Stroke is one of the leading causes of death in the world, resulting mostly from the sudden rupture of atherosclerotic plaques. From a biomechanical view, plaque rupture can be considered as a mechanical failure caused by extremely high plaque stress. In this PhD project, we are aiming to predict 3D plaque stress based on in-vivo MRI by using fluid structure interaction (FSI) method, and provide information for plaque rupture risk assessment. Fluid structure interaction was implemented with ANSYS 11.0, followed by a parameter study on fibrous cap thickness and lipid core size with realistic carotid plaque geometry. Twenty patients with carotid plaques imaged by in-vivo MRI were provided in the project. A framework of reconstructing 3D plaque geometry from in-vivo multispectral MRI was designed. The followed reproducibility study on plaque geometry reconstruction procedure and its effect on plaque stress analysis filled the gap in the literature on imaging based plaque stress modeling. The results demonstrated that current MRI technology can provide sufficient information for plaque structure characterization; however stress analysis result is highly affected by MRI resolution and quality. The application of FSI stress analysis to 4 patients with different plaque burdens has showed that the whole procedure from plaque geometry reconstruction to FSI stress analysis was applicable. In the study, plaque geometries from three patients with recent transient ischemic attack were reconstructed by repairing ruptured fibrous cap. The well correlated relationship between local stress concentrations and plaque rupture sites indicated that extremely high plaque stress could be a factor responsible for plaque rupture. Based on the 20 reconstructed carotid plaques from two groups (symptomatic and asymptomatic), fully coupled fluid structure interaction was performed. It was found that there is a significant difference between symptomatic and asymptomatic patients in plaque stress levels, indicating plaque stress could be used as one of the factors for plaque vulnerability assessment. A corresponding plaque morphological feature study showed that plaque stress is significantly affected by fibrous cap thickness, lipid core size and fibrous cap surface irregularities (curvedness). A procedure was proposed for predicting plaque stress by using fibrous cap thickness and curvedness, which requires much less computational time, and has the potential for clinical routine application. The effects of residual stress on plaque stress analysis and arterial wall material property characterization by using in-vivo MRI data were also discussed for patient specific modeling. As the further development, histological study of plaque sample has been combined with conventional plaque stress analysis by assigning material properties to each computational element, based on the data from histological analysis. This method could bridge the gap between biochemistry and biomechanical study of atherosclerosis plaques. In conclusion, extreme stress distributions in the plaque region can be predicted by modern numerical methods, and used for plaque rupture risk assessment, which will be helpful in clinical practice. The combination of plaque MR imaging analysis, computational modelling, and clinical study/ validation would advance our understandings of plaque rupture, prediction of future rupture, and establish new procedures for patient diagnose, management, and treatment.

This thesis focuses on special arterial fluid mechanics techniques developed for patient-specific computer modeling of blood flow in cerebral arteries with aneurysm and stent. These techniques are used in conjunction with the core computational technique, which is the space–time version of the variational multiscale (VMS) method and is called “DST/SST-VMST.” The special techniques include using NURBS for the spatial representation of the surface over which the stent mesh is built, mesh generation techniques for both the finite- and zero-thickness representations of the stent, techniques for generating refined layers of mesh near the arterial and stent surfaces, and models for representing double stent. We compute the unsteady flow patterns in the aneurysm and investigate how those patterns are influenced by the presence of single and double stents. We also compare the flow patterns obtained with the finite- and zero-thickness representations of the stent.

Three-dimensional (3-D) medical imaging creates notable opportunities as input toward engineering analyses, whether for basic understanding of the normal function or patho-physiology of an organ, or for the simulation of virtual surgical procedures. These analyses most often require finite element (FE) models to be constructed from patient-specific 3-D medical images. However, creation of such models can be extremely labor-intensive; in addition, image processing and mesh generation are often operator-dependent, lack robustness and may be of suboptimal quality. Focusing on the human aorta, the goal of the present work is to create a fast and robust methodology for quadrilateral surface and hexahedral volume meshing from 3-D medical images with minimal user input. By making use of the segmentation capabilities of the 3-D gradient vector flow field combined with original ray-tracing and orientation control algorithms, we will demonstrate that it is possible to incrementally grow a structured quadrilateral surface mesh of the inner wall of the aorta. The process does not only require minimal input from the user, it is also robust and very fast compared to existing methods; it effectively combines segmentation and meshing into one single effort. After successfully testing the methodology and measuring the quality of the meshes produced by it from synthetic as well as real medical image datasets, we will make use of the surface mesh of the inner aortic wall to derive hexahedral meshes of the aortic wall thickness and of the fluid domain inside the aorta. We will finally outline a tentative approach to merge several structured meshes to process the main branches of the aorta.

