en moyenne plus élevée que le dose planifiée. Les analyses dose-volume effets ont montré des corrélations significatives entre D0.1cm3 et D2cm3 et la probabilité de survenue d’une morbidité tardive urinaire ou rectale. De la même manière, des corrélations significatives ont été établies entre la D90 des CTV à haut risque et à risque intermédiaire et la probabilité d’obtention du contrôle local. Divers caractéristiques tumorales (largeur au diagnostic, volume du CTV-HR, stade FIGO), impactent ces relations, de même que l’étalement total du traitement.Conclusion : Des corrélations dose-volume effets ont été établies entre les paramètres dosimétriques modernes et la probabilité d’obtenir le contrôle local ou d’entraîner une morbidité tardive. En ce qui concerne le contrôle tumoral, les objectifs de prescription doivent être personnalisés en fonction de critères carcinologiques. Pour les organes à risque, de contraintes de dose basées sur l’expérience de la curiethérapie 3D peuvent être établies, mais doivent être affinées dans de futures études en fonction de cofacteurs tels que les comorbidités. Les points gardent un intérêt en recherche clinique, pour l’étude de la morbidité vésicale ou vaginale.Ce travail a l'objet de 6 publications dans des revues internationales à comité de lecture. La septième est présentée sous forme de manuscrit. / Objectives: To establish dose-volume effects correlations between volumetric dosimetric parameters proposed by the GEC-ESTRO and the probability of occurrence of events such as tumor control or radiation-induced toxicity.Methods: Clinical and dosimetric data of patients treated at Gustave Roussy and in different centers have been reviewed. At first step, dosimetric parameters of image-guided brachytherapy were compared with those of conventional brachytherapy. Secondly, the topography of the most exposed areas of the organs at risk, and the impact of the movements of the bladder, rectum, and sigmoid colon on the assessment of the delivered dose, were studied. Finally, analyzes dose-volume effects were performed.Results: The values of volumetric dosimetric parameters (D2cm3) of the bladder and rectum appeared weakly correlated and significantly higher than the doses evaluated at ICRU points of bladder and rectum , an even in an alternative bladder point. The most exposed areas of the bladder and rectum appeared located above the points of the ICRU. The movements of the organs around the implant during the delivery of the treatment appeared marginal for the bladder and sigmoid, apart from individual variations. However, the mean delivered dose to the rectum was higher than the planned dose. Dose-volume effects correlations showed significant correlations between D0.1cm3 and D2cm3 and the probability of occurrence of urinary or rectal late morbidity. Similarly, significant correlations have been established between the D90 of the high risk, intermediate risk-CTV and the probability of achieving local control. Various tumor characteristics (width, HR-CTV volume, FIGO stage) impact these relationships, as well as the treatment time.Conclusion: Dose-volume effects correlations have been established between modern dosimetric parameters and the probability of achieving local control or cause late morbidity. Regarding tumor control, prescription aims must be customized according to oncologic criteria. For organs at risk, new dose constraints based on 3D brachytherapy experience can be established but should be refined in future studies based on cofactors such as comorbidities. The points retain an interest in clinical research for the study of bladder or vaginal morbidity.

