Search

Abnormal vascular structure has been identified as one of the major characteristics of tumours. In this thesis, we carry out quantitative analysis on different tumour vascular structures and research the relationship between vascular structure and its transportation efficiency. We first study segmentation methods to extract the binary vessel representations from microscope images. We found that local phase-hysteresis thresholding is able to segment vessel objects from noisy microscope images. We also study methods to extract the centre lines of segmented vessel objects, a process termed as skeletonization. We modified the conventional thinning method to regularize the extremely asymmetrical structure found in the segmented vessel objects. We found this method is capable to produce vessel skeletons with satisfactory accuracy. We have developed a software for 3D vessel structural analysis. This software is consisted of four major parts: image segmentation, vessel skeletonization, skeleton modification and structure quantification. This software has implemented local phase-hysteresis thresholding and structure regularization-thinning methods. A GUI was introduced to enable users to alter the skeleton structures based on their subjective judgements. Radius and inter branch length quantification can be conducted based on the segmentation and skeletonization results. The accuracy of segmentation, skeletonization and quantification methods have been tested on several synthesized data sets. The change of tumour vascular structure after drug treatment was then investigated. We proposed metrics to quantify tumour vascular geometry and statistically analysed the effect of tested drugs on normalizing tumour vascular structure. finally, we developed a spatio-temporal model to simulate the delivery of oxygen and 3-18 F-fluoro-1-(2-nitro-1-imidazolyl)-2-propanol (Fmiso), which is the hypoxia tracer that gives out PET signal in an Fmiso PET scanning. This model is based on compartmental models, but also considers the spatial diffusion of oxygen and Fmiso. We validated our model on in vitro spheroid data and simulated the oxygen and Fmiso distribution on the segmented vessel images. We contend that the tumour Fmiso distribution (as observed in Fmiso PET imaging) is caused by the abnormal tumour vascular structure which is further aroused from tumour angiogenesis process. We depicted a modelling framework to research the relationships between tumour angiogenesis, vessel structure and Fmiso distribution, which is going to be the focus of our future work.

Image registration methods underpin many analysis techniques in neuroimaging. They are essential in group studies when images of different individuals or different modalities need to be brought into a common reference frame. This thesis explores the potential of brain connectivity- driven alignment and develops surface registration techniques for magnetic resonance imaging (MRI), which is a noninvasive neuroimaging tool for probing function and structure of the human brain. The first part of this work develops a novel surface registration framework, based on free mesh deformations, which aligns cortical and subcortical surfaces by matching structural connectivity patterns derived using probabilistic tractography (diffusion-weighted MRI). Structural, i.e. white matter, connectivity is a good predictor of functional specialisation and structural connectivity-driven registration can therefore be expected to enhance the alignment of functionally homologous areas across subjects. The second part validates developed methods for cortical surfaces. Resting State Networks are used in an innovative way to delineate several functionally distinct regions, which were then used to quantify connectivity-driven registration performance by measuring the inter- subject overlap before and after registration. Consequently, the proposed method is assessed using an independent imaging modality and the results are compared to results from state-of-the-art cortical geometry-driven surface registration methods. A connectivity-driven registration pipeline is also developed for, and applied to, the surfaces of subcortical structures such as the thalamus. It is carefully validated on a set of artificial test examples and compared to another novel surface registration paradigm based on spherical wavelets. The proposed registration pipeline is then used to explore the differences in the alignment of two groups of subjects, healthy controls and Alzheimer's disease patients, to a common template. Finally, we propose how functional connectivity can be used instead of structural connectivity for driving registrations, as well as how the surface-based framework can be extended to a volumetric one. Apart from providing the benefits such as the improved functional alignment, we hope that the research conducted in this thesis will also represent the basis for the development of templates of structural and functional brain connectivity.

Stroke is the largest cause of severe adult complex disability, caused when the blood supply to the brain is interrupted, either by a clot or a burst blood vessel. It is characterised by deficiencies in movement and balance, changes in sensation, impaired motor control and muscle tone, and bone deformity. Clinically applied stroke management relies heavily on the observational opinion of healthcare workers. Despite the proven validity of a few clinical outcome measures, they remain subjective and inconsistent, and suffer from a lack of standardisation. Motion capture of the upper limb has also been used in specialised laboratories to obtain accurate and objective information, and monitor progress in rehabilitation. However, it is unsuitable in environments that are accessible to stroke patients (for example at patients’ homes or stroke clubs), due to the high cost, special set-up and calibration requirements. The aim of this research project was to validate and assess the sensitivity of a relatively low cost, wearable, compact and easy-to-use monitoring system, which uses inertial sensors in order to obtain detailed analysis of the forearm during simple functional exercises, typically used in rehabilitation. Forearm linear and rotational motion were characterised for certain movements on four healthy subjects and a stroke patient using a motion capture system. This provided accuracy and sensitivity specifications for the wearable monitoring system. With basic signal pre-processing, the wearable system was found to report reliably on acceleration, angular velocity and orientation, with varying degrees of confidence. Integration drift errors in the estimation of linear velocity were unresolved. These errors were not straightforward to eliminate due to the varying position of the sensor accelerometer relative to gravity over time. The cyclic nature of rehabilitation exercises was exploited to improve the reliability of velocity estimation with model-based Kalman filtering, and least squares optimisation techniques. Both signal processing methods resulted in an encouraging reduction of the integration drift in velocity. Improved sensor information could provide a visual display of the movement, or determine kinematic quantities relevant to the exercise performance. Hence, the system could potentially be used to objectively inform patients and physiotherapists about progress, increasing patient motivation and improving consistency in assessment and reporting of outcomes.

