Esclerose Múltipla (MS) é uma doença neurodegenerativa que tem ganhado grande atenção nas últimas décadas, sendo o diagnóstico por imagens de ressonância magnética (MRI) um grande aliado para a avaliação da doença.Porém, um dos principais desafios é garantir uma maior sensibilidade e especificidade para detecção de diferentes lesões no sistema nervoso central (CNS) e assim classificar as diferentes variantes da MS, auxiliando na tomada de decisão para o tratamento farmacológico. Nas últimas décadas, a técnica de imagens ponderadas por difusão (DWI), em especial a técnica de imagem por tensor de difusão (DTI), têm sido evidenciada com grande potencial para o estudo da MS, apresentando uma melhora significativa para a detecção de lesões sutis, ainda em estágios iniciais da MS. Desta forma, as técnicas de processamento de imagens estão em constante aprimoramento para que sejam adaptados às novas modalidades de aquisição de imagens. Neste estudo focamos o desenvolvimento de uma técnica de processamento digital de imagens multimodais a fim de proporcionar uma solução viável para a rotina de diagnóstico por imagem em MS. Um conjunto de 25 pacientes de uma variante de MS foi selecionado aleatoriamente do banco de imagens do HCFMRP. Três modalidades de imagens foram coletadas para a avaliação da segmentação automatica (T1, T2-FLAIR e DTI), assim como a segmentação manual do especialista para cada paciente. Três métodos de segmentação multimodal automática de lesões foram analisados (Bayesiano, Frequentista e Agrupamento) afim de analisar a sensibilidade e especificidade de detecção de lesões na substância branca aparentemente normal (NAWM). Os resultados sugerem que o método de segmentação Bayesiano apresenta maior robustes e precisão na definição tanto de lesões visivelmente contrastantes em T1 e T2-FLAIR (i.e. lesões hipo e hiperintensas) assim como lesões da NAWM evidentes nos mapas quantitativos de DTI (FA e ADC). O erro associado à técninca automática de segmentação ficou em torno de 1.51 +- 0.51 % do volume total de lesões marcadas pelo especialista. Concluímos que o uso de ferramentas multimodais de segmentação de imagens MRI alcançou patamares razoáveis de detecção de lesões de MS, tornando assim uma ferramenta computacional hábil para uso no diagnóstico clínico. / Multiple sclerosis (MS) is a neurodegenerative disease that has gained great attention in the last decades, which magnetic resonance imaging (MRI) have shown as an important tool for the disease evaluation. However, one of the main challenges is guaranteeing greater lesion detection sensitivity and specificity in the whole central nervous system (CNS) and thus classify the different variants of MS, which aids in decision making for pharmacological treatment. In the last decades, the diffusion-weighted imaging (DWI) technique, especially the diffusion tensor imaging approach (DTI), has been evidenced with great potential for the study of MS, presenting a significant improvement for the detection of lesions even in early stages of MS. Hence, the techniques of image processing are constantly improving in order to be adapted on a multimodal image evaluation. In this study, the development of a multimodal digital image processing technique to provide a viable solution to the MS imaging routine was focused. A set of 25 patients from a MS variant was randomly selected from the HCFMRP imaging database. Three MR imaging modalities were collected for the evaluation of our automatic segmentation (T1, T2-FLAIR and DTI), as well as manual segmentation of the specialist for each patient. Three methods of automatic multimodal segmentation of MS lesions were analyzed (Bayesian, Frequentist and Clustering) in order to analyze the sensitivity and specificity of lesion detection in the apparently normal white matter (NAWM). The results suggest that the Bayesian segmentation method presented greater robustness and precision in the definition of visibly contrasting lesions in T1 and T2-FLAIR (i.e. hypo and hyperintense lesions) as well as NAWM lesions that are evident in quantitative DTI (FA and ADC). The error associated with the automatic segmentation technique was around 1.51 +- 0.51 % of the total lesion volume being evaluated by the a specialist. We conclude that the use of multimodal MRI images can be used in an automatic segmentation tools, reaching reasonable levels of MS lesion detection, thus making it a useful tool for clinical diagnosis.