Das entwickelte Bildgeführte Präzisionsbestrahlungsgerät für Kleintiere (engl. Small Animal Image-Guided Radiation Therapy – SAIGRT) dient der schnellen, hochauflösenden Röntgenbildgebung und präzisen, konformalen Bestrahlung von Kleintieren im Rahmen präklinischer in-vivo Experimente für die translationale Krebsforschung. Speziell programmierte Softwares zur Gerätesteuerung sowie zur Bildkorrektur- und Bildrekonstruktion auf dem zentralen leistungsfähigen Arbeitsplatz-PC stellen alle Gerätefunktionen zur Verfügung und ermöglichen durch automatisierte Abläufe und intuitive grafische Nutzeroberflächen eine einfache, sichere Bedienung. Für die Bestrahlungsplanung wird eine vollwertige, aus der humanen klinischen Strahlentherapie adaptierte 3D-Bestrahlungsplanungssoftware eingesetzt, die etablierte Werkzeuge für den Transfer und die Koregistrierung multimodaler Bilddaten, die Konturierung und Segmentierung von Zielvolumina und Risikoorganen sowie die Erstellung und Validierung von Bestrahlungsplänen enthält. Die resultierende Dosisverteilung wird darin basierend auf dem individuellen CT-Datensatz des Versuchstieres und einem auf das SAIGRT angepassten Maschinenmodell mittels eines Monte-Carlo-Algorithmus exakt und realitätsnah simuliert. Durch geometrische Kalibrierungen und vielfältige Basisdatenmessungen für die Bildgebung und Bestrahlung im Rahmen der Gerätekommissionierung ist eine Zielgenauigkeit von ca. ±0,1 mm mit hoher geometrischer Abbildungstreue und guter Bildqualität bei Bildgebungsdosen vergleichbar denen klinischer Radiografie- und CT-Geräte möglich. Die Dosisverteilung zur Bestrahlung der Versuchstiere spiegelt bei der definierten Strahlungsqualität größenskaliert die humane Strahlentherapie mit hochenergetischer Photonenstrahlung von klinischen Linearbeschleunigern wider. Ein umfassendes Qualitätssicherungsprogramm bestehend aus regelmäßiger Wartung und wiederkehrenden Konstanzprüfungen der Bildgebung und Bestrahlung sichert dauerhaft den technisch einwandfreien Zustand und die ordnungsgemäße Verfügbarkeit aller Gerätefunktionen in gleichbleibender Güte. Das SAIGRT ist somit nachweislich geeignet, bildgeführte Bestrahlungen mit einem Ablauf analog dem einer modernen klinischen Strahlentherapie am Menschen in präklinischen in-vivo Experimenten präzise an Kleintieren zu applizieren. Es leistet dadurch einen essentiellen Beitrag zur translationalen Krebsforschung in Dresden, indem die klinische Situation realistischer modelliert und so potenziell die Übertragbarkeit der Ergebnisse auf Krebspatienten verbessert werden kann. / The Small Animal Image-Guided Radiation Therapy (SAIGRT) platform facilitates fast, high resolution X-ray imaging and precise, conformal irradiation of small animals in preclinical in-vivo experiments for translational cancer research. Dedicated software for device control as well as image correction and reconstruction on a central high performance PC provide all device functions and allow simple and safe operation by automated procedures and intuitive graphical user interfaces. A fully 3D treatment planning software adapted from human clinical radiation therapy is used for treatment planning, containing established tools and methods for the transfer and registration of multimodality imaging data, contouring and segmentation of target volumes and organs at risk as well as creation and evaluation of treatment plans. Based on an individual CT scan of the small animal and a machine model adapted for the SAIGRT, the resulting dose distribution is simulated by a Monte-Carlo algorithm in a precise and realistic manner. Geometrical calibrations as well as manifold basic data measurements for X-ray imaging and irradiation during commissioning resulted in a targeting and imaging accuracy of about ±0.1 mm, a correct representation of imaging geometry and a good image quality with imaging doses comparable with those of clinical radiography and CT systems. Dose distribution of the defined beam quality used for irradiation of small animals reflects a downsized human radiation therapy using high energy photon beams of clinical linear accelerators. A comprehensive quality assurance program comprising regular maintenance and periodic constancy tests of X-ray imaging and irradiation ensures permanent technically perfect condition and proper availability of all implemented functions in a stable high quality. The SAIGRT platform is feasible for image-guided irradiations precisely applied to small animals in preclinical in-vivo experiments using a workflow of modern human radiation oncology. Thus, it significantly contributes to translational cancer research by more realistic modelling the clinical situation and potentially brings the results closer to their clinical implementation.

