MRI is a versatile tool for systematically assessing anatomical and functional changes in small animal models of human disease. Its noninvasive nature makes MRI an ideal candidate for longitudinal evaluation of disease progression in mice; however achieving the desired level of statistical power can be expensive in terms of imaging time. This is particularly true for cancer studies, where dynamic contrast-enhanced (DCE-) MRI, which involves the repeated acquisition of anatomical images before, during, and after the injection of a paramagnetic contrast agent, is used to monitor changes in tumor vasculature. A means of reducing the overall time required to scan multiple cohorts of animals in distinct experimental groups is therefore highly desirable. Multiple-mouse MRI, in which several animals are simultaneously scanned in a common MRI system, has been successfully used to improve study throughput. However, to best utilize the next generation of small-animal MRI systems that will be equipped with an increased number of receive channels, a paradigm shift from simultaneously scanning as many animals as possible to scanning a more manageable number, at a faster rate, must be considered. Given a small-animal MRI system with 16 available receive channels, the simulations described in this work explore the tradeoffs between the number of animals scanned at once and the number of array elements dedicated to each animal for maximizing throughput. An array system consisting of 15 receive and 5 transmit coils allows throughput-optimized acceleration of a DCE-MRI protocol by a combination of multi-animal and parallel imaging techniques. The array system was designed and fabricated for use on a 7.0-T / 30-cm MRI system, and tested for high-throughput imaging performance in phantoms. Results indicate that up to a nine-fold throughput improvement is possible without sacrificing image quality compared to standard single-animal imaging hardware. A DCE-MRI study throughput improvement of just over six times that achieved with conventional single-mouse imaging was realized. This system will lower the barriers for DCE-MRI in preclinical research and enable more thorough sampling of disease pathologies that progress rapidly over time. / text

Acquired data from dynamic contrast enhanced MRI measurements can be used to non-invasively assess tumour vascular characteristics through pharmacokinetic modelling. The modelling requires an arterial input function which is the concentration of contrast agent in the blood reaching the volume of interest as a function of time. The aim of this work is testing and optimizing a turboFLASH sequence to appraise its suitability for measuring the arterial input function by measuring phase. Contrast concentration measurements in a phantom were done with both phase and relaxivity techniques. The results were compared to simulations of the experiment conditions to compare the conformance. The results using the phase technique were promising, and the method was carried on to in-vivo testing. The in-vivo data displayed a large signal loss which motivated a new phantom experiment to examine the cause of this signal reduction. Dynamic measurements were made in a phantom with pulsatile flow to mimic a blood vessel with a somewhat modified turboFLASH sequence. The conclusions drawn from analyzing the data were used to further improve the sequence and this modified turboFLASH sequence was tested in an in-vivo experiment. The obtained concentration curve showed significant improvement and was deemed to be a good representation of the true blood concentration. The conclusion is that phase measurements can be recommended over relaxivity based measurements. This recommendation holds for using a slice selective saturation recovery turboFLASH sequence and measuring the arterial input function in the neck. Other areas of application need more thorough testing.

No description available.

Contrast-enhanced mammography (CEM) relies on visualizing the growth of new blood vessels (i.e. tumor angiogenesis) to provide sufficient materials for cell proliferation during the development of cancer. Since cancers will accumulate an injected contrast agent more than other tissues, it is possible to use one of several methods to enhance the area of lesions in the x-ray image and remove the contrast of normal tissue. Large area flat panel detectors may be used for CEM wherein the subtraction of two acquired images is used to create the resulting enhanced image. There exist several methods to acquire the images to be subtracted, which include temporal subtraction (pre- and post-contrast images) and dual-energy subtraction (low- and high-energy images), however these methods suffer from artifacts due to patient motion between image acquisitions. In this research the use of a multilayer flat panel detector is examined for CEM that is designed to acquire both (low- and high-energy) images simultaneously, thus avoiding motion artifacts in the resulting subtracted image. For comparison, a dual-energy technique prone to motion artifacts that uses a single-layer detector is also investigated. Both detectors are evaluated and optimized using amorphous selenium as the x-ray to charge conversion material, however the theoretical analysis could be extended to other conversion materials. Experimental results of single pixel prototypes of both multilayer and single-layer detectors are also discussed and compared to theoretical results. For a more comprehensive analysis, the motion artifacts present in dual-exposure techniques are modeled and the performance degradation due to motion artifacts is estimated. The effects of noise reduction techniques are also evaluated to determine potential image quality improvements in CEM images.

