The individual soft tissues in the human body, such as liver, prostate, thyroid and breast, can each be characterized by a set of mechanical properties. Among these properties, the stiffness, or Young’s modulus, is of particular interest, as disease processes or abnormal growths introduce changes in the tissue stiffness. For example, cirrhosis is associated with an increase in stiffness in the affected region(s) of the liver, and the severity has a strong positive correlation with the measured liver tissue stiffness. Although the conventional ultrasound image is produced by changes in acoustic properties, most notably acoustic impedance (equal to density times sound speed), it is in fact possible to measure tissue strain ultrasonically, by performing ultrasound imaging while the tissue region of interest is mechanically perturbed. Although in principle incorrect, such strain imaging methods are commonly referred to as ultrasound elastography imaging. While tissue strain can reveal the presence of stiffness changes, its diagnostic value is limited due to the inability to reveal the magnitude of the stiffness change. Still, strain imaging is a feature on several commercial scanners. There does in fact exist an elegant, but complex and quite expensive, quantitative ultrasound method of imaging the elasticity of soft tissues, called Supersonic Shear imaging (SSI). However, a much lower cost method of quantitatively imaging tissue elasticity would be useful, especially if the method can be implemented with only minor modifications to existing ultrasound scanner design. This dissertation research deals with an attempt of designing and testing such a method. Ultrasound elastography encompasses a number of diverse techniques, roughly categorized by the mechanical perturbation method into two main groups: quasi-static and dynamic methods. Dynamic elastography requires a vibrating source, either separate or integrated with a transducer, making the imaging system cumbersome, especially for the portable systems. Quasi-static elastography only requires conventional ultrasound hardware, however current techniques remain qualitative with unknown stress distribution. This dissertation focuses on the investigation of free hand quantitative quasi-static elastography, aiming to real time assessment. Our proposed low cost real-time ultrasound elastography system is based on determining an axial strain and an axial stress over a region of interest, i.e., an axial strain image and an axial stress image are required. By taking the axial stress/axial strain ratio for each pixel in the image, an actual elasticity image is established. To achieve this goal, our system needs to ultrasonically measure the mechanical strain fast and accurately over a specified image plane; likewise, the system needs to be able to calculate the mechanical stress over the same image plane in real time. Now, the stress imaging will require us to apply a quasi-static force function and also to be able to quantify this force function. There are two major research efforts we have made to implement a low cost real-time ultrasound elastography system. The first important topic of this dissertation involves the development of a novel displacement and strain estimator based on analytical phase tracking (APT), which has been demonstrated to give better performance in terms of accuracy, resolution and computational efficiency (approximately 40 times faster than the standard time domain cross correlation method). The second important topic is the stress field reconstruction, with efforts in: 1) integrate force sensors into a single linear array transducer probe, with the goal of quantifying the applied force function; 2) propose a superposition method based on Love’s analytical equation to calculate the stress distribution, where this solution is computationally fast enough to allow real time stress field estimation; 3) analyze the accuracy of the proposed stress method using finite element analysis as a reference on different simulated phantoms. The final objective is to combine the strain and stress information together for quantitative elastography. Correspondingly, we have implemented experiments to evaluate the method on homogeneous and inhomogeneous phantoms of various types. Results show that this method is able to distinguish medium with different stiffness. We have conducted experiments to study the feasibility and improve the accuracy of this estimation technique based on phantoms with known elasticity. In principle, such a technique could be used to image the distribution of Young’s modulus under quasi-static compression, with specific applications to medical imaging.

In this study, we have first described the normal liver structure including the hepatic acinus that is characterized by its structural and functional heterogeneity. Second, we have addressed the pathogenesis of liver fibrosis: the major source of excess extracellular matrix appears to be perisinusoidal stellate cells. The concept of reversibility of liver fibrosis opens the way for new therapeutic perspectives. We have then analyzed the different methods of assessment of liver fibrosis. Liver biopsy is the current reference standard. However, it is invasive and subject to sampling error. Consequently, many attempts are made to develop non-invasive tests: biochemical tests and imaging methods, including functional MR imaging with perfusion, diffusion or spectroscopy, have been proposed. Among the imaging methods, elastography by measuring directly the liver stiffness appears as one of the most promising techniques. Lastly, we have described our research that was focused on MR elastography. Our results show that MR elastography is a feasible, accurate and reproducible method to stage liver fibrosis, and that it is superior to biochemical testing with aspartate-to-platelets ratio index and ultrasound elastography to stage liver fibrosis. Further studies remain to be done to decrease the long examination time of MR elastography and, consequently, to integrate it into a comprehensive hepatic MR protocol.