In recent years, texture synthesis has been a heated topic in computer graphics, and the development of advanced algorithms for generating high-quality 3D textures is an area of active research. A recently proposed model, Neural Cellular Automata, can synthesize realistic 2D texture images or videos. However, due to the complexity and non-differentiable nature of 3D rendering and the lack of definition of the neighborhood on 3D mesh objects, no one has extended the 2D Neural Cellular Automata to the 3D scenario. In this master’s thesis, we propose a novel method for modeling the neighborhood relationship on 3D mesh objects, drawing inspiration from a graph variant of the Neural Cellular Automata. We also design an end-to-end 3D texture synthesis pipeline, leveraging a differentiable renderer to enable the Graph Neural Cellular Automata to learn to synthesize desired 3D textures. Our method allows users to either give the text description of the target textures or present the target texture images as the objectives. We evaluate the effectiveness of our proposed method both qualitatively and quantitatively, comparing it with the state-of-the-art method to demonstrate that it achieves comparable or better results. Furthermore, we explore the homology between the graph variant of Neural Cellular Automata and the 2D model, examining whether our proposed model preserves critical properties of the 2D model such as zero-shot generalization and self-regeneration. Finally, we analyze the limitations and potential drawbacks of our proposed method and suggest directions for future research. In summary, this thesis proposes a novel approach to synthesizing high-quality 3D textures using the Graph Neural Cellular Automata model and a differentiable renderer. Our work provides a foundation for future research in this area, and we believe that our findings will contribute to the development of advanced algorithms for 3D texture synthesis. / Under de senaste åren har textursyntes varit ett hett ämne inom datorgrafik, och utvecklingen av avancerade algoritmer för att generera högkvalitativa 3D-texturer är ett aktivt forskningsområde. En nyligen föreslagen modell, Neural Cellular Automata, kan syntetisera realistiska 2D-texturbilder eller videor. Dock, på grund av komplexiteten och den icke-differentierbara naturen av 3D-rendering och bristen på definition av grannskapet på 3D-meshobjekt, har ingen utvidgat 2D Neural Cellular Automata till 3D-scenariot. I den här masteruppsatsen föreslår vi en ny metod för att modellera grannskapsrelationen på 3D-meshobjekt, inspirerade av en grafvariant av Neural Cellular Automata. Vi utformar också en ände-till-ände 3D-textursyntes pipeline, genom att utnyttja en differentierbar renderer för att möjliggöra för Graph Neural Cellular Automata att lära sig syntetisera önskade 3D-texturer. Vår metod tillåter användare att antingen ge textbeskrivningen av måltexturerna eller presentera måltexturbilderna som målen. Vi utvärderar effektiviteten av vår föreslagna metod både kvalitativt och kvantitativt, jämför den med den mest avancerade metoden för att visa att den uppnår jämförbara eller bättre resultat. Dessutom utforskar vi homologin mellan grafvarianten av Neural Cellular Automata och 2D-modellen, undersöker om vår föreslagna modell bevarar kritiska egenskaper hos 2D-modellen som zero-shot generalisering och självregenerering. Slutligen analyserar vi begränsningarna och eventuella nackdelar med vår föreslagna metod och föreslår riktningar för framtida forskning. Sammanfattningsvis föreslår denna avhandling en ny metod för att syntetisera högkvalitativa 3D-texturer med hjälp av Graph Neural Cellular Automata-modellen och en differentierbar renderer. Vårt arbete ger en grund för framtida forskning inom detta område, och vi tror att våra fynd kommer att bidra till utvecklingen av avancerade algoritmer för 3D-textursyntes.

No description available.

No description available.