This thesis presents a novel set of image analysis tools developed for the purpose of assisting radiologists with the task of detecting and characterizing breast lesions in image data acquired using magnetic resonance imaging (MRI). MRI is increasingly being used in the clinical setting as an adjunct to x-ray mammography (which is, itself, the basis of breast cancer screening programs worldwide) and ultrasound. Of these imaging modalities, MRI has the highest sensitivity to invasive cancer and to multifocal disease. MRI is the most reliable method for assessing tumour size and extent compared to the gold standard histopathology. It also shows great promise for the improved screening of younger women (with denser, more radio opaque breasts) and, potentially, for women at high risk. Breast MRI presently has two major shortcomings. First, although its sensitivity is high its specificity is relatively poor; i.e. the method detects many false positives. Second, the method involves acquiring several high-resolution image volumes before, during and after the injection of a contrast agent. The large volume of data makes the task of interpretation by the radiologist both complex and time-consuming. These shortcomings have motivated the research and development of the computer-aided detection systems designed to improve the efficiency and accuracy of interpretation by the radiologist. Whilst such systems have helped to improve the sensitivity/specificity of interpretation, it is the premise of this thesis that further gains are possible through automated image analysis. However, the automated analysis of breast MRI presents several technical challenges. This thesis investigates several of these, noise filtering, parametric modelling of contrast enhancement, segmentation of suspicious tissue and quantitative characterisation and classification of suspicious lesions. In relation to noise filtering, a new denoising algorithm for dynamic contrast-enhanced (DCE-MRI) data is presented, called the Dynamic Non-Local Means (DNLM). The DCE-MR image data is inherently contaminated by Rician noise and, additionally, the limited acquisition time per volume and the use of fat-suppression diminishes the signal-to-noise ratio. The DNLM algorithm, specifically designed for the DCE-MRI, is able to attenuate this noise by exploiting the redundancy of the information between the different temporal volumes, while taking into account the contrast enhancement of the tissue. Empirical results show that the algorithm more effectively attenuates noise in the DCE-MRI data than any of the previously proposed algorithms. In relation to parametric modelling of contrast enhancement, a new empiric model of contrast enhancement has been developed that is parsimonious in form. The proposed model serves as the basis for the segmentation and feature extraction algorithms presented in the thesis. In contrast to pharmacokinetic models, the proposed model does not rely on measured parameters or constants relating to the type or density of the tissue. It also does not assume a particular relationship between the observed changes in signal intensity and the concentration of the contrast agent. Empirical results demonstrate that the proposed model fits real data better than either the Tofts or Brix models and equally as well as the more complicated Hayton model. In relation to the automatic segmentation of suspicious lesions, a novel method is presented, based on seeded region growing and merging, using criteria based on both the original image MR values and the fitted parameters of the proposed model of contrast enhancement. Empirical results demonstrate the efficacy of the method, both as a tool to assist the clinician with the task of locating suspicious tissue and for extracting quantitative features. Finally, in relation to the quantitative characterisation and classification of suspicious lesions, a novel classifier (i.e. a set of features together with a classification method) is presented. Features were extracted from noise-filtered and segmented-image volumes and were based both on well-known features and several new ones (principally, on the proposed model of contrast enhancement). Empirical results, based on routine clinical breast MRI data, show that the resulting classifier performs better than other such classifiers reported in the literature. Therefore, this thesis demonstrates that improvements in both sensitivity and specificity are possible through automated image analysis.

No description available.

No description available.