Magnetic resonance imaging plays an important yet underutilized role in determining the natural history and prognosis of oral carcinoma. Depth of tumour invasion is an emergent factor in the oral cancer literature. However, problems exist with the definition of cut-points suitable for inclusion in TNM staging criteria. Statistical methodology represents a possible explanation but is underexplored. In this work, a review of the depth of invasion literature is conducted with emphasis on statistical technique. As well, statistical simulation is used to explore the implications of the of the minimum p-value method. The results demonstrate that the use of continuous variable categorization and multiple testing is widespread, and contributes to cut-point variability and false-positive tests. Depth, as a predictor of OCLNM and survival, must be questioned. The volume of tumour invasion is a promising prognostic factor that has not been fully investigated in the oral carcinoma literature. In this work, the volume of tumour invasion is measured on MRI and compared to thickness and maximum diameter in its capacity to predict 2-year all-cause, disease-related and disease-free survival, as well as occult cervical lymph node metastasis prediction. As part of a comprehensive approach, morphometric factors are incorporated into multifactor predictive models using regression, artificial neural networks and recursive partitioning. It is evident that MRI-based volume is superior all other linear measurements for both occult cervical lymph node metastasis and survival prediction. Artificial neural networks wee superior to all other techniques for survival prediction. There is a case for a unified artificial neural networks model for survival prediction that uses volume, midline invasion and N-stage to determine prognosis. This model can be used to determine individualized probabilities of 2-year survival. The lateral extrinsic muscles of the tongue lie just beneath the surface of the lateral tongue, yet their invasion is a criterion for T4 classification using the TNM staging system. In this work, the Visible Human Female is used to conduct an anatomic study of the extrinsic muscles of the tongue. Linear measurement is used to quantify the distance from the surface mucosa to the most superficial muscle fibres of the styloglossus and genioglossus. Further, the lateral extrinsic muscles are poorly demonstrated on MRI. An anatomic atlas of the tongue is fused with MRI images of oral carcinoma to demonstrate lateral muscle invasion. The results demonstrate that the styloglossus and hyoglossus lie very close to the surface of the lateral tongue, in some cases passing within 1 mm of the surface mucosa. These extrinsic muscles are readily invaded by even small tumours of the lateral tongue. Strict application of the TNM T4a criteria leads to unnecessary upstaging as these carcinomas do not warrant the prognosis and aggressive treatment of Stage IV disease. Extrinsic muscle invasion should be removed as a T4a criterion for the oral cavity. A separate category, T4a (oral tongue) specifying invasion of the genioglossus is also recommended. This work presented in this thesis is an original contribution to the field of oral cavity cancer research and has determined that there is capacity for improvement in current efforts to determine the natural history and prognosis of oral cavity squamous cell carcinoma. This thesis is the first to examine the role of statistical methodology in oral carcinoma depth of invasion cut-point variability. Further, this work presents an original approach to the prediction of regional metastasis and survival using advanced multivariate modeling techniques. No other work explored MRI-measured volume using the substantial sample size gathered in this thesis. Finally, this work is the first to demonstrate that lateral extrinsic muscle invasion is an unnecessary component of the T4a (oral cavity) classification criteria and should be reconsidered.

This thesis describes the design a novel type of particle accelerator for charged particle therapy. The accelerator is called a non-scaling, Fixed Field Alternating Gradient (ns-FFAG) accelerator, and will accelerate both protons and carbon ions to energies required for clinical use. The work is undertaken as part of the PAMELA project. An existing design for a ns-FFAG is taken as a starting point and analysed in terms of its ability to suit the charged particle therapy application. It is found that this design is particularly sensitive to alignment errors and would be unable to accelerate protons and carbon ions at the proposed acceleration rate due to betatron resonance crossing phenomena. To overcome this issue, a new type of non-linear ns-FFAG is developed which avoids resonance crossing and meets the requirements provided by clinical considerations. Two accelerating rings are required, one for protons up to 250 MeV and fully stripped carbon ions to 68 MeV/u, the other to accelerate the carbon ions up to 400-430 MeV/u. Detailed studies are undertaken to show that this new type of accelerator is suitable for the application. An alignment accuracy of 50 micrometers will not have a detrimental effect on the beam and the dynamic aperture for most lattice configurations is found to be greater than 50 pi.mm.mrad normalised in both the horizontal and vertical plane. Verification of the simulation code used in the PAMELA lattice design is carried out using experimental results from EMMA, the world's first ns-FFAG for 10-20 MeV electrons built at Daresbury Laboratory, UK. Finally, it is shown that the described lattice can translate into realistic designs for the individual components of the accelerator. The integration of these components into the PAMELA facility is discussed.