Elastography is a non-invasive imaging modality that uses ultrasound to estimate the elasticity of soft tissues. The resulting images are called 'elastograms'. Elastography techniques are promising as cost-effective tools in the early detection of pathological changes in soft tissues. The quality of elastographic images depends on the accuracy of the local displacement estimates. Cross-correlation based displacement estimators are precise and sensitive. However cross-correlation based techniques are computationally intense and may limit the use of elastography as a real-time diagnostic tool. This study investigates the use of parallel general purpose graphics processing unit (GPGPU) engines for speeding up generation of elastograms at real-time frame rates while preserving elastographic image quality. To achieve this goal, a cross-correlation based time-delay estimation algorithm was developed in C programming language and was profiled to locate performance blocks. The hotspots were addressed by employing software pipelining, read-ahead and eliminating redundant computations. The algorithm was then analyzed for parallelization on GPGPU and the stages that would map well to the GPGPU hardware were identified. By employing optimization principles for efficient memory access and efficient execution, a net improvement of 67x with respect to the original optimized C version of the estimator was achieved. For typical diagnostic depths of 3-4cm and elastographic processing parameters, this implementation can yield elastographic frame rates in the order of 50fps. It was also observed that all of the stages in elastography cannot be offloaded to the GPGPU for computation because some stages have sub-optimal memory access patterns. Additionally, data transfer from graphics card memory to system memory can be efficiently overlapped with concurrent CPU execution. Therefore a hybrid model of computation where computational load is optimally distributed between CPU and GPGPU was identified as an optimal approach to adequately tackle the speed-quality problem in real-time imaging. The results of this research suggest that use of GPGPU as a co-processor to CPU may allow generation of elastograms at real time frame rates without significant compromise in image quality, a scenario that could be very favorable in real-time clinical elastography.

In the study of the mechanical behavior of biological tissues, many complex tissues are often modeled as poroelastic systems due to their high fluid content and mobility. Fluid content and fluid transport mechanisms in tissues are known to be highly correlated with several pathologies. Thus, imaging techniques capable of providing accurate information about these mechanisms can potentially be of great diagnostic value. Ultrasound elastography is an imaging modality that is currently used as a complement to sonographic methods to detect a variety of tissue pathologies. Poroelastography is a new elastographic technique that has been recently proposed to image the mechanical behavior of tissues that can be modeled as poroelastic media. The few poroelastographic studies retrievable focus primarily on homogeneous poroelastic media. In this study, a statistical analysis of the performance of poroelastographic techniques in a non-homogeneous poroelastic simulation model under different loading conditions was carried out. The two loading conditions simulated were stress relaxation (application of constant strain) and creep compression (application of constant stress), both of which have been commonly used in the field of poroelastography. Simulations were performed using a FE poroelastic simulation software combined with ultrasound simulation software techniques and poroelastography processing algorithms developed in our laboratory. The non-homogeneous poroelastic medium was modeled as a cube (background) containing a cylindrical inclusion (target). Different permeability, Young’s modulus and Poisson’s ratio contrasts between the underlying matrix of the background and the target were considered. Both stress relaxation and creep compression loading conditions were simulated. The performance of poroelastography techniques was quantified in terms of accuracy, elastographic contrast–to–noise ratio and contrast transfer efficiency. The results of this study show that, in general, image quality of both axial strain and effective Poisson’s ratio poroelastograms is a complex function of time, which depends on the contrast between the poroelastic material properties of the background and the poroelastic material properties of the target and the boundary conditions. The results of this study could have important implications in defining the clinical range of applications of poroelastographic techniques and in the methodologies currently deployed.

