La Spectroscopie et l'Imagerie Spectroscopique de Résonance Magnétique (ISRM) jouent un rôle émergent parmi les outils cliniques, en donnant accès, d'une manière complètement non-invasive, aux concentrations des métabolites in vivo. Néanmoins, les inhomogénéités du champ magnétique, ainsi que les courants de Foucault, produisent des distorsions significatives de la forme de raie des spectres, induisant des conséquences importantes en termes de biais lors de l'estimation des concentrations. Lors des traitements post-acquisition, cela est habituellement traité à l'aide des méthodes de pré-traitement, ou bien par l'introduction de fonctions analytiques plus complexes. Cette thèse se concentre sur la prise en compte de distorsions arbitraires de la forme de raie, dans le cas des méthodes qui utilisent une base de métabolites comme connaissance a priori. L'état de l'art est évalué, et une nouvelle approche est proposée, fondée sur l'adaptation de l'amortissement de la base des métabolites au signal acquis. La forme de raie présumée commune à tous les métabolites est estimée et filtrée à l'aide de la méthode LOWESS. L'approche est validée sur des signaux simulés, ainsi que sur des données acquises in vitro. Finalement, une deuxième approche novatrice est proposée, fondée sur l'utilisation des propriétés spectrales de la forme de raie commune. Le nouvel estimateur est testé seul, mais aussi associé avec l'estimateur classique de maximum de vraisemblance, démontrant une réduction significative du biais dans le cas des signaux à haut rapport signal-sur-bruit. / Magnetic Resonance Spectroscopy (MRS) and Spectroscopic Imaging (MRSI) play an emerging role in clinical assessment, providing in vivo estimation of disease markers while being non-invasive and applicable to a large range of tissues. However, static magnetic field inhomogeneity, as well as eddy currents in the acquisition hardware, cause important distortions in the lineshape of acquired NMR spectra, possibly inducing significant bias in the estimation of metabolite concentrations. In the post-acquisition stage, this is classically handled through the use of pre-processing methods to correct the dataset lineshape, or through the introduction of more complex analytical model functions. This thesis concentrates on handling arbitrary lineshape distortions in the case of quantitation methods that use a metabolite basis-set as prior knowledge. Current approaches are assessed, and a novel approach is proposed, based on adapting the basis-set lineshape to the measured signal.Assuming a common lineshape to all spectral components, a new method is derived and implemented, featuring time domain local regression (LOWESS) filtering. Validation is performed on synthetic signals as well as on in vitro phantom data. Finally, a completely new approach to MRS quantitation is proposed, centred on the use of the compact spectral support of the estimated common lineshape. The new metabolite estimators are tested alone, as well as coupled with the more common residual-sum-of-squares MLE estimator, significantly reducing quantitation bias for high signal-to-noise ratio data.

<p>Lipids and sodium (²³Na) are two essential components of the human body. They play a role in almost all biological systems. However, an increase in their levels is associated with metabolic diseases. The elevation of their contents can cause similar health disorders. Examples of prevalent disorders that share an increase of musculoskeletal lipids and ²³Na are hypertension and diabetes. However, the relationship between in vivo lipid and sodium levels in pathophysiology has not been studied enough and therefore is still unclear. Additionally, the available quantification methods to facilitate such a study may not be practical. They are either invasive, not sensitive enough, or require an impractical measurement time.</p> <p>Therefore, in this work, our aims were to develop practical in vivo methods to quantify the absolute sodium concentration as well as the concentration of each lipid component individually, and to study the correlation between them within the skeletal muscles.</p> <p>Since lipids and ²³Na have different nuclear magnetic resonance properties, their quantification by magnetic resonance (MR) techniques face different challenges. Thus, we optimized different MR spectroscopic imaging (MRSI) techniques for lipids and ²³Na. </p> <p>Our proposed proton MRSI was able to provide eight lipid fat fraction (FF) maps representing each musculoskeletal lipid component (fatty acid) detected by our MRSI technique, and demonstrated a superior sensitivity compared to the conventional MR imaging methods.</p> <p>For ²³Na, our developed ²³Na-MRSI was able to measure and map the absolute ²³Na concentration with values agreeing with those reported previously in biopsy studies, and with a high repeatability (CV < 6 %) within significantly shorter acquisition time compared to other available techniques.</p> <p> Finally, the ²³Na concentration and the fat fractions of each lipid component within healthy skeletal muscles were measured and correlated using our developed MRSI methods. Our findings suggest a positive regional relationship between ²³Na and lipids and negative correlation between ²³Na and BMI under healthy conditions.</p>