Cardiovascular magnetic resonance imaging (MRI) is a well established technique in clinical cardiology. Different MRI sequences are routinely used to assess cardiac anatomy, function, viability and other parameters that aid diagnosing cardiac disease. Conversely, cardiac magnetic resonance spectroscopy (MRS), the only available method for a non-invasive study of human cardiac metabolism, has not evolved into a clinical tool yet. The combination of both techniques holds great potential to gain insight into the causality of cardiomyopathy diseases or other medical conditions with high cardiovascular risk profile, like diabetes or obesity and improve the clinical management of cardiac diseases. Nowadays, high field clinical MR systems have the great potential of improving the low spatial and temporal resolution and reproducibility of MRS. The aim of this thesis was to develop and implement a cardiac 1H-MRS method at 3 T that can be applied in clinical routine for the assessment of creatine and lipid levels in the human myocardium. The methodological developments to advance cardiac MRS are presented first. A robust 1H-MRS method comprising an optimized single-voxel technique, phased-array coil combination routine, optimized water suppression, breath-hold averaging and post-processing methods were developed. First, reproducibility and feasibility of the method were validated in vivo by acquiring 1H-MRS of the liver in almost one hundred healthy subjects. Subsequently, myocardial lipids levels were obtained in healthy volunteers by single breath-hold 1H-MRS triggered to mid-diastole, showing good reproducibility in an acquisition time less than 12 s. The good spectral resolution achieved using this method was demonstrated by the ability to differentiate for the first time two pools of myocardial lipids in spectra from the septum of patients with suspected myocardial lipid excess. Finally, creatine levels for healthy volunteers were investigated using multiple breath-hold acquisitions. Thus, this study shows the practicality and feasibility to incorporate this rapid cardiac 1H-MRS method into clinical studies of the human myocardium.

The effectiveness of radiotherapy treatment could be significantly improved if tumour cells could be rendered more sensitive to ionising radiation without altering the sensitivity of normal tissues. However, many of the key mechanisms that determine intrinsic tumour radiosensitivity are largely unknown. This thesis is concerned with the identification of novel determinants of tumour radiosensitivity. A siRNA screen of 200 genes involved in DNA damage repair was conducted using γH2AX foci post-irradiation as a marker of cell damage. This screen identified POLQ as a potential tumour-specific contributor to radioresistance. Subsequent investigations demonstrated that POLQ knockdown resulted in radiosensitisation of a panel of tumour cell lines, whilst having little or no effect on normal tissue cell lines. It was subsequently shown that POLQ depletion rendered tumour cells significantly more sensitive to several classes of cytotoxic agents. Following exposure to etoposide, it was found that tumour cells depleted of POLQ had reduced RAD51 foci formation, suggesting that POLQ is involved in homologous recombination. A homologous recombination assay was used to confirm that POLQ depletion does indeed result in reduced homologous recombination efficiency. These findings led to the investigation of the clinical significance of tumour overexpression of POLQ. The clinical outcomes of patients with early breast cancer were correlated with tumour expression levels of POLQ. It was found that POLQ overexpression was correlated with ER negative disease and high tumour grade, both of which are associated with poor clinical outcomes. POLQ overexpression was associated with extremely poor relapse free survival rates, independently of any other clinical or pathological feature. The mechanism that causes this adverse outcome may in part arise from resistance to adjuvant chemotherapy and radiotherapy treatment. These findings, combined with the limited normal tissue expression of POLQ, make it an appealing target for possible clinical exploitation.

A patient-specific model of the liver can supply accurate volume measurements for oncologists and lesion locations and liver visualisation for surgeons. Our work seeks to enable an automatic computational tool for liver quantification. To create this model, the liver shape must be segmented from 3D CT images. In doing so, we can quantify liver volume and restrict the region of interest to ease the task of tumour and vascular segmentation. The main objective of liver segmentation developed into a mission to fluently describe liver shape a priori in level-set methods. This thesis looks at the utility of an implicit shape representation based on the Poisson equation to describe highly variable shapes, with application to image segmentation. Our first contribution is analyses on four implicit shape representations based on the heat equation, the signed distance function, Poisson’s equation, and the logarithm of odds. For four separate shape case studies, we summarise the class of shapes through their shape representation using Principal Component Analysis (PCA). Each shape class is highly variable across a population, but have a characteristic structure. We quantitatively compare the implicit shape representations, within each class, by evaluating its compactness, and in the last case, also completeness. To the best of our knowledge, this study is novel in comparing several shape representations through a single dimension reduction method. Our second contribution is a hybrid region-based level set segmentation that simultaneously infers liver shape given the image data, integrates the Poisson-based shape function prior into the segmentation, and evolves the level set according to the image data. We test our algorithm on exemplary 2D liver axial slices. We compare results for each image to results from (a) level-set segmentation without a shape prior and (b) level-set segmentation with a shape prior based on the Signed Distance Transform (SDT). In both priors, shapes are projected from shape space through the sample population mean and its modes of variation (the minimum number of principal components to comprise at least 95% of the cumulative variance). We compare results on four individual cases using the Dice coefficient and the Hausdorff distance. This thesis introduces an implicit shape representation based on Poisson’s equation in the field of medical image segmentation, showing its influence on shape space summary and projection. We analyse the shape space for compactness, showing that it is more compact in each of our case studies by at least two-fold and as much as three-fold. For 3D liver shapes, we show that it is more complete than the other three implicit shape representations. We utilise its description efficiency for use in 2D liver image segmentation, implementing the first shape function prior based on the Poisson equation. We show a qualitative and quantitative improvement over segmentation results without any shape prior and comparable results to segmentation with a SDT shape prior.