Three-dimensional (3-D) medical imaging creates notable opportunities as input toward engineering analyses, whether for basic understanding of the normal function or patho-physiology of an organ, or for the simulation of virtual surgical procedures. These analyses most often require finite element (FE) models to be constructed from patient-specific 3-D medical images. However, creation of such models can be extremely labor-intensive; in addition, image processing and mesh generation are often operator-dependent, lack robustness and may be of suboptimal quality. Focusing on the human aorta, the goal of the present work is to create a fast and robust methodology for quadrilateral surface and hexahedral volume meshing from 3-D medical images with minimal user input. By making use of the segmentation capabilities of the 3-D gradient vector flow field combined with original ray-tracing and orientation control algorithms, we will demonstrate that it is possible to incrementally grow a structured quadrilateral surface mesh of the inner wall of the aorta. The process does not only require minimal input from the user, it is also robust and very fast compared to existing methods; it effectively combines segmentation and meshing into one single effort. After successfully testing the methodology and measuring the quality of the meshes produced by it from synthetic as well as real medical image datasets, we will make use of the surface mesh of the inner aortic wall to derive hexahedral meshes of the aortic wall thickness and of the fluid domain inside the aorta. We will finally outline a tentative approach to merge several structured meshes to process the main branches of the aorta.

Medical imaging technologies have allowed for in vivo evaluation of the human musculoskeletal system. With advances in both medical imaging and computing, patient-specific model development of anatomic structures is becoming a reality. Three-dimensional surface models are useful for patient-specific measurements and finite element studies. Orthopaedics is closely tied to engineering in the analysis of injury mechanisms, design of implantable medical devices, and potentially in the prediction of injury. However, a disconnection exists between medical imaging and orthopaedic analysis; whereby, the ability to generate three-dimensional models from an imaging dataset is difficult, which has restricted its application to large patient populations. We have compiled image processing, image segmentation, and surface generation tools in a single software package catered specifically to image-based orthopaedic analysis. We have also optimized an automated segmentation technique to allow for high-throughput bone segmentation and developed algorithms that help to automate the cumbersome process of mesh generation in finite element analysis. We apply these tools to evaluate graft placement in anterior cruciate ligament reconstruction in a multicenter study that aims to improve the patient outcomes of those that undergo this procedure.

Patient-specific multi-physics simulations have the potential to improve the diagnosis, treatment, and scientific inquiry of heart valve dynamics. It has been shown that the flow characteristics within the left ventricle are important to correctly capture the aortic and mitral valve motion and corresponding fluid dynamics, motivating the use of patient-specific imaging to describe the aortic and mitral valve geometries as well as the motion of the left ventricle (LV). The LV position can be captured at several time points in the cardiac cycle, such that its motion can be prescribed a priori as a Dirichlet boundary condition during a simulation. Valve leaflet motion, however, should be computed from soft-tissue models and incorporated using fully-coupled Fluid Structure Interaction (FSI) algorithms. While FSI simulations have in part or wholly been achieved by multiple groups, to date, no high-throughput models have been developed, which are needed for use in a clinical environment. This project seeks to enable patient-derived moving LV boundary conditions, and has been developed for use with a previously developed immersed boundary, fixed Cartesian grid FSI framework. One challenge in specifying LV motion from medical images stems from the low temporal resolution available. Typical imaging modalities contain only tens of images during the cardiac cycle to describe the change in position of the left ventricle. This temporal resolution is significantly lower than the time resolution needed to capture fluid dynamics of a highly deforming heart valve, and thus an approach to describe intermediate positions of the LV is necessary. Here, we propose a primarily Eulerian means of representing LV displacement. This is a natural extension, since an Eulerian framework is employed in the CFD model to describe the large displacement of the heart valve leaflets. This approach to using Eulerian interface representation is accomplished by applying “morphing” techniques commonly used in the field of computer graphics. For the approach developed in the current work, morphing is adapted to the unique characteristics of a Cartesian grid flow solver which presents challenges of adaptive mesh refinement, narrow band approach, parallel domain decomposition, and the need to supply a local surface velocity to the flow solver that describes both normal and tangential motion. This is accomplished by first generating a skeleton from the Eulerian interface representation, and deforming the skeleton between image frames to determine bulk displacement. After supplying bulk displacement, local displacement is determined using the Eulerian fields. The skeletons are also utilized to automate the simulation setup to track the locations upstream and downstream where the system inflow/outflow boundary conditions are to be applied, which in the current approach, are not limited to Cartesian domain boundaries.