Neste trabalho apresentamos e comparamos diferentes técnicas de tomografia bidimensional por RMN implementados num sistema de tomografia de campo magnético ultra-baixo (O.05T). A partir da seqüência convencional \"Spin Echo\" (SE), utilizada rotineiramente, implementamos a seqüência \"Gradient Recalled Echo\" (GRE) e duas seqüências que utilizam o principio de Steady-State Free Precession (SSFP), sendo \"Fast Low Angle Shot\" (FLASH) e \"Fast Acquisition Double Echo\" (FADE). Com as seqüências de SSFP conseguimos diminuir drasticamente a duração de um exame de tomografia convencional (seqüência SE). A seqüência FADE ainda permite adquirir duas imagens com contrastes claramente diferentes sem aumentar significativamente a duração do exame. Desenvolvemos procedimentos de calibração indispensáveis para as técnicas de SSFP que também melhoraram a relação sinal ruído de 15 por cento na técnica SE . Analisamos teórica e experimentalmente o comportamento do contraste das seqüências apresentadas. Fizemos ainda uma serie de imagens de um phantom e da cabeça de um voluntário com as diferentes seqüências e sugerimos algumas combinações dos parâmetros (protocolos) como tempo de repetição, tempo ao eco e angulo de excitação. Um destes protocolos esta sendo testado em casos clínicos para comparar a utilidade das seqüências apresentadas no diagnóstico medico. Com este trabalho ganhamos uma experiência na utilização de técnicas rápidas indispensáveis para a elaboração de uma metodologia para obter imagens tridimensionais / In this work we present and compare different techniques for bidimensional tomography in NMR which were implemented on a ultra low magnetic field (O.05T) tomographic system. Based on the conventional spin echo pulse sequence (SE), which is routinely used, we implemented the gradient recalled echo (GRE) pulse sequence and two sequences that use the principle of Steady-State Free Precession (SSFP) being \"Fast Low Angle Shot\" (FLASH) and \"Fast Acquisition Double Echo\" (FADE). With the SSFP sequences we shorten drastically the duration of the conventional SE tomography sequence. Besides this the FADE sequence gives two images with clearly different contrast without extending significantly the duration of the experiment. Needed for the SSFP techniques, we developed calibration procedures, which improved as well the SE sequence signal to noise ratio of about 15 percent. We analyzed theoretically and experimentally the behavior of the contrast of the presented sequences. Moreover we acquired various images of a phantom and the human brain of a normal volunteer using the different sequences and proposed some combinations of the parameters (protocols) repetition time, time to echo and flip angle. One of these protocols is being tested in clinical cases to compare the usefulness of the presented techniques for medical diagnostic. With this work we gained a wide experience in using SSFP techniques that will be indispensable in the elaboration of three-dimensional tomography methodologies

Apresentamos um estudo e a implementação de três técnicas destinadas a seleção de múltiplos planos, com objetivo de otimizar a tomografia por RMN. A primeira utiliza a excitação simultânea de múltiplos planos, sendo cada plano codificado com uma fase pré-determinada e a informação destes decodificada a posteriori através da combinação linear dos dados de n experimentos realizados. A segunda utiliza-se da excitação simultânea de múltiplos planos, como na técnica anterior, porém esses planos são adquiridos na presença de um gradiente de leitura oblíquo, que permite a obtenção simultânea dos sinais destes planos sem necessidade de pós-processamento. A terceira técnica usa a multiplexação de freqüências de excitação no tempo, isto é, diferentemente das teorias anteriores, excitam-se diferentes planos sucessivamente, durante o tempo de repetição de uma seqüência de aquisição, permitindo a aquisição de vários planos ao tempo de um único. A construção e o uso de um phantom destinado a caracterização do equipamento também é discutida. Esse phantom possibilita por exemplo: a determinação da largura do plano selecionado, o espaçamento entre os planos e o perfil destes, entre outros. Abordamos também, vários aspectos técnicos necessários a uma melhor performance do tomógrafo, tais como: circuitos de recepção e ressoadores. Apresentamos finalmente, uma discussão introdutória e os primeiros resultados experimentais já obtidos com a técnica de excitação adiabática com gradiente modulado (GMAX), utilizando bobinas de superfície / We present the study and the implementation of three techniques HIS for the selection of multi-slice, aiming the optimization of the NMR tomography . The first technique uses the simultaneous excitation of multiple slices, being each plane encoded with a pre determined phase for a latter decodification of the