In this thesis, we develop a motion correction method to overcome the degradation of image quality introduced by respiratory motion in positron emission tomography (PET), so that diagnostic performance for lung cancer can be improved. Lung cancer is currently the most common cause of cancer death both in the UK and in the world. PET/CT, which is a combination of PET and CT, providing clinicians with both functional and anatomical information, is routinely used as a non-invasive imaging technique to diagnose and stage lung cancer. However, since a PET scan normally takes 15-30 minutes, respiration is inevitable in data acquisition. As a result, thoracic PET images are substantially degraded by respiratory motion, not only by being blurred, but also by being inaccurately attenuation corrected due to the mismatch between PET and CT. If these challenges are not addressed, the diagnosis of lung cancer may be misled. The main contribution of this thesis is to propose a novel process for respiratory motion correction, in which non-attenuation corrected PET images (PET-NAC) are registered to a reference position for motion correction and then multiplied by a voxel-wise attenuation correction factor (ACF) image for attenuation correction. The ACF image is derived from a CT image which matches the reference position, so that no attenuation correction artefacts would occur. In experiments, the motion corrected PET images show significant improvements over the uncorrected images, which represent the acquisitions typical of current clinical practice. The enhanced image quality means that our method has the potential to improve diagnostic performance for lung cancer. We also develop an automatic lesion detection method based on motion corrected images. A small lung lesion is only 2 or 3 voxels in diameter and of marginal contrast. It could easily be missed by human observers. Our method aims to provide radiologists with a map of potential lesions for decision so that diagnostic efficiency can be improved. It utilises both PET and CT images. The CT image provides a lung mask, to which lesion detection is confined, whereas the PET image provides distribution of glucose metabolism, according to which lung lesions are detected. Experimental results show that respiratory motion correction significantly increases the success of lesion detection, especially for small lesions, and most of the lung lesions can be detected by our method. The method can serve as a useful computer-aided image analysing tool to help radiologists read images and find malignant lung lesions. Finally, we explore the possibility of incorporating temporal information into respiratory motion correction. Conventionally, respiratory gated PET images are individually registered to the reference position. Temporal continuity across the respiratory period is not considered. We propose a spatio-temporal registration algorithm, which models temporally smooth deformation in order to improve the registration performance. However, we discover that the improvement introduced by temporal information is relatively small at the cost of a much longer computation time. Spatial registration with regularisation yields similar results but is superior in speed. Therefore, it is preferable for respiratory motion correction.

Cavitation is a phenomenon present during many ultrasound therapies, including the thermal ablation of malignant tissue using high intensity focused ultrasound (HIFU). Inertial cavitation, in particular, has been previously shown to result in increased heat deposition and to be associated with broadband noise emissions that can be readily monitored using a passive receiver without interference from the main ultrasound signal. The present work demonstrates how an array of passive receivers can be used to generate maps of cavitation distribution during HIFU exposure, uncovering a new potential method of monitoring HIFU treatment. Using a commercially available ultrasound system (z.one, Zonare, USA), pulse transmission can be switched off and data from 64 elements of an array can be simultaneously acquired to generate passive maps of acoustic source power. For the present work, a 38 mm aperture 5-10 MHz linear array was used, with the 64 elements chosen to span the entire aperture. Theory and simulations were used to show the spatial resolution of the system, the latter showing that the broadband nature of inertial cavitation makes passive maps robust to interference between cavitating bubbles. Passive source mapping was first applied to wire scatterers, demonstrating the ability of the system to resolve broadband sources. With the array transversely placed to the HIFU axis, high-resolution passive maps are generated, and emissions from several cavitating bubbles are resolved. The sensitivity of passive mapping during HIFU exposure is compared with that of an active cavitation detector following exposure. The array was then placed within a rectangular opening in the centre of the HIFU transducer, providing a geometric setup that could be used clinically to monitor HIFU treatment. Cavitation was instigated in continuous and disjoint regions in agar tissue mimicking gel, with the expected regions of cavitation validating the passive maps obtained. Finally, passive maps were generated for samples of ox liver exposed to HIFU. The onset of inertial cavitation as detected by the passive mapping approach was found to provide a much more robust indicator of lesioning than post-exposure B-mode hyperecho, which is in current clinical use. Passive maps based on the broadband component of the received signal were able to localize the lesions both transversely and axially, however cavitation is generally indicated 5 mm prefocal to the lesions. Further work is needed to establish the source of this discrepancy. It is believed that with use of an appropriately designed cavitation detection array, passive mapping will represent a major advance in ultrasound-guided HIFU therapy. Not only can it be utilized in real-time during HIFU exposure, without the need to turn the therapeutic ultrasound field off, but it has also been shown in the context of the present work to provide a strong indicator of successful lesioning and high signal-to-noise compared to conventional B-mode ultrasound techniques.