La lombalgie est un problème de santé commun qui touche une grande partie de la population des pays industrialisés. Au cours des années, la modélisation numérique a été largement étudiée pour étudier la biomécanique du rachis lombaire pour aider fortement les cliniciens dans le diagnostic et les traitements de cette pathologie. Ce travail présente une méthodologie pour la modélisation éléments finis spécifique au patient prenant en compte à la fois la géométrie individualisée et les propriétés des matériaux des structures biologiques. Dans cette étude, le maillage est piloté par des connaissances des matériaux personnalisées qui sont extraites de l'imagerie médicale avancée. En outre, un logiciel convivial comprenant du traitement d'images, des maillages « material-driven » et de l'affectation des propriétés des matériaux, nommé C3M pour le «Computed Material-driven Mesh Model», a été développé pour générer efficacement des modèles FE spécifiques aux sujets à partir d'images médicales. Ce procédé est appliqué pour générer un modèle FE spécifique au patient du rachis lombaire à partir d'images issues par Résonance Magnétique (IRM) ou par tomodensitométrie 3D (CT). Cette approche ouvre une nouvelle perspective pour améliorer le processus de maillage à l'aide de connaissances du matériel dérivées d'images médicales. Le modèle proposé permet un assemblage précis et simple de vertèbres et des disques intervertébraux en tenant en compte à la fois la géométrie et les propriétés mécaniques des matériaux reflétant la spécificité du patient. / Low back pain is a common health problem which impacts a large part of the population in industrialized countries. Over the years, numerical modeling has been widely studied to investigate the biomechanics of lumbar spine for strongly assisting clinicians in diagnosis and treatments of this spinal pathology. In recent years, there has been a growing interest in researching and developing patient specific computer modeling which has proven its ability to provide great promises for developing realistic model of individual subject. However, still the specificity of these models is not fully described or is often limited to patient geometry. In fact, few models consider appropriate material properties derived from tissue characterization obtained from medical images. Furthermore, patient specific models can be obtained with geometry and mechanical properties derived from CT, but few from MRI which is well-suited for examining soft tissues. Therefore, development of the high-fidelity, patient-specific finite element model of the lumbar spine still presents the challenge. In this context of patient-specific finite element modeling, mesh generation is a crucial issue which requires an accurate representation of the geometry with well-shaped and sized elements and a relevant distribution of materials. This work presents a methodology for patient-specific finite element modeling which takes both individualized geometry and material properties of biological structures into consideration. In this study, the mesh is driven by personalized material knowledge which is extracted from advanced medical imaging. Additionally, a user-friendly program including image processing, material-driven meshing and material properties assignment, named C3M for “Computed Material-driven Mesh Model”, has been developed to generate efficiently subject-specific FE models derived from medical images. This process is applied to generate a patient specific FE model of lumbar spine based on both MRI and CT images. This approach opens a new direction to improve the meshing process using material knowledge derived from medical images. The proposed model allows an accurate and straightforward assembly of vertebrae and IVDs considering both geometry and material properties reflecting patient-specificity.

<p>Breast cancer affects 12.5\% of women over their life time and tissue expansion (TE) is the most common technique for breast reconstruction after mastectomy. However, the rate of complications with TE can be as high as 15\%. Even though the first documented case of TE happened in 1957, there has yet to be a standardized procedure established due to the variations among patients and the TE protocols are currently designed based on surgeon's experience. There are several studies of computational and theoretical framework modeling skin growth in TE but these tools are not used in the clinical setting. This dissertation focuses on bridging the gap between the already existing skin growth modeling efforts and it's potential application in the clinical setting.</p> <p><br></p> <p>We started with calibrating a skin growth model based on porcine skin expansions data. We built a predictive finite element model of tissue expansion. Two types of model were tested, isotropic and anisotropic models. Calibration was done in a probabilistic framework, allowing us to capture the inherent biological uncertainty of living tissue. We hypothesized that the skin growth rate was proportional to stretch. Indeed, the Bayesian calibration process confirmed that this conceptual model best explained the data. </p> <p><br></p> <p>Although the initial model described the macroscale response, it did not consider any activity on the cellular level. To account for the underlying cellular mechanisms at the microscopic scale, we have established a new system of differential equations that describe the dynamics of key mechanosensing pathways that we observed to be activated in the porcine model. We calibrated the parameters of the new model based on porcine skin data. The refined model is still able to reproduce the observed macroscale changes in tissue growth, but now based on mechanistic knowledge of the cell mechanobiology.  </p> <p><br></p> <p>Lastly, we demonstrated how our skin growth model can be used in a clinical setting. We created TE simulations matching the protocol used in human patients and compared the results with clinical data with good agreement. Then we established a personalized model built from 3D scans of a patient unique geometry. We verified our model by comparing the skin growth area with the area of the skin harvested in the procedure, again with good agreement.</p> <p><br></p> <p>Our work shows that skin growth modeling can be a powerful tool to aid surgeons design TE procedures before they are actually performed. The simulations can help with optimizing the protocol to guarantee the correct amount of skin is growth in the shortest time possible without subjecting the skin to deformations that can compromise the procedure.</p>