Background: Drugs targeting angiogenic pathway remain the mainstay of treatment for metastatic renal cell carcinoma (mRCC). Tyrosine Kinase Inhibitors (TKI) as Sunitinib, Pazopanib as single agents and humanised monoclonal antibody bevacizumab (Bev) in combination with Interferon- α2a (IFN) have established as the first-line therapy for mRCC. Despite improvements in treatment, there are multiple questions which remain unanswered. In the combination of Bev and IFN, the respective role of each drug and whether any additional anti-angiogenic activity is gained by adding IFN to Bev remains unknown. As the clinical benefit obtained with these cytostatic agents does not always correlate with the conventional response assessment techniques as RECIST, it is necessary to reconsider the methods by which we assess benefit from these therapies. In this thesis, I report three studies aiming to answer these questions. Methods: With the clinical trial reported here, I explore whether Bev induced changes in vascular parameters measured by Dynamic Contrast Enhanced MRI (DCE-MRI) is significantly enhanced by the addition of IFN. In a phase II, randomised, open labelled, multicentre trial, treatment naïve mRCC patients were randomised to receive Bev on its own or in combination with a low dose (3MU) or standard dose (9MU) IFN. DCE-MRI was used to assess the changes in vascularity with the primary endpoint being, changes in transfer coefficient (Ktrans) after six weeks of treatment. I also report two retrospective imaging-based studies, using contrast-enhanced CT scans, performed to improve the methodology of response assessment for these antiangiogenic therapeutics. Here I explore the use of a) combining changes in size and arterial phase contrast enhancement measured using CT scan and b) changes in CT texture as methods of therapeutic response assessment in mRCC patients treated with TKI. Results: With the phase 2 clinical trial, we faced significant difficulty in recruitment as a result of restrictions in access to treatment in NHS, other competing studies and restrictions proposed by the DCE-MRI inclusion criteria. With slow recruitment, an unplanned analysis was performed after 21 patients were recruited. Analysis of primary endpoint showed no trend in the difference between arms with no correlation found between change in Ktrans and addition of IFN to bevacizumab. Effect size analysis performed due to the small numbers recruited failed to show any significance in the observed difference in Ktrans. Change in Ktrans and Kep may identify a group of patients likely to have PFS > 6 months, but this observation needs to evaluation in a larger sample size. Measuring size and change in arterial phase enhancement retrospectively using CT, a new criterion "modified" Choi, which prerequisite a combination of a decrease in arterial phase density by 15% and a decrease in size by 10% for response was proposed. Response assessment was measured with RECIST, Choi and modified Choi individually in 20 evaluable patients retrospectively and clinical benefit compared with Kaplan-Meier statistics and Log-Rank test. Response assessment as defined by the modified Choi criteria successfully identified patients who received clinical benefit from the treatment. Time to progression (TTP) was 448 days for the partial response and 89 days for stable disease as per the new criteria which were statistically significant with a p-value of 0.002. The second retrospective analysis explored the textural changes in enhanced CT scan. Performed in collaboration with researchers from Brighton University who developed the software algorithm used to assess changes in entropy and uniformity, 87 metastases from 39 patients with mRCC were analysed at baseline and after two cycles of TKI treatment. Textural parameters and response assessment criteria were correlated with TTP. After two cycles of TKI, the decrease in tumour entropy was 3%-45%, and increase in uniformity was 5%-21%. At a threshold change of -2% with uniformity, on a coarse scale of 2.5, the textural change was able to separate responders from non-responders. With Kaplan-Meier analysis comparing all four criteria, the percentage change in uniformity was statistically more significant than for RECIST, Choi, and Modified Choi criteria. Cox regression analysis showed that texture uniformity was an independent predictor of time to progression. Discussion: With the studies reported here, I was able to demonstrate the importance of improving the methodology in assessment of therapeutic response to targeted anti-angiogenic therapy in metastatic renal cell carcinoma. Even though the clinical trial, terminated early due to slow recruitment, did not reach its primary endpoint, changes in other vascular parameters as Kep combined with changes Ktrans showed tendency towards identifying a group of patients who derived clinical benefit of >6months with these therapies. This is particularly exciting as given the vascular stabilisation effect proposed for bevacizumab, the effusion parameter Kep may be a better tool in assessing response rather than Ktrans and warrants further assessment in a larger cohort. Modified choi criterion and textural analysis are two important methodological improvements in response assessment of cytostatic anti-angiogenic therapy. In the analyses reported here, both techniques have shown superiority over RECIST in response assessment and differentiating mRCC patients who is likely to gain clinical benefit by TKI therapy. Validation of these criteria on a larger patient cohort is important. As these criterions are assessed on standard enhanced CT scans, incorporating these criteria, especially modified choi criterion, as part of standard CT assessment could be performed and will provide a real world validation. Retrospective assessment using larger cohort of patients from previous phase 3 trials or inclusion of these parameters prospectively in phase 3 trials would also help us in evaluating these modalities further.