information by the linear combination of n experiments. The second one makes use of the multi-slice simultaneous excitation, like the first one, but the slices are acquired in the presence of an oblique reading gradient , which provides us the multi-slice signal without any further computer processing. The third technique uses frequency multiplexed excitation, that is, different slices are successively excited during the repetition time of an acquisition sequence, making possible multi-slice acquisition at the same time of a single slice The construction and the use of a phantom for the equipment characterizations are discussed too. With this phantom we can determine the thickness of the selected slice, the spacing between the slices and the their shapes. Many technical aspects necessary for an improvement of the tomography performance, like reception circuits and resonators, are discussed. At last, we present a brief introduction to the gradient modulated adiabatic excitation (GMAX) technique and the first results ever obtained with it, using surface coils

As bobinas ativamente blindadas (actively shielded) melhoram a performance das bobinas de gradiente ao cancelar o campo magnético na parte externa da bobina primária, evitando a degradação do campo no seu interior devido às correntes induzidas (eddy currents). O presente trabalho resultou na construção de um conjunto de bobinas de gradientes ativamente blindadas (Gx, Gy e Gz ), de geometria cilíndrica e acesso axial, com diâmetro de volume útil de 10cm para imagens (field of view), com campos de gradiente de aproximadamente 8.0 G/cm para r=0, com uma máxima distorção de campo de 5% neste volume, onde foi observado um efetivo cancelamento do campo externo, o que evita correntes induzidas nas partes metálicas internas do magneto OXFORD de 2.0 Tesla horizontal de 30cm de diâmetro, no qual serão empregadas. Alguns resultados prévios foram obtidos com um protótipo transversal, em apenas uma direção, e também com um conjunto de bobinas de gradiente para utilização em obtenção de imagens microscópicas de ESR pulsado. As bobinas foram projetadas utilizando-se a técnica de densidade de corrente, que se baseia na abordagem de campo alvo (target field) e também foi empregado o método de \"simulated annealing\" para otimização do desenho. Este sistema de gradientes apresenta também uma baixa indutância, o que permitirá chaveamentos rápidos (~0.1 ms). Na análise do campo de gradiente gerado foi utilizado o método de diferença de fase, confirmando a eficiência da técnica de densidade de corrente utilizada nos cálculos para a construção destas bobinas de gradiente / Actively shielded coils are designed to improve the performance of a gradient coil set by canceling the magnetic field in the region outside the body of the coil, avoiding degradation of the field in its interior due to eddy currents induced in the electric vicinity. The present work resulted on the construction of a shielded gradient coil set (Gx, Gy e Gz) with cylindrical geometry and axial access, with volume diameter of 1 O cm (imaging FOV) and gradient field intensity of 8.0 Gauss/cm in r=0. The maximum gradient distortion aim was 5% inside that volume, along with an effective cancellation of the field close to the inner metallic parts of the OXFORD 2.0 Tesla 31 cm bore horizontal magnet. Preliminary results were obtained with a prototype, which included only one transverse channel, and a miniature set with the same geometry was specially designed to perform micro imaging experiments using pulsed Electron Paramagnetic Resonance. Coils were designed using the current density technique, which is based on the target field approach and simulated annealing for optimization. The gradient systems described here present also a low inductance necessary to perform fast switching (~0.1 ms). For the analysis of the performance of the gradients we used the method of phase reference, confirming the efficiency of the current density technique used on the design of the gradient sets

Introdução: O carcinoma de mama é uma doença altamente prevalente e incidente. Em nosso meio, ainda cerca de metade dos casos são diagnosticados em estadios localmente avançados e/ou disseminados. Nesta situação, a quimioterapia neoadjuvante é o tratamento padrão.. Infelizmente, o padrão de resposta ao tratamento sistêmico é variável e existe um parcela de pacientes que não apresentam redução tumoral significativa e outros que apresentam progressão da doença. A identificação prévia do grupo de pacientes que mais se beneficiariam da quimioterapia neoadjuvante poderia evitar os efeitos adversos dos quimioterápicos no grupo de pacientes com baixa probabilidade de resposta bem como otimizar custos de tratamento. A ressonância magnética (RM) devido a sua análise tridimensional das imagens, alta sensibilidade e resolução espacial vem sendo