De nos jours, les progrès de l’informatisation de l’imagerie médicale assistent au plus près les médecins dans leur soin au patient. Des modèles personnalisés computationnels sont utilisés pour le diagnostique, prognostique et planification du traitement, en diminuant lesrisques pour le patient, et potentiellement les frais médicaux.Heartflow est l’exemple même d’une compagnie qui réussit ce service dans le domaine cardiovasculaire. À partir d’un modèle extrait d’images tomographiques rayons X, les lésions avec impact fonctionnel sont identifiées dans les artères coronaires. Cette analyse qui combine l’anatomie à la fonction est néanmoins limitée par la résolution de l’image. En aval de ces larges vaisseaux, un examen fonctionnel dénommé Imagerie de Perfusion du Myocarde (IPM) met en évidence les régions du myocarde affectées par un déficit de flux sanguin. Cependant, l’IPM n’établie pas de relation fonctionnelle avec les larges vaisseaux coronaires lésés en amont.L’objectif de ce projet est de construire la connexion fonctionnelle entre les coronaires et le myocarde, en extrapolant l’analyse fonctionnelle depuis les larges vaisseaux vers le lit capillaire. À cette fin, il faut étendre le modèle vasculaire jusqu'aux microvaisseaux, et mener une analyse fonctionnelle en direction du comportement myocardique.Nous étendons une méthode de génération d’arbre vasculaire basée sur la satisfaction de principes fonctionnels, nommée Constrained Constructive Optimization (Optimization Constructive sous Contraintes), pour qu’elle s’applique à de multiples arbres vasculaires en compétition. L’algorithme simule l’angiogénèse avec minimisation du volume vasculaire sous contraintes de flux et de géométrie adaptant la croissance simultanée des arbres aux caractéristiques du patient. Cette méthode fournit un modèle hybride composé de coronaires épicardiales extraites d’images et de vaisseaux synthétiques jusqu’aux artérioles, emplissant le ventricule gauche du myocarde.Puis, nous construisons un pipeline d’analyse fonctionnelle multi-échelle pour étendre la simulation de flux depuis les coronaires vers le myocarde. Cela consiste en un modèle de flux coronaire 1D compatible avec la vasculature hybride, et l’analyse de la distribution spatiale des flux provenant des segments terminaux. Cette dernière est réalisée dans une nomenclature similaire à celle de l’IPM pour permettre la comparaison avec des données de vérité terrain fonctionnelles.Nous avons relié l’anatomie du réseau vasculaire à la distribution de flux dans le myocarde pour plusieurs patients. Cette analyse multi-échelle permet d’identifier des pistes pour affiner les méthodes de génération vasculaire et de simulation de flux. Cette extrapolation anatomique et fonctionnelle personnalisée est une première passerelle pour la simulation de perfusion du myocarde à partir d’imagerie tomographique scanner. La construction d’un tel modèle computationnel personnalisé pourrait aider à la compréhension de la physio-pathologie cardiovasculaire complexe et, enfin, à la santé du patient. / Recent advances in medical image computing have allowed automatedsystems to closely assist physicians in patient therapy. Computationaland personalized patient models benefit diagnosis, prognosisand treatment planning, with a decreased risk for the patient,as well as potentially lower cost. HeartFlow Inc. is a successfull exampleof a company providing such a service in the cardiovascularcontext. Based on patient-specific vascular model extracted from XrayCT images, they identify functionally significant disease in largecoronary arteries. Their combined anatomical and functional analysisis nonetheless limited by the image resolution. At the downstreamscale, a functional exam called Myocardium Perfusion Imaging (MPI)highlights myocardium regions with blood flow deficit. However,MPI does not functionally relate perfusion to the upstream coronarydisease.The goal of our project is to build the functional bridge betweencoronary and myocardium, by extrapolating the functional analysisfrom large coronary toward the capillary bed. This objective requiresextension from the coronary model down to the microvasculaturecombined with a functional analysis leading to the myocardium compartment.We expand a tree generation method subjected to functional principles,named Constrained Constructive Optimization, to generate multiplecompeting vascular trees. The algorithm simulates angiogenesisunder vascular volume minimization with flow-related and geometricalconstraints, adapting the simultaneous tree growths to patientpriors. This method provides a hybrid image-based and synthetic geometricmodel, starting from segmented epicardium coronary downto synthetic arterioles, filling the left ventricle myocardium.We then build a multiscale functional analysis pipeline to allowblood flow simulation from the coronaries to the myocardium. Thisis achieved with a 1D coronary model compatible with the hybridvasculature, and a spatial blood flow distribution analysis of the terminalsegments. The latter is performed using a similar nomenclatureto MPI, to enable patient-specific comparison with functional groundtruthdata.We connected the vascular anatomy to blood flow distribution inthe myocardium on several patient datasets. This multiscale frameworkpoints out several leads to refine the vascular network generationand fluid simulation methods. This patient-specific anatomicaland functional extrapolation is a first gateway toward myocardiumperfusion from X-ray CT data. Building such personalized computational model of patient could potentially help investigating cardiovascularcomplex physio-pathology, and, finally, improve the patientcare.

