• sodium MRI
• sodium quantification
• Ultrashort Echo Time imaging

Magnetic Resonance Imaging (MRI) is a powerful imaging technique based on the inter- action between a large external magnetic field and the 1-hydrogen spins within the human body. MRI is an imaging modality widely appreciated for its non-invasivity (due to the usage of non-ionizing radiations), for its capability to acquire images weighted with different kinds of parameters (T1, T2, proton density) and its versatility in providing structural, functional and metabolic informations with an unique exam. The last two decades have seen the widespread diffusion of MRI scanners with growing field strengths: 3T systems have become more and more diffused in clinical practice, while the research has moved toward Ultra High Field strengths (UHF, B0 ≥ 7T). In 2011, the first and only 7T system for human studies in Italy has been installed at the research center IMAGO7, which is located within the IRCCS Stella Maris (Calambrone, PI). Thanks to their better SNR with respect to conventional scanners, Ultra High Field systems allow faster imaging and higher spatial and spectral resolution. Moreover, conveniently tuning the resonant frequency and using specific detectors, these scanners have the capability to detect the signal from nuclei other than 1-hydrogen (such as 23-sodium) which cannot be studied at lower strengths due to their low abundance and small gyromagnetic ratio. The subject of the present thesis is to explore several UHF techniques to perform sodium imaging and to obtain quantitative Tissue Sodium Concentration (TSC) maps, in order to establish an imaging and a quantification protocol for human studies. The feasibility of 23-sodium imaging is important since sodium is the second most abundant MR-active nucleus in human body, and it is involved in a large set of physiological processes such as nerve signal transmission and muscle action. For example, sodium concentration in cartilage is correlated with the biochemical components which regulate the cartilage homeostasis; an alteration of cartilage sodium content is a marker for pathologies such as osteoarthritis. Moreover, the balance between intra- and extra-cellular sodium concentration is kept fixed by several mechanism, so that disturbances of this balance are often signs of disorders (e.g tumors, strokes) which diagnosis would be improved from the acquisition of quantitative TSC maps.
Since 23-sodium is a 3/2 spin nucleus, it exhibits a nuclear quadrupole moment which interacts with local electric field gradients due to the surrounding environment, resulting in relaxation times of the sodium MR signal that are considerably shorter than those of 1hydrogen. This represents a technical challenge for sodium MRI, which has to be overcome 1with non-conventional techniques such as non-cartesian imaging and Ultrashort Echo Time (UTE) imaging. In this work two UTE techniques with different non-cartesian trajectories, 3D Radial and 3D Cones, have been implemented and compared in terms of SNR, acquisition time and sodium quantification capability; the implementation and optimization of these methods have been carried out performing phantom experiments. The results of the in-phantom studies have then been applied for the imaging and quantification of a human knee cartilage of an healthy volunteer, in order to analyze the method performance and limitations in a potential clinical setting (e.g osteoarthritis diagnosis).
The thesis is divided in two main parts; the first (Chapters 1 − 2 − 3 − 4) covers the theoretical background of the Magnetic Resonance Imaging, with a specific focus on the topic of Sodium MRI. The second part (Chapters 5 − 6 − 7) reports the experimental part of the work.
In details, a method for creating TSC maps, which includes post-processing and calibration of the raw images, has been developed and tested in a phantom experiment, starting from the implementation and optimization of the sequences. Then, some methods to reduce the acquisition time were applied and their impact on the quantification accuracy was verified. The measures were carried out using three different acquisition schemes, in order to define a flow chart for a clinical study.
The whole experiment was repeated on two dual-tuned (1-hydrogen/23-sodium) surface radiofrequency coils: a commercial coil, which represents a gold standard, and a custom coil built at IMAGO7, to ensure the reliability of the latter. The measurements for SNR estimation were carried on each of the three acquisition schemes and for both the two coils.
Then, the results in terms of protocol (parameters, trajectories, post-processing and analysis) obtained in the phantom experiment were tested on a human knee, in order to evaluate the in-vivo reliability of the measure and the limitations of the technique.
Finally, the problem of the image distortions due to the eddy currents, which are a major issue of the non-cartesian techniques used for sodium MRI, was investigated; more specifically, the impact of these distortions on the accuracy of the quantifications, has been evaluated via computer simulations.
At the end of this work, an optimized protocol for sodium MRI is available on the 7T scanner for in-vitro and in-vivo applications. Phantom experiments confirmed the possibility of evaluating the sodium content in a tissue by quantitative maps and provided information on their accuracy and reproducibility. Although some hardware limitations could affect in-vivo quantification, associated errors have been estimated and some methods for corrections have been proposed. Further studies, oriented on both hardware and software implementations, will improve the absolute in-vivo quantification of sodium concentrations.