La séquence d’IRM de perfusion pondérée T1 après injection de Gadolinium (Gd), appelée dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) fait partie du protocole d’IRM multiparamétrique (IRM-mp) réalisée pour le bilan d’extension du cancer prostatique (CaP). Le rationnel pour l’utilisation de cette séquence est d’une part le rôle capital de la néoangiogénèse dans le développement et la dissémination du CaP et d’autre part la possibilité d’imager l’angiogénèse in vivo et de façon non invasive. L’analyse quantitative nécessite un post-traitement multi-étapes complexe, dont le principe repose sur la modélisation pharmacocinétique (PC) de la biodistrubtion du Gd. Elle permet de calculer des paramètres PC reflétant la perméabilité capillaire et/ou la perfusion. Dans le CaP, ces paramètres PC ont montré leur potentiel pour évaluer l’agressivité tumorale, le pronostic, l’efficacité d’un traitement et/ou pour déterminer la dose efficace d’une nouvelle molécule anti-angiogéniques ou antivasculaires en cours de développement. Néanmoins, ils manquent de reproductibilité, notamment du fait des différentes techniques de quantifications utilisées par les logiciels de post-traitement.Nous avons développé au sein du laboratoire un outil de quantification capable de calculer une cartographie T1(0) à partir de la méthode des angles de bascule variables, nécessaire pour convertir les courbes du signal en courbe de concentration du Gd (Ct); de déterminer la fonction d’entrée artérielle (AIF – arterial input function) dans l’artère fémorale (Indivuduelle – Ind) ou lorsque cela n’était pas possible, d’utiliser une AIF issue de la littérature, telle que celle de Weinmann (W) ou de Fritz-Hansen (FH) ; et d’utiliser deux modèles PC, celui de Tofts et celui de Tofts modifié. Le logiciel a été validé sur des données simulées et sur une petite série clinique.Nous avons ensuite étudié l’impact du choix de la fonction d’entrée artériel sur les paramètres PC et notamment sur leur capacité à distinguer le CaP du tissu sain. 38 patients avec un CaP (>0,5cc) de la zone périphérique (ZP) ont été rétrospectivement inclus. Chaque patient avait bénéficié d’une IRM-mp sur laquelle deux régions d’intérêt (RI) : une tumorale et une bénigne ont été sélectionnées en utilisant une corrélation avec des cartes histo-morphométriques obtenues après prostatectomie radicale. En utilisant trois logiciels d’analyse quantitative différents, les valeurs moyennes de Ktrans (constante de transfert), ve (fraction du volume interstitiel) and vp (fraction du volume plasmatique) dans les RI ont été calculées avec trois AIF différentes (AIF Ind, AIF de W et AIF de FH). Ktrans était le paramètre PC qui permettait de mieux distinguer le CaP du tissu sain et ses valeurs étaient significativement supérieures dans le CaP, quelque soit l’AIF ou le logiciel. L’AIF de W donnait des aires sous les courbes ROC (Receiver Operating Characteristic) significativement plus grandes que l’AIF de FH (0.002≤p≤0.045) et plus grandes ou égales à l’AIF Ind (0.014≤p≤0.9). L’AIF Ind et de FH avaient des aires sous les courbes ROC comparables (0.34≤p≤0.81). Nous avons donc montré que les valeurs de Ktrans et sa capacité à distinguer CaP du tissu sain variaient significativement avec le choix de l’AIF et que les meilleures performances étaient obtenues avec l’AIF de W. / Dynamic contrast enhanced (DCE)-MRI is a T1 weighted sequence performed before, during and after a bolus injection of a contrast agent (CA). It is included in the multi-parametric prostate MRI (mp-MRI) protocol using to assess the extent of prostate cancer (PCa). The rationale for using DCE-MRI in PCa is that on one hand angiogenesis has been shown to play a central role in the PCa development and metastasis and on the other hand that DCE-MRI is a non invasive method able to depict this angiogenesis in vivo. The quantitative analysis of DCE-MRI data is a complex and multi-step process. The principle is to use a pharmacokinetic (PK) model reflecting the theoretical distribution of the CA in a tissue to extract PK parameters that describe the perfusion and capillary permeability. These parameters are of growing interest, especially in the field of oncology, for their use in assessing the aggressiveness, the prognosis and the efficacy of anti-angiogenic or anti-vascular treatments. The potential utility of these parameters is significant; however, the parameters often lack reproducibility, particularly between different quantitative analysis software programs.Firstly, we developed a quantitative analysis software solution using the variable flip angle method to estimate the T1 mapping which is needed to convert the signal-time curves to CA concentration-time curves; using three different arterial input functions (AIF): an individual AIF (Ind) measured manually in a large artery, and two literature population average AIFs of Weinmann (W) and of Fritz-Hansen (FH); and using two PK models (Tofts and modified Tofts). The robustness of the software programs was assessed on synthetic DCE-MRI data set and on a clinical DCE-MRI data set. Secondly, we assessed the impact of the AIF selection on the PK parameters to distinguish PCa from benign tissue. 38 patients with clinically important peripheral PCa (≥0.5cc) were retrospectively included. These patients underwent 1.5T multiparametric prostate MR with PCa and benign regions of interest (ROI) selected using a visual registration with morphometric reconstruction obtained from radical prostatectomy. Using three pharmacokinetic (PK) analysis software programs, the mean Ktrans, ve and vp of ROIs were computed using three AIFs: Ind-AIF, W-AIF and FH-AIF. The Ktrans provided higher area under the receiver operating characteristic curves (AUROCC) than ve and vp. The Ktrans was significantly higher in the PCa ROIs than in the benign ROIs. AUROCCs obtained with W-AIF were significantly higher than FH-AIF (0.002≤p≤0.045) and similar to or higher than Ind-AIF (0.014≤p≤0.9). Ind-AIF and FH-AIF provided similar AUROCC (0.34≤p≤0.81).We have then demonstrated that the selection of AIF can modify the capacity of the PK parameter Ktrans to distinguish PCa from benign tissue and that W-AIF yielded a similar or higher performance than Ind-AIF and a higher performance than FH-AIF.