A three-dimensional, incompressible, Rayleigh damped magnetic resonance elastography (MRE) material property reconstruction algorithm capable of reconstructing the spatial distribution of both the real and imaginary parts of the shear modulus, density and bulk modulus from full-field MR-detected harmonic motion data was developed. The algorithm uses a subzone-based implementation of motion error minimization techniques, using 27 hexahedral finite elements, and is written in FORTRAN to run on high performance distributed computing systems. The theory behind the methods used is presented in a form that is directly applicable to the code's structure, to serve as a reference for future research building on this algorithm. Globally defined Rayleigh damping parameter reconstructions using simulated data showed that it is possible to reconstruct the correct combination of Rayleigh parameters under noise levels comparable to MR measurements. The elastic wave equation is used to demonstrate that use of a one parameter damping model to fit a Rayleigh damped material can lead to artefacts in the reconstructed damping parameter images, a prediction that is verified using simulated reconstructions. Initial results using MR-detected motion data from both gelatine phantoms and in-vivo cases produced good reconstructions of real shear modulus, as well as showing promise for successful imaging of damping properties. An initial investigation into an alternative elemental basis function approach to supporting the material property distribution produced some promising results, as well as highlighting some significant issues with large variations across the elements.

Mudanças nas características mecânicas de tecidos biológicos geralmente estão relacionadas com algum tipo de patologia. Técnicas de imagens elastográficas são métodos quantitativos de se estimar as propriedades mecânicas de tecidos. Em geral, o objetivo destas técnicas de imagem é medir o movimento do tecido provocado por uma força interna ou externa. Por meio desse movimento, parâmetros viscoelásticos do meio em análise são reconstruídos. A força de excitação pode ser tanto quasi-estática, como dinâmica. O trabalho apresentado nesta tese aborda as técnicas de elastografia dinâmica e quasi-estática. Na abordagem quasi-estática, a elasticidade não-linear é estudada através de phantoms com características que simulam as do tecido humano. Na abordagem dinâmica, o movimento dinâmico promovido por força de radicação acústica é avaliado através de técnicas ultrassônicas e magnéticas. O desenvolvimento de materiais para serem usados como phantoms para elastografia por ultrassom é descrito. O comportamento elástico não-linear desses materiais foi analisado através de resultados de ensaios mecânicos. Esses materiais foram desenvolvidos para apresentarem uma relação tensão/deformação que não dependesse do módulo de cisalhamento para pequenas deformações, e foram projetados para serem usados em phantoms em que configurações heterogêneas são empregadas (por exemplo, phantoms com inclusões esféricas). O efeito da não-linearidade elástica dos materiais sobre o contraste, a relação sinal ruído e a relação contraste ruído de imagens elastográficas de um phantom contendo inclusões esféricas, sofrendo altas deformações (até 20%) foi investigada. Foi demonstrada a viabilidade de se medir movimentos vibratórios induzidos por feixes acústicos confocais através de um ultrassom Doppler que utiliza ondas contínuas. A interferência de feixes de ultrassom com pequena diferença de frequências provoca o aparecimento de uma força dinâmica no alvo. Foi demonstrada a formação de imagens de uma esfera rígida imersa em um phantom viscoelástico, através da varredura de ambos os transdutores (confocal e Doppler) pelo plano focal do transdutor confocal. O comportamento dinâmico de uma esfera magnetizada induzido por força de radicação acústica foi investigado. A esfera foi suspensa em água em configuração de pêndulo. Forcas estática de longa (poucos segundos) e curta (poucos milisegundos) duração foram utilizadas. O movimento da esfera foi medido através de um sensor magnetoresistivo. A partir da nova posição de equilíbrio em resposta à força de radicação de longa duração, a amplitude dessa força foi estimada. Para se estimar a viscosidade da água, o movimento de relaxação da esfera após a força ter sido desligada foi ajustado a um modelo de movimento-harmônico amortecido. O movimento de uma esfera rígida imersa em um phantom feito de gelatina, deslocada por força de radicação acústica, foi avaliado por meio de ecos ultrassônicas obtidos com um sistema pulso/eco. A teoria utilizada para se estimar os parâmetros viscoelásticos do phantom, usando o movimento induzido na esfera, é uma extensão da teoria usada para se estimar a viscosidade da água. / Changes in the mechanical properties of soft tissues may be related with pathological disorders. Elasticity imaging is a quantitative method of estimating the mechanical properties of the tissue. In general, the