Functional Magnetic Resonance Imaging (fMRI) is a noninvasive imaging method that reflects local changes in brain activity. FMRI group studies involves the analysis of the functional images acquired for each of a group of subjects under the same experimental conditions. We propose a spatial marked point-process model for the activation patterns of the subjects in a group study. Each pattern is described as the sum of individual centres of activation. The marked point-process that we propose allows the researcher to enforce repulsion between all pairs of centres of an individual subject that are within a specified minimum distance of each other. It also allows the researcher to enforce attraction between similarly-located centres from different subjects. This attraction helps to compensate for the misalignment of corresponding functional areas across subjects and is a novel method of addressing the problem of imperfect inter-subject registration of functional images. We use a Bayesian framework and choose prior distributions according to current understanding of brain activity. Simulation studies and exploratory studies of our reference dataset are used to fine-tune the prior distributions. We perform inference via Markov chain Monte Carlo. The fitted model gives a summary of the activation in terms of its location, height and size. We use this summary both to identify brain regions that were activated in response to the stimuli under study and to quantify the discrepancies between the activation maps of subjects. Applied to our reference dataset, our measure is successful in separating out those subjects with activation patterns that do not agree with the overall group pattern. In addition, our measure is sensitive to subjects with a large number of activation centres relative to the other subjects in the group. The activation summary given by our model makes it possible to pursue a range of inferential questions that cannot be addressed with ease by current model-based approaches.

One approach that has demonstrated success in the field of molecular imaging utilizes microparticles of iron oxide (MPIO) conjugated to specific antibodies and/or peptides to provide contrast effects on MRI in relation to the molecular expression of a specified target. The experimental aims of this thesis were 1) to investigate the ability of VCAM-1 and P-selectin targeted MPIO to detect the expression of VCAM-1 and P-selectin on the activated endothelium in-vitro and in-vivo in mouse models of renal and cerebral ischemia reperfusion injury, and 2) develop a novel contrast agent for imaging αvβ3-integrin expression in angiogenesis using RGD peptide conjugated MPIO (RGD-MPIO) in-vitro. MPIO (1.0 µm) were conjugated to monoclonal antibodies against VCAM-1 (VCAM-MPIO) or P-selectin (PSEL-MPIO). In vitro, MPIO bound in a dose-dependent manner to tumor necrosis factor (TNF)-alpha stimulated sEND-1 endothelial cells when conjugated to VCAM-1 (R² = 0.88, P<0.01) and P-selectin antibodies (R² = 0.93, P<0.01), reflecting molecular VCAM-1 and P-selectin mRNA and protein expression. Mice subjected to unilateral, transient (30 minutes) renal ischemia and subsequent reperfusion received intravenous VCAM-MPIO and PSEL-MPIO (4.5 mg iron/kg body weight). In ischemic kidneys, MR related contrast effects of VCAM-MPIO were 4-fold higher than unclamped kidneys (P<0.01) and 1.5-fold higher than clamped kidneys of PSEL-MPIO injected mice (P<0.05). VCAM-MPIO binding was less evident in IRI kidneys pre-treated with VCAM-1 antibody (P<0.001). VCAM-1 mRNA expression and VCAM-MPIO contrast volume were highly correlated (R² = 0.901, P<0.01), indicating that quantification of contrast volume reflected renal VCAM-1 transcription. In mice subjected to cerebral ischemia, contrast volume was 11-fold greater in animals injected with VCAM-MPIO versus control IgG-MPIO (P<0.05). Finally, S-nitroso-N-acetylpenicillamine (SNAP) stimulated HUVEC-C cells, which express αvβ3-integrin, showed 44-fold greater RGD-MPIO binding than unstimulated cells (P<0.001) and 4-fold greater RGD-MPIO binding than SNAP stimulated cells blocked with soluble RGD peptide (P<0.001) in-vitro. This thesis demonstrated that targeted MPIO exhibited contrast effects that defined and quantified the molecular expression of specific targets through the use of high-resolution MRI in in-vitro and in-vivo models of vascular inflammation.

Bayesian learning methods are the basis of many powerful analysis techniques in neuroimaging, permitting probabilistic inference on hierarchical, generative models of data. This thesis primarily develops Bayesian analysis techniques for magnetic resonance imaging (MRI), which is a noninvasive neuroimaging tool for probing function, perfusion, and structure in the human brain. The first part of this work fits nonlinear biophysical models to multimodal functional MRI data within a variational Bayes framework. Simultaneously-acquired multimodal data contains mixtures of different signals and therefore may have common noise sources, and a method for automatically modelling this correlation is developed. A Gaussian process prior is also used to allow spatial regularization while simultaneously applying informative priors on model parameters, restricting biophysically-interpretable parameters to reasonable values. The second part introduces a novel data fusion framework for multivariate data analysis which finds a joint decomposition of data across several modalities using a shared loading matrix. Each modality has its own generative model, including separate spatial maps, noise models and sparsity priors. This flexible approach can perform supervised learning by using target variables as a modality. By inferring the data decomposition and multivariate decoding simultaneously, the decoding targets indirectly influence the component shapes and help to preserve useful components. The same framework is used for unsupervised learning by placing independent component analysis (ICA) priors on the spatial maps. Linked ICA is a novel approach developed to jointly decompose multimodal data, and is applied to combined structural and diffusion images across groups of subjects. This allows some of the benefits of tensor ICA and spatially-concatenated ICA to be combined, and allows model comparison between different configurations. This joint decomposition framework is particularly flexible because of its separate generative models for each modality and could potentially improve modelling of functional MRI, magnetoencephalography, and other functional neuroimaging modalities.