empregada na monitorização do tratamento do câncer de mama, utilizando-se, principalmente, técnicas funcionais, como a avaliação farmacocinética após injeção do meio contraste paramagnético e as imagens ponderadas em difusão (DWI). A difusão pode ser utilizada para calcular o valor do coeficiente aparente de difusão (ADC). Recentemente, o ADC vem sendo utilizado como ferramenta diagnóstica e prognóstica no câncer de mama.. Alguns estudos mostraram que pacientes respondedoras à quimioterapia neoadjuvante apresentaram um aumento dos valores médios do ADC tumoral logo após o segundo ciclo do tratamento. Porém, poucos estudos, e com alguns resultados discordantes, avaliaram a capacidade do ADC em discriminar, previamente ao tratamento quimioterápico, aqueles tumores que terão melhor resposta patológica. Objetivos: comparar os diversos parâmetros derivados do ADC de neoplasias mamárias para avaliar possíveis preditores de resposta patológia à quimioterapia neoadjuvante. Materiais e métodos: Estudo retrospectivo envolvendo 57 pacientes com carcinoma de mama localmente avançado que realizaram RM previamente à quimioterapia neoadjuvante. Múltiplos métodos de medida do ADC foram realizados, sendo estudadas as validades das medidas calculando-se a área sobre a curva ROC e estabelecendo-se os valores de sensibilidade e especificidade. Os dados das diferentes mensurações foram analisados através do teste ANOVA. Para avaliação da variabilidade inter-observador foi utilizado o teste de Kappa. Resultados: As diferentes mensurações dos valores de ADC dos tumores não evidenciou diferença estatística significativa entre o grupo respondedor e não respondedor. Nenhum dos parâmetros analisados pode ser considerado como preditor de resposta. Não houve diferença significativa na obtenção dos diferentes parâmetros derivados do ADC entre a medida de um único corte na região central da lesão, quando comparada com a medida de toda a lesão (volumetria). Conclusão: Os diferentes parâmetros derivados da medida do ADC, pré-quimioterapia neoadjuvante não predizem resposta ao tratamento em pacientes com tumores de mama localmente avançados / Introduction : Breast cancer is a highly prevalent and incident disease. In Brazil , although about half of cases are diagnosed in locally advanced or disseminated stages . In this situation, the neoadjuvant chemotherapy is a standard treatment. Unfortunately, the pattern of response to systemic treatment is variable and there is a proportion of patients with no significant tumor reduction and others that have disease progression. Prior identification of the group of patients who would most benefit of neoadjuvant chemotherapy could avoid the adverse effects of chemotherapy in patients with low probability of response and optimize treatment costs. Magnetic resonance imaging ( MRI) due to its three-dimensional image analysis, high sensitivity and spatial resolution has been used for monitoring the treatment of breast cancer, using mainly functional techniques such as pharmacokinetic evaluation after injection of paramagnetic contrast medium and diffusion weighted imaging (DWI) . It can be used to calculate the apparent diffusion coefficient (ADC). Recently, the ADC has been used as a diagnostic and prognostic tool in breast cancer. Some studies have shown that patients who respond to neoadjuvant chemotherapy showed an increase of the mean ADC tumor values after the second cycle of treatment. However, few studies and with some discordant results, reviewed the ADC \'s ability to discriminate, prior to chemotherapy, those tumors that have better pathological response. Objectives: To compare the various parameters derived ADC of breast tumors to evaluate possible predictors of response to neoadjuvant chemotherapy pathology. Methods : a retrospective study of 56 patients with locally advanced breast carcinoma who underwent MRI prior to neoadjuvant chemotherapy. Multiple ADC measurement methods were performed and the validity of the measures was studied by calculating the area under the ROC curve and setting up the sensitivity and specificity values. The data of different measurements were analyzed using ANOVA test. To evaluate the interobserver variability was used Kappa test. Results: The different measurements of ADC values of tumors showed no statistically significant difference between the responder group and the group with no response. None of the analyzed parameters can be suggested as predictor of response. There was no significant difference in obtaining the various parameters derived from the ADC either measuring a single slice, at the central region of the lesion or measuring the whole lesion (volumetry). Conclusion: The different parameters derived from the measurement of the ADC prior to neoadjuvant chemotherapy do not predict response to therapy in patients with locally advanced breast cancers..