aim of this technique is to measure tissue motion caused by external or internal forces and use it to reconstruct the viscoelastic parameters of the medium. The excitation stress used can be (quasi-) static or dynamic. Both elastographic approaches are explored in this thesis work. In the quasi-static approach, the nonlinear elasticity is studied through tissue-mimicking phantom experiments. In the dynamic approach, the dynamic motion provided through acoustic radiation force is evaluated using ultrasonic and magnetic techniques. The development of phantom materials for elasticity imaging is reported. These materials were specifically designed to provide nonlinear stress/strain relationship that can be controlled independently of the small strain shear modulus of the material, and were designed for use in phantoms where heterogeneous configurations (e.g, spherical targets in a uniform background) are employed. The effects of phantom materials nonlinearity over the strain contrast, signal-to-noise ratio and contrast-to-noise ratio of a phantom containing spherical inclusions undergoing large deformations (up to 20%) were investigated.The feasibility of measuring vibration movement, through a mono-channel continuous wave Doppler system, induced by focused confocal beams, is demonstrated. The interference of two ultrasonic beams promotes a dynamic force to the target. The ability to form images of a rigid spherical inhomogeneity embedded in viscoelastic phantom by scanning both ultrasonic transducers (confocal and Doppler) across the confocal transducer focal plane is presented. The dynamic behavior of a rigid magnetic sphere induced by an acoustic radiation force was investigated. The sphere was suspended in water in a simple pendulum configuration. Steady forces of long (few seconds) and short (few milliseconds) durations were used. The movement of the magnetic sphere was tracked using a magnetoresistive sensor. From the new equilibrium position of the sphere in response to the long-duration static radiation force, the amplitude of this force was estimated. To access the water viscosity, the relaxation movement after the acoustic force had stopped was fitted to a harmonic-motion model. The motion of a rigid sphere embedded in gelatin phantom, displaced by acoustic radiation force, was evaluated using the ultrasonic echoes from a pulse-echo system. The theory used to estimate the viscoelastic parameters of the phantom, from the oscillation of the rigid sphere is an extension of the relation used to estimate the water viscosity.

Heart disease is the leading cause of death globally. Early diagnosis is the key to successful treatment. By providing noninvasive, non-ionizing, and real-time imaging, echocardiography plays a critical role in identifying heart disease. Compared to other imaging modalities, ultrasound has unparalleled temporal resolution. High frame-rate imaging has enabled the development of new metrics to characterize myocardial mechanics. Strain imaging measures the heart's deformation throughout the cardiac cycle, providing a quantitative assessment of cardiac health. The intention of this dissertation is to bring Myocardial Elastography (ME) closer to clinical realization. ME is a high frame-rate strain imaging technique for transthoracic and intracardiac echocardiography. This work consists of four Aims. There is a fundamental trade-off between spatial and temporal resolution in strain imaging. In Aim 1, the optimal transmit sequence that generates the most accurate and precise strain estimate was determined. Two common approaches to coherent compounding (full and partial aperture) were compared in simulation and in transthoracic imaging of healthy human subjects (n=5). The optimized subaperture compounding sequence (25-element subperture, 90° angular aperture, 10 virtual sources, 300 Hz frame rate) was compared to the optimized steered compounding sequence (60° angular aperture, 15° tilt, 10 virtual sources, 300 Hz frame rate) and was found to measure strain in healthy human subjects with equivalent precision. The optimal compounding configuration was then evaluated against two other high-frame rate transmit strategies, ECG-gated focused imaging, and wide-beam imaging, in simulation and in healthy subjects (n=7). Achieving the highest level of strain precision, ECG-gated focused imaging was determined to be the preferred imaging approach in patients capable of sustaining a breath hold, with compounding preferred in those unable to do so. Rapid diagnosis is essential to successful treatment of myocardial infarction. In Aim 2, ME's ability to track infarct formation and recovery, and localize infarct using regional strain measurments, was investigated in a large animal survival model (n=11). Infarcts were generated via ligation of the left anterior descending, imaging regularly for up to 28 days. A radial strain-based metric, percentage of healthy myocardium by strain (PHM_ε), was developed as a marker for healthy myocardial tissue. PHM_ε was strongly