Evolution of vulnerable carotid plaques are crucial reason for cerebral ischemic strokes and identifying them in the early stage can become very important in avoiding the risk of stroke. In order to improve the identification and quantification accuracy of infancy plaques better visualization techniques are needed. Improving the visualization and quantification of neovascularization in carotid plaque using contrast enhanced ultrasound imaging still remains a challenging task. In this thesis work, three optimization techniques are proposed, which showed an improvement in the sensitivity of contrast agents when compared to the conventional clinical settings and insonation strategies. They are as follows:1) Insonation at harmonic specific (2nd harmonic) resonance frequency instead of resonance frequency based on maximum energy absorption provides enhanced nonlinear contribution.2) At high frequency ultrasound imaging, shorter pulse length will provide improved harmonic signal content when compared to longer pulse lengths. Applying this concept to multi- pulse sequencing (Pulse Inversion and Cadence contrast pulse sequencing) resulted in increased magnitude of the remaining harmonic signal after pulse summations.3) Peak negative pressure optimization of Pulse Inversion and Cadence contrast pulse sequencing was showed to further enhance the nonlinear content of the backscattered signal from contrast microbubbles without increasing the safety limits, defined by the mechanical index.The results presented in this thesis are based on computational modeling (Bubblesim software) and as a future continuation we plan to verify the simulation results with vitro studies.