Le système perceptif d’un chirurgien doit s’adapter aux contraintes multisensorielles liées à la chirurgie guidée par l’image. Trois expériences sont conçues pour explorer ces contraintes visuelles et haptiques pour l’apprentissage guidé par l’image. Les résultats montrent que les sujets sont plus rapides et plus précis avec une vision directe. La stéréoscopie 3D n’améliore pas les performances des débutants complets. En réalité virtuelle, la variation de la longueur, largeur, position et complexité de l'objet affecte les performances motrices. La force de préhension appliquée sur un système robotique chirurgical dépend de l'expérience de l'utilisateur. En conclusion, le temps et la précision sont importants, mais la précision doit rester une priorité pour un apprenti. L'homogénéité des groupes d'étude est important pour la recherche sur la formation chirurgicale. Les résultats ont un impact direct sur le suivi des compétences individuelles pour les applications guidées par l'image. / Perceptual system of a surgeon must adapt to conditions of multisensorial constrains regard to planning, control, and execution of the image-guided surgical operations. Three experimental setups are designed to explore these visual and haptic constraints in the image-guided training. Results show that subjects are faster and more precise with direct vision compared to image guidance. Stereoscopic 3D viewing does not represent a performance advantage for complete beginners. In virtual reality, variation in object length, width, position, and complexity affect the motor performance. Applied grip force on a surgical robot system depends on the user experience level. In conclusion, both time and precision matter critically, but trainee gets as precise as possible before getting faster should be a priority. Study group homogeneity and background play key role in surgical training research. The findings have direct implications for individual skill monitoring for image-guided applications.

The work of this Thesis is focused on image guided focal therapy of prostate cancer by High Intensity Focused Ultrasound (HIFU). Currently MRI is the only imaging technique that can locate the tumor in prostate. In contrast, the tumor is not visible in the ultrasound image which is used to guide the HIFU planning and therapy. The aim of the Thesis is to provide registration techniques of T2 MRI to ultrasound. Two approaches were explored: 1) Region-based registration. More particularly, we studied an ultrasound texture descriptors based on moments invariant to rotation and scaling. These descriptors are sensitive to speckle distribution regardless of the scale or the orientation. As we expected, some of these descriptors can be used to characterize regions sharing a similar speckle spatial distribution. But, we also found that some other descriptors were sensitive to the contours of these regions. This property seems very useful to adapt the classical boundary-based or mixed region/boundary-based segmentation methods (active contours, graph cut, etc.) to process US images. 2) Surface-based registration approach.. We adapted the Optimal Definition Surface (OSD) method to the segmentation of the prostate in T2 MRI, Furthermore, we proposed the multiple-objects OSD which is a concurrent segmentation of the prostate, bladder and rectum. Finally we used the prostate surface extracted from the ultrasound volume and from T2 MRI in a surface-to-surface elastic registration scheme. This registration allowed us to merge the preoperative MR information in the peroperative US volume.

This thesis addresses three problems related to image registration, prediction and tracking, applied to Angiography and Oncology. For image analysis, various probabilistic models have been employed to characterize the image deformations, target motions and state estimations. (i) In Digital Subtraction Angiography (DSA), having a high quality visualization of the blood motion in the vessels is essential both in diagnostic and interventional applications. In order to reduce the inherent movement artifacts in DSA, non-rigid image registration is used before subtracting the mask from the contrast image. DSA image registration is a challenging problem, as it requires non-rigid matching across spatially non-uniform control points, at high speed. We model the problem of sub-pixel matching, as a labeling problem on a non-uniform Markov Random Field (MRF). We use quad-trees in a novel way to generate the non uniform grid structure and optimize the registration cost using graph-cuts technique. The MRF formulation produces a smooth displacement field which results in better artifact reduction than with the conventional approach of independently registering the control points. The above approach is further improved using two models. First, we introduce the concept of pivotal and non-pivotal control points. `Pivotal control points' are nodes in the Markov network that are close to the edges in the mask image, while 'non-pivotal control points' are identified in soft tissue regions. This model leads to a novel MRF framework and energy formulation. Next, we propose a Gaussian MRF model and solve the energy minimization problem for sub-pixel DSA registration using Random Walker (RW). An incremental registration