A dosimetria 3D utilizando gel à base de monômeros é uma importante ferramenta utilizada em casos de Radioterapia, em que há necessidade de uma alta resolução da distribuição espacial de dose. Neste trabalho, desenvolvemos um gel à base de monômeros do acido metacrilico, denominado MAGIC®, com a finalidade de avaliar a distribuição espacial de dose em simulações de tratamento de radioterapia, utilizando duas técnicas de imagens por ressonância magnética, Magnetic Resonance Imaging (MRI): a relaxometria (RT) e a transferência de magnetização (Magnetization Transfer, MT). Para tal, o trabalho foi desenvolvido em três etapas. Na primeira, desenvolvemos fantomas para serem irradiados com radiação ? e x, sobre os quais imagens de ressonância foram obtidas e analisadas. Desenvolvemos também um software, escrito em Matlab®, para analisar as imagens e traçar curvas de isodose dos fantomas irradiados. Em seguida, aperfeiçoamos o preparo, a composição, e as maneiras de adquirir e processar as imagens do gel irradiado. Nessa etapa incorporamos o formaldeido ao gel MAGIC, aumentando sua sensibilidade em 15%. Por fim, simulamos cinco configurações de tratamento de radioterapia e avaliamos as isodoses adquiridas, comparando-as com aquelas obtidas pelo software de planejamento radioterápico virtual, TPS®. As avaliações feitas com a técnica de relaxometria corresponderam às nossas expectativas. Com ela foi possível caracterizar o gel MAGIC, com alguns testes dosimétricos, bem como avaliar as 5 simulações de tratamentos radioterápicos propostas. O gel se mostrou linear até 20 Gy, mas não apresentou níveis aceitáveis de reprodutibilidade, necessitando de uma curva de calibração em cada teste. O gel possui número atômico efetivo próximo ao da água não necessitando de fatores de correções. Na maioria dos casos simulados, as isodoses das simulações feitas com o gel MAGIC reproduziram as simulações virtuais. Infelizmente, não obtivemos, por enquanto, o mesmo sucesso com a avaliação por MT. As imagens de MT não se mostraram confiáveis e por esse motivo as simulações com o gel só foram feitas com a técnica de RT. / The 3D monomeric gel dosimetry is an important tool in radiation therapy cases which needs high spatial dose resolution. In this work we developed a methacrilic acid monomeric gel called MAGIC® to evaluate spatial dose distributions in simulations of radiation therapy treatments. We used two Magnetic Resonance Imaging (MRI) techniques, the relaxometry (RT) and the magnetization transfer (MT). The work was developed in three stages. First of all, the phantoms were created to be irradiated using and x radiations, the MRI were acquired and analyzed. We also developed a software, programmed in Matlab®, to analyze the images and to draw isodoses curves of irradiated phantoms. Following it, we improved the way the gel was prepared and its composition, as well the image acquisition and processing. In this part we added formaldehyde to the gel, improving its sensibility in 15%. Finally, we simulated 5 different radiation therapy treatments and compared the isodoses measured with the isodoses of the radiation therapy treatment planning software (TPS®). The results of the RT technique corresponded to our expectations. Using the RT we characterized the gel with dosimetric tests and evaluated five different radiation therapy treatments. The gel showed a linear relation with the dose until 20 Gy, but the results of the tests were not reproducibles, because of that we made a calibration curve for each test. The effective atomic number of the gel is close to the water, so it was not necessary any correction. In most of the simulated cases, the isodoses measured with the gel reproduced the virtual simulations. Unfortunately, until now, we didn\' t have the same success using the MT technique. The MT images were not reliable and because of that the simulations were only made with the RT technique.

INTRODUÇÃO: A ultrassonografia (US) e a ressonância magnética (RM) representam os principais métodos de diagnósticos por imagem na avaliação do mecanismo extensor do joelho. Há carência de trabalhos que avaliem por US e RM a presença de alterações no mecanismo extensor de jogadores de futebol assintomáticos. OBJETIVO: Avaliar, por meio de RM e US, o tendão quadríceps e o ligamento patelar de jogadores de futebol assintomáticos e dos participantes do grupo-controle. Identificar se há alterações de imagens ou lesões. Determinar se essas sofrem influências pelos seguintes critérios: canhoto ou destro, quanto à dominância ao chute, posição em campo, idade do atleta e tempo de prática esportiva. MÉTODO: Foram avaliados um total de 248 joelhos: 112 eram de 56 de jogadores de futebol assintomáticos e 136 eram do grupo-controle. O grupo composto por jogadores de futebol apresentava idades entre 14 e 34 anos. Já o grupo-controle era composto por indivíduos com idade também entre 14 e 34 anos, principalmente, universitários e funcionários do hospital, com gênero igual aos dos jogadores selecionados, mas sem atividades físicas de impacto, como vôlei, futebol e salto, por exemplo. RESULTADO: Constatou-se a presença de alterações de imagens ou de lesões em nove ligamentos patelares e apenas duas nos tendões quadríceps de jogadores de futebol assintomáticos, tanto pela US quanto pela RM, e nenhum caso alterado no grupo-controle. Houve concordância significativa entre RM e US quanto aos