hide
Free keywords:
-
Abstract:
Introduction
Recent studies at ultra-high field [1,2] show that FID based MR spectroscopic imaging (FID-MRSI) avoids in-plane chemical shift displacement and minimizes T2 related signal losses. Yet, FID-MRSI spectra suffer from lipid contaminations (arising from subcutaneous fatty tissue) and phase distortions, caused by the FID truncation (due to the acquisition delay necessary for the excitation pulse and gradients). Our aim was to examine the feasibility of FID-MRSI of the healthy human brain at 9.4T. In the presented approach the missing FID points were reconstructed with an autoregressive model [3] and the lipid contaminations were minimized by improving the point spread function which was achieved by increasing the spatial resolution and Hamming filter as in [2]. Total acquisition time of FID-MRSI sequence was shortened by optimizing the water suppression (water suppression enhanced through T1 effects-WET) [4] scheme.
Methods
Spectra were collected at a 9.4T MR scanner (Siemens, Erlangen, Germany) using a custom built coil [5]. Water suppressed spectra (with original and optimized WET) were acquired from the brain of two volunteers. In-vivo measurements were approved by the local ethical committee. Acquisition parameters: TR: 340 and 240ms (original and optimized WET, respectively), acquisition delay: 2.3ms, 64×64 voxels, acquisition duration: 128ms, spectral bandwidth: 4kHz, nominal voxel size 3.1×3.1×10mm. These acquisition parameters resulted in a total acquisition time of 21 (original) and 15 min (optimized WET).
Results
The remaining water signal after suppression is depicted in Fig.1. It can be seen that in the case of optimized WET (c), not only the residual water signal is smaller, but also the suppression is more homogenous against B1+ variations (a). Acquisition at high in-plane resolution combined with Hamming filtering minimized contaminations with lipid signals. This was even the case for voxels closely located to the skull (Fig. 2, graphs 14), where the spectral range of interest between 2 and 4 ppm was unaffected.
Conclusions
We showed that FID-MRSI at a field strength of 9.4T is feasible. It was possible to reconstruct the missing FID points, so that the phase distortions present in the spectra could be reduced. Optimization of the water suppression scheme allowed significant reduction of total acquisition time. In conclusion, FID-MRSI at 9.4T is highly promising, as it addresses the most critical problems of chemical shift displacement and signal losses due to fast T2 relaxation.