approach is developed using quad-tree based MRF structure and RW, wherein the density of control points is hierarchically increased at each level M depending of the features to be used and the required accuracy. A novel numbering scheme of the control points allows us to reuse the computations done at level M in M + 1. Both the models result in an accelerated performance without compromising on the artifact reduction. We have also provided a CUDA based design of the algorithm, and shown performance acceleration on a GPU. We have tested the approach using 25 clinical data sets, and have presented the results of quantitative analysis and clinical assessment. (ii) In External Beam Radiation Therapy (EBRT), in order to monitor the intra fraction motion of thoracic and abdominal tumors, the lung diaphragm apex can be used as an internal marker. However, tracking the position of the apex from image based observations is a challenging problem, as it undergoes both position and shape variation. We propose a novel approach for tracking the ipsilateral hemidiaphragm apex (IHDA) position on CBCT projection images. We model the diaphragm state as a spatiotemporal MRF, and obtain the trace of the apex by solving an energy minimization problem through graph-cuts. We have tested the approach using 15 clinical data sets and found that this approach outperforms the conventional full search method in terms of accuracy. We have provided a GPU based heterogeneous implementation of the algorithm using CUDA to increase the viability of the approach for clinical use. (iii) In an adaptive radiotherapy system, irrespective of the methods used for target observations there is an inherent latency in the beam control as they involve mechanical movement and processing delays. Hence predicting the target position during `beam on target' is essential to increase the control precision. We propose a novel prediction model (called o set sine model) for the breathing pattern. We use IHDA positions (from CBCT images) as measurements and an Unscented Kalman Filter (UKF) for state estimation. The results based on 15 clinical datasets show that, o set sine model outperforms the state of the art LCM model in terms of prediction accuracy.

Peripheral vascular grafts are used for the treatment of peripheral arterial disease and arteriovenous grafts for vascular access in end stage renal disease. The development of neo-intimal hyperplasia and thrombosis in the distal anastomosis remains the main reason for occlusion in that region. The local haemodynamics produced by a graft in the host vessel is believed to significantly affect endothelial function. Single spiral flow is a normal feature in medium and large sized vessels and it is induced by the anatomical structure and physiological function of the cardiovascular system. Grafts designed to generate a single spiral flow in the distal anastomosis have been introduced in clinical practice and are known as spiral grafts. In this work, spiral peripheral vascular and arteriovenous grafts were compared with conventional grafts using ultrasound and computational methods to identify their haemodynamic differences. Vascular-graft flow phantoms were developed to house the grafts in different surgical configurations. Mimicking components, with appropriate acoustic properties, were chosen to minimise ultrasound beam refraction and distortion. A dual-beam two-dimensional vector Doppler technique was developed to visualise and quantify vortical structures downstream of each graft outflow in the cross-flow direction. Vorticity mapping and measurements of circulation were acquired based on the vector Doppler data. The flow within the vascular-graft models was simulated with computed tomography based image-guided modelling for further understanding of secondary flow motions and comparison with the experimental results. The computational assessments provided a three-dimensional velocity field in the lumen of the models allowing a range of fluid dynamic parameters to be predicted. Single- or double-spiral flow patterns consisting of a dominant and a smaller vortex were detected in the outflow of the spiral grafts. A double- triple- or tetra-spiral flow pattern was found in the outflow of the conventional graft, depending on model configuration and Reynolds number. These multiple-spiral patterns were associated with increased flow stagnation, separation and instability, which are known to be detrimental for endothelial behaviour. Increased in-plane mixing and wall shear stress, which are considered atheroprotective in normal vessels, were found in the outflow of the spiral devices. The results from the experimental approach were in agreement with those from the computational approach. This study applied ultrasound and computational methods to vascular-graft phantoms in order to characterise the flow field induced by spiral and conventional peripheral vascular and arteriovenous grafts. The results suggest that spiral grafts are associated with advanced local haemodynamics that may protect endothelial function and thereby may prevent their outflow anastomosis from neo-intimal hyperplasia and thrombosis. Consequently this work supports the hypothesis that spiral grafts may decrease outflow stenosis and hence improve patency rates in patients.