resultados obtidos. Além disso, os jogadores de futebol apresentaram anormalidades (alterações de imagens ou lesões) no tendão quadríceps ou no ligamento patelar significativamente (p<0,05) maior do que no grupo-controle; a idade média dos jogadores com alguma anormalidade é significativamente menor do que a dos jogadores sem lesão (p<0,003); o tempo médio de prática desportiva dos jogadores que apresentam anormalidade é menor do que o tempo médio dos jogadores que não têm anormalidade (p<0,001); os jogadores destros apresentaram anormalidades, enquanto que os jogadores canhotos não (p<0,05). Dentre os jogadores, há correlação entre mais alterações de imagem ou lesão de joelho e menor tempo como jogador e menor idade (R² de Nagelkerke = 0,700). Destros têm risco aumentado de ter anormalidade nas referidas estruturas (Razão de Chances = 15,204) quanto menor for o tempo como jogador de futebol. Os jogadores da defesa têm 4,76 vezes mais chance de desenvolver anormalidade do que atacantes e goleiros (p<0,04). CONCLUSÃO: Os jogadores de futebol assintomáticos apresentam anormalidades no tendão do quadríceps e ligamento patelar detectadas pela RM e US. As alterações estão relacionadas a menor idade, menor tempo de prática desportiva, ser destro e ser jogador de defesa / INTRODUCTION: Ultrasonography (US) and magnetic resonance imaging (MRI) are the main diagnostic imaging methods used in evaluation of the extensor mechanism of the knee. Theres a scarcity of works assessing alterations on the extensor mechanism of the knee of asymptomatic soccer players using US and MRI. OBJECTIVE: To assess, by means of MRI and US imaging, the quadriceps tendon and patellar ligament of asymptomatic soccer players and healthy control group; identify abnormal imaging findings or injuries; ascertain whether these are affected by leg dominance (left or right), field position, age, and time spent practicing the sport. METHOD: A total of 248 knees were evaluated: 112 of 56 asymptomatic soccer players and 136 of healthy controls. The players ages ranged from 14 to 34 years. The control group was composed mostly of college students and hospital staff members, also with ages between 14 and 34, matched for gender to the selected soccer players, and who did not engage in high-impact activities such as volleyball, soccer, or jumping. RESULTS: Abnormal imaging findings or injuries were detected, by US and MRI alike, in nine patellar ligaments and two quadriceps tendons of asymptomatic soccer players. No abnormal findings were detected in the control group. CONCLUSION: There was significant consistency between MRI and US findings. Soccer players were much more likely (p<0.05) to have abnormal imaging findings or injuries of the quadriceps tendon and patellar ligament as compared with the control group; the average age of players presenting some abnormality is significantly lower than of those without injuries (p<0.003); the average time spent practicing the sport for players presenting abnormalities is lower than those who dont present injuries (p<0.001); a correlation between the dominant leg and the affected knee was observed, with the players with a dominant right leg presenting injuries while the ones with a dominant left leg not showing abnormalities (p<0.05). Amongst the players there is a correlation between more abnormal imaging findings or injuries on the knee and less professional practicing time and age (Nagelkerke R² = 0.700).Players with a dominant right leg have an increased chance of presenting abnormalities on the quadriceps tendon and patellar ligament (Odds Ratio = 15.204) with a lower average time spent practicing the sport. Defense players have 4.76 times more chance to develop abnormalities than forward players and goalkeepers (p<0.04). CONCLUSION: The asymptomatic soccer players present abnormalities on the quadriceps tendon and patellar ligament identified using ultrasonography (US) and magnetic resonance imaging (MRI). The alterations are related with less age, less time of professional practice, right leg dominance and defense field position

Human structural and functional brain architecture is increasingly studied by applying the mathematical framework of complex networks to data from magnetic resonance imaging. Connections (edges) in such brain networks are commonly constructed using correlations of features between pairs of brain regions, such as regional morphology (across participants) or neurophysiological time series (within participants). Subsequent analyses frequently focus on summary network statistics calculated using the strongest correlations, but often neglect potential underlying shifts within the correlation distribution. This thesis presents methods for the construction and analysis of correlation-derived structural and functional brain networks, focusing on the implications of changes within the correlation distribution. First, schizophrenia is considered as an example disease which is known to present a reduction in mean correlation between regional neurophysiological time series. Previous studies reported increased network randomisation in schizophrenia, but these results may have been driven by inclusion of a greater number of noisy edges in patients’ networks, based on retention of a fixed proportion of the strongest edges during network thresholding. Here, a novel probabilistic thresholding procedure is applied, based on the realisation