linearly correlated with actual infarct size as determined by gross pathology (R2 = 0.80). ME was capable of diagnosing individual myocardial segments as non-infarcted or infarcted with high sensitivity (82%), specificity (92%), and precision (85%) (ROC AUC = 0.90), and tracked infarct recovery from collateral reperfusion through time. Noninvasive strain imaging at rest can improve pre-test probability accuracy, and reduce unnecessary stress testing. In Aim 3, ME's potential to provide early diagnosis of coronary artery disease was investigated in an ongoing study. Patients undergoing myocardial perfusion imaging were recruited (n=126). Perfusion scores were used as the reference standard. Morphological transformations were integrated into the processing pipeline to reduce variability in the strain measurements. PHM_ε was reintroduced and used to differentiate between patients with and without coronary artery disease. ME was capable of distinguishing between normal patients and those with significant ischemia or infarct (subjects with perfusion defects at rest) with statistical significance (p < 0.05), although a greater sample size is needed to confirm the results. One of the most common treatments for arrhythmia, catheter ablation, can fail if the lesion line intended to terminate the abnormal rhythm is non-contiguous. In Aim 4, the gap resolution and clinical feasibility of Intracardiac Myocardial Elastography (IME) strain imaging, an ablation monitoring technique, was investigated. Lesion size estimation and gap resolution was evaluated in an open chest canine model (n=3), wherein lesion lines consisting of three lesions and two gaps were generated in each canine left ventricle via epicardial ablation. All gaps were resolvable. Average lesion and gap areas were measured with high agreement (33 ± 14 mm2 and 30 ± 15 mm2, respectively) when compared against gross pathology (34 ± 19 mm2 and 26 ± 11 mm2, respectively). Gaps as small as 11 mm2 (3.6 mm on epicardial surface) were identifiable. Patients undergoing ablation to treat typical cavotricuspid isthmus atrial flutter (n=5) were imaged throughout the procedure. In all patients, strain decreased in the cavotricuspid isthmus after ablation (mean paired difference of -17 ± 11 %, p < 0.05). Together, these Aims intend to translate a promising imaging method from research to clinical reality.

The non-invasive differential diagnosis of breast masses through ultrasound imaging motivates the following class of elastic inverse problems: Given one or more measurements of the displacement field within an elastic material, determine the material property distribution within the material. This thesis is focused on uncertainty quantification in inverse problem solutions, with application to inverse problems in linear and nonlinear elasticity. We consider the inverse nonlinear elasticity problem in the context of Bayesian statistics. We show the well-known result that computing the Maximum A Posteriori (MAP) estimate is consistent with previous optimization formulations of the inverse elasticity problem. We show further that certainty in this estimate may be quantified using concepts from information theory, specifically, information gain as measured by the Kullback-Leibler (K-L) divergence and mutual information. A particular challenge in this context is the computational expense associated with computing these quantities. A key contribution of this work is a novel approach that exploits the mathematical structure of the inverse problem and properties of conjugate gradient method to make these calculations feasible. A focus of this work is estimating the spatial distribution of the elastic nonlinearity of a material. Measurement sensitivity to the nonlinearity is much higher for large (finite) strains than for smaller strains, and so large strains tend to be used for such measurements. Measurements of larger deformations, however, tend to show greater levels of noise. A key finding of this work is that, when identifying nonlinear elastic properties, information gain can be used to characterize a trade-off between larger strains with higher noise levels and smaller strains with lower noise levels. These results can be used to inform experimental design. An approach often used to estimate both linear and nonlinear elastic property distributions is to do so sequentially: Use a small strain deformation to estimate the linear properties, and a large strain deformation to estimate the nonlinearity. A key finding of this work is that accurate characterization of the joint posterior probability distribution over both linear and nonlinear elastic parameters requires that the estimates be performed jointly rather than sequentially. All the methods described above are demonstrated in applications to problems in elasticity for both simulated data as well as clinically measured data (obtained in vivo). In the context of the clinical data, we evaluate repeatability of measurements and parameter reconstructions in a clinical setting.