Ces travaux de thèse s'inscrivent dans le contexte des gestes médico-chirurgicaux assistés par ordinateur pour le traitement endovasculaire de l'anévrisme de l'aorte abdominale. La complexité de ce type de procédure et l'évolution permanente des dispositifs endovasculaires posent des difficultés liées notamment à la sélection des endoprothèses les mieux adaptées à l'anatomie du patient et à la localisation précise des outils endovasculaires lors de l'intervention. L'objectif de ces travaux thèse est d'apporter aux cliniciens une aide à la décision lors de la planification et de la réalisation de la procédure. L'approche envisagée consiste à mettre en correspondance les examens d'imagerie préopératoire et peropératoire et à les combiner avec des modèles d'interaction dispositifs / tissus pour améliorer le positionnement des endoprothèses dans des structures vasculaires déformables. Nous envisageons tout d'abord une solution permettant de positionner interactivement des modèles d'endoprothèse dans la structure vasculaire préopératoire afin de vérifier l'adéquation des endoprothèses sélectionnées avec l'anatomie du patient. La méthode s'appuie sur un modèle géométrique ou mécanique approché de l'endoprothèse placée dans une structure vasculaire non déformable. Nous proposons ensuite une solution originale de simulation des interactions outil-tissu dans le but d'anticiper les déformations vasculaires provoquées par l'insertion des outils, relativement rigides, avant le déploiement des endoprothèses. Le guidage de l'intervention par navigation endovasculaire augmentée est abordé au travers du recalage 3D/2D. Une méthode polyvalente est proposée afin de mettre en correspondance le scanner ainsi que les modèles préopératoires avec les images peropératoires. L'évaluation de ces méthodes et leur application sur données patients ont permis de montrer la compatibilité de l'approche proposée avec le workflow clinique. / This thesis deals with computer-assisted surgery in the context of endovascular repair of abdominal aortic aneurysm. The complexity of this procedure and the ongoing development of endovascular devices pose challenges such as the selection of the most appropriate stent-grafts for the patient anatomy and the precise location of endovascular tools during surgery. The aim of this thesis is to provide clinicians with decision support for the planning and the performance of the procedure. The proposed approach consists in matching preoperative and intraoperative image data and in combining them with tool-tissue interaction models in order to improve the positioning of stent-grafts in deformable vascular structures. We consider a solution for positioning stent-grafts interactively in the preoperative vascular structure to verify the adequacy of the selected stent-grafts with the patient anatomy. The method is based on a geometrical or simplified mechanical model of the stent graft placed in a rigid vascular structure. We propose an original solution for simulating tool-tissue interactions in order to anticipate vascular deformations caused by the insertion of stiff tools before stent-graft deployment. Augmented reality guidance for endovascular interventions is addressed through 3D/2D registration. A versatile method is proposed for the matching of the CT-scan and preoperative models with intraoperative images. The evaluation of these methods and the results obtained on patient data have shown the compatibility of the proposed approach with the clinical workflow.