Sodium (23Na) Imaging at 9.4 Tesla
Although hydrogen is the most commonly imaged nucleus in MRI due to its high concentration in biological tissues in the form of water and fat, any other nucleus having a net nuclear spin, e.g. 23Na, 17O, 19F, etc., can be used for imaging and potentially gaining additional information. Even though sodium is the second most abundant nucleus in the human body, the signal produced by it is considerably lower than the one obtained from hydrogen. This is due, on the one hand, to the underlying physics characterising the relaxation of the excited sodium nucleus and, on the other hand, to the 1000 fold lower in vivo concentration.
Sodium imaging benefits greatly from the advent of MR scanners with ultra-high magnetic field strengths (>4T). The concomitant increase in sensitivity permits the acquisition of sodium images exhibiting a higher spatial resolution and signal-to-noise ratio (SNR), while maintaining a clinically acceptable scan time. As sodium ions play an important role in cellular homeostasis and cell viability, 23Na MRI can serve as a means for the assessment of cell integrity. Furthermore, degenerative changes, e.g. in articular cartilage, might be assessed in the future by the aid of sodium imaging.
The relaxation of the excited sodium nucleus is characterized by a fast (bi-) exponential decay, leading to a large decrease in signal intensity within only a few milliseconds. As standard imaging protocols used for proton imaging are not well suited to cope with this problem, specialized imaging modalities have to be implemented and optimized for the use at ultra-high field strength. In this project, ultra-short echo time (UTE) sequences, such as a stack-of-spirals, were implemented to acquire sodium images at 9.4 T.
The use of non-Cartesian imaging sequences enables a more efficient acquisition of the sodium signal. However, the reconstruction of the acquired data becomes more challenging because the data samples are in most cases not anymore evenly spaced in k-space. As a matter of fact the image cannot be reconstructed straightaway by a standard Fourier transformation. Instead, reconstruction algorithms such as gridding must be used to first interpolate the acquired data to a Cartesian grid.
In most cases sodium images are acquired with long repetition times in order to minimize relaxation weighting. Nevertheless, fast imaging sequences can also be used for sodium imaging and achieve a higher SNR in a shorter time as demonstrated by the images shown on the right-hand side. An SNR comparison was performed for three spiral imaging sequences which used either RF spoiling, gradient spoiling (FISP ? fast imaging with steady-state precession) or balanced gradients (bSSFP ? balanced steady-state free precession). The sequence parameters of the used stack of spirals were chosen as follows: nominal resolution 1.5x1.5x4.0 mm3, partitions 52, spiral interleaves per partition 130, TR 10 ms, readout time 3 ms (RF spoiled acquisition and FISP) and 5 ms (bSSFP), acquisition time (TA) 10 min. The duration of the hard excitation pulse was set to 2 ms in order to achieve sufficient flip angle while adhering to the prescribed SAR limits. The bSSFP images exhibit excellent image quality and only a few banding artefacts can be seen. The relative SNR measured in a large region of interest (ROI) in a central slice was: 42 (spoiled acquisition), 44 (FISP), and 56 (bSSFP), respectively.
Tissue sodium concentration maps
Tissue sodium concentration (TSC) maps can be obtained based on UTE images if adequate reference standards are available for calibration. The figure on the left-hand side shows TSC maps acquired at 9.4 T in a healthy volunteer. The nominal resolution of these images is 1x1x5 mm3 and the measurement time was 30 min. Six reference standards were placed below the subject's head in order to relate the measured signal intensity to the sodium concentration. An average sodium concentration of 36±2 and 31±1 mmol/L of wet tissue were for gray and white matter, respectively.
|2015-2016||Scientist at the Max-Planck-Institute for Biological Cybernetics|
|2011-2016||PhD student at the University of Tübingen, guest scientist at the , MPI for Biological Cybernetics|
|2005-2011||Diploma in Physics at RWTH Aachen University, Aachen, Germany|
|2008-2009||Erasmus Exchange Program at Imperial College London, London, United Kingdom|