Uncertainty in the sizing of breast cancer masses is a major issue in breast screening programs, as there is a tendency to severely underestimate the sizing of malignant masses, especially with ultrasound imaging as part of the standard triple assessment. Due to this issue about 20% of all surgically treated women have to undergo a second resection, therefore the aim of this thesis is to address this issue by developing novel image analysis methods. Ultrasound elasticity imaging has been proven to have a better ability to differentiate soft tissues compared to standard B-mode. Thus a novel segmentation algorithm is presented, employing elasticity imaging to improve the sizing of malignant breast masses in ultrasound. The main contributions of this work are the introduction of a novel filtering technique to significantly improve the quality of the B-mode image, the development of a segmentation algorithm and their application to an ongoing clinical trial. Due to the limitations of the employed ultrasound device, the development of a method to improve the contrast and signal to noise ratio of B-mode images was required. Thus, an autoregressive model based filter on the radio-frequency signal is presented which is able to reduce the misclassification error on a phantom by up to 90% compared to the employed device, achieving similar results to a state-of-the art ultrasound system. By combining the output of this filter with elasticity data into a region based segmentation framework, a computationally highly efficient segmentation algorithm using Graph-cuts is presented. This method is shown to successfully and reliably segment objects on which previous highly cited methods have failed. Employing this method on 18 cases from a clinical trial, it is shown that the mean absolute error is reduced by 2 mm, and the bias of the B-Mode sizing to underestimate the size was overcome. Furthermore, the ability to detect widespread DCIS is demonstrated.

The advent of hyperpolarized magnetic resonance (MR) has provided new potential for real-time visualization of in vivo metabolic processes. The aim of the work in this thesis was to use hyperpolarized substrates to study rapid metabolic processes occurring in the healthy and diseased rat heart. Initial work, described in Chapter 2, optimized the hyperpolarization process to reproducibly generate tracers. Chapter 3 describes use of hyperpolarized 1-13C-pyruvate to investigate in vivo flux through the regulatory enzyme pyruvate dehydrogenase (PDH). Cardiac PDH activity was altered in several physiological and pathological states, namely fasting, type 1 diabetes, and high-fat feeding, and in vivo flux through PDH was measured using hyperpolarized MR. These measurements correlated with measurements of in vitro PDH activity obtained using a validated biochemical assay. The work in Chapter 4 investigated the physiological interaction between hyperpolarized tracer and cardiac tissue. The effect of hyperpolarized 1-13C-pyruvate concentration on its in vivo metabolism was analyzed using modified Michaelis-Menten kinetics. It was found that hyperpolarized MR could non-invasively follow mechanisms of metabolic regulation, in addition to reporting enzyme activity. In Chapter 5, hyperpolarized MR was incorporated into the isolated perfused rat heart. 1-13C-pyruvate in normal and ischaemic hearts revealed significant differences in lactate metabolism, and provided the foundation for a novel intracellular pH probe. Infusion of 2-13C-pyruvate in the isolated rat heart enabled the first real-time visualization of Krebs cycle intermediates. In summary, the work in this thesis has highlighted the potential of hyperpolarized MR to reveal novel information on heart disease.

Nonrigid image registration has received much attention in the medical imaging and computer vision research communities, because it enables a wide variety of applications. Feature tracking, segmentation, classification, temporal image differencing, tumour growth estimation, and pharmacokinetic modeling are examples of the many tasks that are enhanced by the use of aligned imagery. Over the years, the medical imaging and computer vision communties have developed and refined image registration techniques in parallel, often based on similar assumptions or underlying paradigms. This thesis focuses on variational registration, which comprises a subset of nonrigid image registration. It is divided into chapters that are based on fundamental aspects of the variational registration problem: image dissimilarity measures, changing overlap regions, regularizers, and computational solution strategies. Key contributions include the development of local versions of standard dissimilarity measures, the handling of changing overlap regions in a manner that is insensitive to the amount of non-interesting background information, the combination of two standard taxonomies of regularizers, and the generalization of solution techniques based on Fourier methods and the Demons algorithm for use with many regularizers. To illustrate and validate the various contributions, two sets of example imagery are used: 3D CT, MR, and PET images of the brain as well as 3D CT images of lung cancer patients.

The estimation of probability density functions (PDF) of intensity values plays an important role in medical image analysis. Non-parametric PDF estimation methods have the advantage of generality in their application. The two most popular estimators in image analysis methods to perform the non-parametric PDF estimation task are the histogram and the kernel density estimator. But these popular estimators crucially need to be ‘tuned’ by setting a number of parameters and may be either computationally inefficient or need a large amount of training data. In this thesis, we critically analyse and further develop a recently proposed non-parametric PDF estimation method for signals, called the NP windows method. We propose three new algorithms to compute PDF estimates using the NP windows method. One of these algorithms, called the log-basis algorithm, provides an easier and faster way to compute the NP windows estimate, and allows us to compare the NP windows method with the two existing popular estimators. Results show that the NP windows method is fast and can estimate PDFs with a significantly smaller amount of training data. Moreover, it does not require any additional parameter settings. To demonstrate utility of the NP windows method in image analysis we consider its application to image segmentation. To do this, we first describe the distribution of intensity values in the image with a mixture of non-parametric distributions. We estimate these distributions using the NP windows method. We then use this novel mixture model to evolve curves with the well-known level set framework for image segmentation. We also take into account the partial volume effect that assumes importance in medical image analysis methods. In the final part of the thesis, we apply our non-parametric mixture model (NPMM) based level set segmentation framework to segment colorectal MR images. The segmentation of colorectal MR images is made challenging due to sparsity and ambiguity of features, presence of various artifacts, and complex anatomy of the region. We propose to use the monogenic signal (local energy, phase, and orientation) to overcome the first difficulty, and the NPMM to overcome the remaining two. Results are improved substantially on those that have been reported previously. We also present various ways to visualise clinically useful information obtained with our segmentations in a 3-dimensional manner.