that the strongest edges are not necessarily most likely to be true following adjustment of edge probabilities for effects of participant in-scanner motion. Probabilistically thresholded functional networks show decreased randomness, and increased consistency across participants. Further, applying probabilistic thresholding eliminates increased network randomisation in schizophrenia, supporting the hypothesis that previously reported group differences originated in the application of standard thresholding approaches to patient networks with decreased functional correlations. Subsequently, healthy adolescent development is studied, to help understand the frequent emergence of psychiatric disorders in this period. Importantly, both structural and functional brain networks undergo maturational shifts in correlation distribution over adolescence. Due to reliance of structural correlation networks on a group of subjects, previous studies of adolescent structural network development divided groups into discrete age-bins. Here, a novel sliding-window method is used to describe adolescent development of structural correlation networks in a continuous manner. Moreover, networks are probabilistically thresholded by retaining edges that are most consistent across bootstrapped samples of participants, leading to clearer maturational trajectories. These structural networks show non-linear trajectories of adolescent development driven by changes in association cortical areas, compatible with a developmental process of pruning combined with consolidation of surviving connections. Robustness of the results is demonstrated using extensive sensitivity analyses. Finally, adolescent developmental changes in functional network architecture are described, focusing on the characterisation of unthresholded (fully weighted) networks. The distribution of functional correlations presents a non-uniform shift over adolescence. Initially strong cortical connections to primary sensorimotor areas further strengthen into adulthood, whereas association cortical and subcortical edges undergo a subtler reorganisation of functional connectivity. Furthermore, individual subcortical regions show distinct maturational profiles. Patterning of maturation according to known functional systems is affirmed by partitioning regions developing at similar rates into maturational modules. Taken together, this thesis comprises novel methods for the characterisation of disease-related and normative developmental changes in structural and functional correlation brain networks. These methods are generalizable to a wide range of scenarios, beyond the specific disease and developmental age-ranges presented herein.

The research presented within this dissertation focused on two major areas of research: monitoring the response to single-dose radiotherapy in a RIF-1 tumor model and the role of cortical spreading depression in the evolution of ischemic stroke. For the research in the first half of this dissertation, quantitative MRI was performed to investigate the spatial correlation between the apparent diffusion coefficient (ADC), spin-spin relaxation times (T2), and proton density (M0) in murine radiation-induced fibrosarcoma (RIF-1) tumors following single-dose (1000cGy) radiotherapy using the k-means (KM) algorithm. An in-depth comparison between KM-derived volume estimates and conventional histology via the hematoxylin-eosin (H&E) staining procedure (for identification of viable tumor versus necrosis), as well as via hypoxic-inducible factor-1alpha (HIF-1alpha) immunohistochemistry (for identification of regions of hypoxia versus well-oxygenated tissue) was performed. The results of this study demonstrated that multispectral (MS) analysis provides: (1) an improved tissue segmentation method over results obtained from conventional single-parameter approaches, (2) subdivision based on the degree of necrosis, as well as delineation between well-oxygenated and hypoxic viable tissue, (3) good correlation with both H&E staining and HIF-1alpha immunohistochemistry, and (4) a method for monitoring the range of tissue viability as a function of time post-treatment, with the potential for predicting therapeutic efficacy. For the research in the second half of this dissertation, manganese-enhanced MRI (MEMRI) was employed for the characterization of both experimental and pathological cortical spreading depression (CSD). In order to determine the utility of manganese ions (Mn2+) as a marker for spreading depression (SD), experimental SD was elicited by chemical stimulation (KCl application to exposed rat cortex) and compared with control conditions. This study demonstrated that (1) Mn2+ is a more accurate marker for SD than DWI or T2* methods, (2) cortical restriction of MEMRI enhancement supports the contention that apical dendrites are necessary for SD propagation. (3) subcortical enhancement was a result of cortical-subcortical neuronal connectivity. Based on these results, preliminary experiments involving the study of SD in ischemia using Mn2+ were performed. Initial results indicate: (1) MEMRI may provide a method for estimating the likelihood of progression to infarction at acute timepoints post onset of stroke. These studies provide a foundation for further investigation into the role of SD in stroke, and the application of Mn2+ towards the design of therapeutic strategies targeting SD inhibition.