Este trabalho apresenta uma análise sobre um novo método de formação de imagem, utilizando aparelhos de ultra-som a elastografia. Esta técnica baseia-se no fato de que quando um meio elástico, como o tecido, é deformado por uma tensão constante e uniaxial, todos os pontos no meio possuem um nível de deformação longitudinal cujo componente principal está ao longo do eixo de deformação. Se elementos do tecido possuem um módulo elástico diferente dos demais, a deformação nestes elementos será relativamente maior ou menor. Elementos mais rígidos geralmente deformam-se menos. Desta forma, pode-se mapear e identificar estruturas com diferentes níveis de dureza. A comparação entre os mapas de RF de pré e pós-deformação foi realizada pela técnica de block matching. Esta técnica consiste em comparar regiões, ou kernels, no mapa de pré-deformação com regiões de mesmo tamanho no mapa de pós-deformação. Esta comparação é feita pela minimização de uma função custo. Nesta técnica, o tamanho do kernel, é um dos principais parâmetros para melhorar a precisão das medidas de deslocamento. O principal objetivo neste trabalho é aperfeiçoar o algoritmo de block matching visando melhorar a precisão da determinação de deslocamento em técnicas de deformação dinâmica e estática, mantendo o custo computacional baixo. Para isto, foram utilizados phantoms com e sem inclusões mais duras que o meio. Os phantoms foram submetidos a deformações estáticas e dinâmicas. Foi possível determinar o comportamento destes phantoms sob estas formas de deformação, e as faixas de kernel e funções custo que forneceram os melhores resultados. Também foram gerados elastogramas do phantom com inclusão. Estas imagens permitiram avaliar a influência dos diferentes kernels sobre a resolução dos elastogramas e a capacidade em diferenciar a lesão do tecido circundante. Comparando os elastogramas obtidos sobre deformação dinâmica, utilizando os kernels que apresentaram o melhor desempenho, com as respectivas imagens em modo B, pôde-se observar que a inclusão estava clara e bem delimitada. / This work provides an analysis about a new method for image formation using ultrasound devices elastography. This technique is based on the fact that when an elastic medium, as the tissue, is deformed under a constant and directional stress, all the points in the medium have a deformation level whose main component is along the deformation axis. If tissues elements have different elastic modules, the deformation in these elements will be higher or lower. Normally harder elements have lower deformations. In this way, one can detect and identify structures with different elastic levels. The comparisons between the pre and post-deformation RF maps were done by the block matching technique. This technique is based on the comparison of regions, or kernels, in the pre-deformation maps with regions of the same size in the post-deformation map. This is done by the minimization of a cost function. In this technique, the kernel size is one of the most important parameters to obtain better resolution and precision in the displacement measurements. The goal of this work is to optimize the block matching algorithm to improve the displacement estimates precision in both dynamic and static deformations, while keeping a low computational cost. To obtain this, we used phantoms with and without inclusions harder than the medium. These phantoms were submitted to both static and dynamic deformations. It was possible to estimate the behavior of these phantoms under these deformations, and the kernel range and cost functions that provided the best results. Also, we generated the elastograms of the phantom with the inclusion. These images allowed us to evaluate the influence of the different kernel sizes under the elastograms resolution and their capability in differentiate the lesion from the embedding tissue. Comparing the elastograms obtained under dynamic deformation that had the best performance, with the B mode images, we could conclude that the inclusion was well delimited and clear.

Breast cancer is the most common cancer in women worldwide. In 2009, a novel imaging modality called Shear Wave Elastography (SWE), an ultrasound technique visualising the elasticity of tissue, was introduced to the field of clinical breast imaging. Because malignant tissues are generally stiffer than benign tissues, SWE supports the differentiation of benign / malignant solid breast lesions. However, no standard has yet been defined for the application and the evaluation of results. Furthermore, image evaluation has to be carried out directly from the ultrasound system, complicating long-term and multi-centre studies. This PhD thesis investigated the influences from the imaging process and image evaluation on SWE measurements. Various parameters were appraised with regard to their diagnostic performance, in order to define the best clinical standard. To define more complex image analysis, taking the parameters investigated into account, algorithms were devised to enable automatic assessment of B-mode and SWE images. In this work, influences from the imaging process and image evaluation on the SWE measurements were demonstrated. The influences investigated included: the impact from the region of interest and the imaging plane used; the individual variation in breast composition; the number of images considered and the pressure applied during imaging. The algorithms described within this work achieved a diagnostic accuracy similar to that of manual assessment by a radiology expert. This thesis demonstrated influences from the imaging process and image evaluation on the SWE measurements obtained. Taking these influences into consideration would complicate the clinical application of SWE imaging. However, automatic image evaluation as presented here would overcome this issue. Using the guidelines defined in this PhD thesis also allows for comparison of results taken from different imaging sites.