High Intensity Focused Ultrasound (HIFU) is a new cancer thermal therapy method which has achieved encouraging results in clinics recently. However, the lack of a temperature monitoring makes it hard to apply widely, safely and efficiently. Conventional ultrasonic temperature estimation based on echo strain suffers from artifacts caused by signal distortion over time, leading to poor estimation and visualization of the 2D temperature map. This thesis presents a novel model-based stochastic framework for ultrasonic temperature estimation, which combines the temperature information from the ultrasound images and a theoretical model of the heat diffusion. Consequently the temperature estimation is more consistent over time and its visualisation is improved. There are 3 main contributions of this thesis related to: improving the conventional echo strain method to estimate temperature, developing and applying approximate heat models to model temperature, and finally combining the estimation and the models. First in the echo strain based temperature estimation, a robust displacement estimator is first introduced to remove displacement outliers caused by the signal distortion over time due to the thermo-acoustic lens effect. To transfer the echo strain to temperature more accurately, an experimental method is designed to model their relationship using polynomials. Experimental results on a gelatine phantom show that the accuracy of the temperature estimation is of the order of 0.1 ◦C. This is better than results reported previously of 0.5 ◦C in a rubber phantom. Second in the temperature modelling, heat models are derived approximately as Gaussian functions which are mathematically simple. Simulated results demonstrate that the approximate heat models are reasonable. The simulated temperature result is analytical and hence computed in much less than 1 second, while the conventional simulation of using finite element methods requires about 25 minutes under the same conditions. Finally, combining the estimation and the heat models is the main contribution of this thesis. A 2D spatial adaptive Kalman filter with the predictive step defined by the shape model from the heat models is applied to the temperature map estimated from ultrasound images. It is shown that use of the temperature shape model enables more reliable temperature estimation in the presence of distorted or blurred strain measurements which are typically found in practice. The experimental results on in-vitro bovine liver show that the visualisation on the temperature map over time is more consistent and the iso-temperature contours are clearly visualised.

Although death due to lethal cardiac arrhythmias is the leading cause of mortality in Western Society, many of the fundamental mechanisms underlying their onset, maintenance and termination, still remain poorly understood. In recent years, experimental techniques such as optical mapping have provided useful high-resolution recordings of cardiac electrical dynamics during complex arrhythmias and defibrillation episodes, which have been combined with detailed computer simulations to further our understanding of these phenomena. However, mechanistic enquiry is severely restricted as the optical mapping technique suffers from a number of distortion effects which compromise the fidelity of the experimental measurements, presenting difficulties in the comparison of experimental data with computational simulations. This Thesis presents a thorough investigation into the distortion effects encountered in optical mapping experiments, guided by the development of a coherent series of computational models. The models presented successfully characterise the specific mechanisms of fluorescent signal distortion due to photon scattering. Photon transport in cardiac tissue is modelled using both continuous (reaction-diffusion) and discrete stochastic (Monte Carlo) approaches to simulate the effects of photon scattering within the myocardium upon the recorded fluorescent signal, which include differing levels of detail and associated computational complexity. Specifically, these models are used to investigate the important role played by the complex ventricular structural anatomy, as well as the specifics of the experimental set-up itself. In addition, a tightly coupled electromechanical model of a contracting cardiac fibre is developed which provides an important first-step towards the development of a model to quantitatively assess the distortion observed when recording from a freely contracting cardiac preparation. Simulation of these distortion effects using the models allows discrimination to be made between those parts of the experimental signal which are due to underlying tissue electrophysiology and those due to artifact, facilitating a more accurate interpretation of experimentally-obtained data. The models presented succeed in two main respects. Firstly, they provide a ‘post-processing’ tool which can be added on to computational simulations of electrical activation, allowing for a more accurate and faithful comparison between simulations and experiments, helping to validate predictions made by electrical models. Secondly, they provide a higher degree of mechanistic insight into the fundemental ways in which optical signals are distorted, showing how this distortion can be maximised or controlled. The understanding and quantification of the fundemental mechanisms of optical mapping signal distortion, provided by this Thesis, therefore fulfils an important role in the study of arrhythmia mechanisms.