"Diffusion-weighted nuclear magnetic resonance has gained widespread use in the characterization of various diseases. Developments in the area of porous media theory have been successfully transferred and adapted for the use in biological tissue. Measurement of the displacement of diffusing water molecules can reveal structural information about the environment in which the molecules translate. The return-to-the-origin (RTO) probability and the apparent diffusion coefficient (ADC) are based on the diffusion behavior of water molecules in a restricted environment. Water molecules in restricted space have smaller displacements, for a given diffusion time, than water molecules diffusing in bulk solution. The cell membranes and organelles in healthy biological tissue impart more restrictions on diffusing water molecules as compared to necrotic tumor tissue. In necrotic tissue the degradation of cellular structures by auto- and/or heterolysis allows water molecules to diffuse over larger distances without encountering restrictions. The spectroscopic measurement of the RTO probability and the RTO enhancement in RIF-1 tumors showed that the RTO probability is sensitive to these changes in structure. The study showed that smaller tumors, which are less necrotic, have a larger RTO probability and enhancement than larger RIF-1 tumors with a higher fraction of necrotic tumor tissue. Extension of the methodology to NMR imaging was used to answer the question if the RTO probability can provide spatial information about the necrotic area within RIF-1 tumors. The necrotic area measured by the ADC and histology were compared. While neither ADC or RTO could show its superiority over the other, both methods showed a good correlation between their mean values and the necrotic area fraction as measured by histology. The mean ADC and the mean RTO enhancement had a correlation with the necrotic tumor fraction, as determined by histology, of r = 0.86 and r = -0.82, respectively. Conventional T2-weighted images of the same tumor slice showed a poorer correlation (r = 0.62) with the necrotic fraction and no visual agreement with the histology. The general features of the ADC and RTO enhancement were in agreement with histology, however, more exact comparisons where not possible due to the large differences in slice thickness between the two techniques. Structural changes similar to those caused by tumor tissue necrosis can be induced by chemo- and radiation therapy and ADC and RTO enhancement were used to monitor these changes non-invasively. RIF-1 tumors were grown on the hind leg of C3H mice and monitored daily by diffusion-weighted MRI. ADC and RTO-enhancement maps were created using data acquired from control animals and animals treated with 100 mg/kg 5-Fluorouracil. Both ADC and RTO proved to be useful in the early detection of the efficacy of treatment as well as for monitoring the progress of therapy. Diffusion measurements by pulsed-field-gradient (PFG) MRI have become an important tool for detecting of pathophysiological changes caused by cancer and stroke. The increasing popularity of diffusion measurements has initiated their use on clinical MRI systems that have limited magnetic-field-gradient strength. These limitations make it necessary to lengthen the diffusion-gradient duration to ensure sufficient signal attenuation for calculating the ADC. Unfortunately, increasing of the diffusion-gradient duration to a large extent violates the theoretical model used in the ADC calculation. The diffusion measurements are not performed in the finite pulse width regime, but rather in the constant gradient regime, requiring a different interpretation of the results. Examination of the differences in the measured diffusion coefficient showed that increasing both the diffusion-gradient duration and the echo time have a significant impact on the results of a diffusion measurement. A different way to assess changes in RIF-1 tumors as a function of treatment is the measurement of the tissue oxygen status. Cell hypoxia has long been linked with treatment resistivity and reoccurrence in cancers, where the oxygen status is a determining factor of treatment outcome. Perfluorocarbons (PFC's) have been used successfully to assess the tumor oxygen status in the past, but required a large MRI slice thickness due to compensate for the low PFC concentration. The tissue oxygen status of the tumor is assessed by intravenous injection of a PFC that is subsequently sequestered in the tumor. The measurement of the T1-relaxation time of the PFC allows the calculation of the oxygen content, which is linearly related to the relaxivity and the temperature. Fluorine-19, multiple-slice, inversion-recovery echo-planar imaging (EPI) allowed high spatial resolution assessment of the tissue oxygen status over the entire tumor. The results demonstrated that there is a large variation in tissue oxygenation between adjacent slices. Comparison of the oxygen distribution between various tumors also showed that there is no common pattern in the spatial distribution of oxygen within the tumor. Monitoring of the oxygen status during chemotherapy showed an increase in hypoxic tissue and a reduction in tumor size in response to the toxicity of the chemotherapeutic agent. As the effects of the treatment subsided, rapid cell proliferation caused the tumor to regrow and a subsequent decrease in tissue oxygen tension was observed. The study clearly demonstrated the changes in oxygen tension in response to chemotherapy and the need for multi-slice MRI acquisition at high spatial resolution to detect these changes."