Introduction: Some past breast cancer radiotherapy regimens led to an increased risk of death from heart disease. Although heart dose from breast cancer radiotherapy has generally reduced over the past few decades, there may still be some cardiac risk. Estimation of future risk for women irradiated today requires both measurement of their cardiac dose and dose-response relationships, which depend on cardiac dosimetry of past regimens, in conjunction with long-term follow-up data. Methods: Virtual simulation and computed tomography 3-dimensional treatment planning on a representative patient were used to estimate mean heart and coronary artery doses for women irradiated since 1950 in 71 randomised trials in the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) overview. Patient-to-patient variability in cardiac dose was assessed. Heart and coronary artery doses were also calculated for breast cancer radiotherapy regimens used since the 1950s in Sweden. Cardiac doses from contemporary (year 2006) radiotherapy were assessed for 55 patients who received tangential breast cancer irradiation at a large UK radiotherapy centre. The maximum heart distance (i.e. the maximum distance between the anterior cardiac contour and the posterior tangential field edges) was measured for the left-sided patients, and its value as a predictor of cardiac doses assessed. Results: Mean heart dose for women irradiated in the EBCTCG trials varied from <1 to 18 Gray, and mean coronary artery dose from <1 to 57 Gray. Patient-to-patient variability was moderate. Mean heart dose for women irradiated in Sweden since the 1950s varied from <1 to 24 Gray, and mean coronary artery dose from <1 to 46 Gray. Heart dose from tangential irradiation has reduced over the past four decades. However, mean heart dose for left-sided patients irradiated in 2006 was 2 Gray and around half of them still received >20 Gray to parts of the heart and left anterior descending coronary artery. For these patients, maximum heart distance was a reliable predictor of cardiac doses. For the other patients, mean heart dose varied little and was usually less than 2 Gray. Conclusions: Cardiac doses from breast cancer radiotherapy can be estimated reliably and are now available for use in deriving dose-response relationships in the EBCTCG data and in a Scandinavian case-control study. Cardiac dose has reduced over the past four decades. Therefore the cardiac risk is also likely to have reduced. Nevertheless, for some patients, parts of the heart still receive >20 Gray in the year 2006.

This biomedical engineering thesis explores the opportunities and challenges of 2D+t contrast echocardiography for left ventricle functional analysis, both clinically and within a computer vision atlas-based deformable template model framework. A database was created for the experiments in this thesis, with 21 studies of contrast Dobutamine Stress Echo, in all 4 principal planes. The database includes clinical variables, human expert hand-traced myocardial contours and visual scoring. First the problem is studied from a clinical perspective. Quantification of endocardial global and local function using standard measures shows expected values and agreement with human expert visual scoring, but the results are less reliable for myocardial thickening. Next, the problem of segmenting the endocardium with a computer is posed in a standard landmark and atlas-based deformable template model framework. The underlying assumption is that these models can emulate human experts in terms of integrating previous knowledge about the anatomy and physiology with three sources of information from the image: texture, geometry and kinetics. Probabilistic atlases of contrast echocardiography are computed, while noting from histograms at selected anatomical locations that modelling texture with just mean intensity values may be too naive. Intensity analysis together with the clinical results above suggest that lack of external boundary definition may preclude this imaging technique for appropriate measuring of myocardial thickening, while endocardial boundary definition is appropriate for evaluation of wall motion. Geometry is presented in a Principal Component Analysis (PCA) context, highlighting issues about Gaussianity, the correlation and covariance matrices with respect to physiology, and analysing different measures of dimensionality. A popular extension of deformable models ---Active Appearance Models (AAMs)--- is then studied in depth. Contrary to common wisdom, it is contended that using a PCA texture space instead of a fixed atlas is detrimental to segmentation, and that PCA models are not convenient for texture modelling. To integrate kinetics, a novel spatio-temporal model of cardiac contours is proposed. The new explicit model does not require frame interpolation, and it is compared to previous implicit models in terms of approximation error when the shape vector changes from frame to frame or remains constant throughout the cardiac cycle. Finally, the 2D+t atlas-based deformable model segmentation problem is formulated and solved with a gradient descent approach. Experiments using the similarity transformation suggest that segmentation of the whole cardiac volume outperforms segmentation of individual frames. A relatively new approach ---the inverse compositional algorithm--- is shown to decrease running times of the classic Lucas-Kanade algorithm by a factor of 20 to 25, to values that are within real-time processing reach.

Our motivation is to develop an automated technique for the segmentation of sub-cortical human brain structures from MR images. To this purpose, models of shape-and-appearance are constructed and fit to new image data. The statistical models are trained from 317 manually labelled T1-weighted MR images. Shape is modelled using a surface-based point distribution model (PDM) such that the shape space is constrained to the linear combination of the mean shape and eigenvectors of the vertex coordinates. In addition, to model intensity at the structural boundary, intensities are sampled along the surface normal from the underlying image. We propose a novel Bayesian appearance model whereby the relationship between shape and intensity are modelled via the conditional distribution of intensity given shape. Our fully probabilistic approach eliminates the need for arbitrary weightings between shape and intensity as well as for tuning parameters that specify the relative contribution between the use of shape constraints and intensity information. Leave-one-out cross-validation is used to validate the model and fitting for 17 structures. The PDM for shape requires surface parameterizations of the volumetric, manual labels such that vertices retain a one-to-one correspondence across the training subjects. Surface parameterizations with correspondence are generated through the use of deformable models under constraints that embed the correspondence criterion within the deformation process. A novel force that favours equal-area triangles throughout the mesh is introduced. The force adds stability to the mesh such that minimal smoothing or within-surface motion is required. The use of the PDM for segmentation across a series of subjects results in a set surfaces that retain point correspondence. The correspondence facilitates landmark-based shape analysis. Amongst other metrics, vertex-wise multivariate statistics and discriminant analysis are used to investigate local and global size and shape differences between groups. The model is fit, and shape analysis is applied to two clinical datasets.