Synopsis

Navigator based motion correction is shown for single voxel spectroscopy at 7T, utilizing fat-selective navigators and static higher order shims, but with linear shims that are switched between 2 sets for the navigator and spectroscopy sub-sequences.

Introduction

Prospective motion correction using volumetric navigators has previously been shown to improve motion robustness for single voxel MR spectroscopy at 3T [1,2]. In these studies, the shim was switched between 1^st order shim sets optimized for the spectroscopy and navigator sub-sequences, respectively.

At 7T, 2^nd-3^rd order shimming is preferred for single voxel MRS, but although that improves the shimming inside the spectroscopy voxel, it typically degrades the shim in the rest of the brain. This makes it challenging to interleave accurate volumetric navigators. Dynamic higher order shimming can alleviate this constraint, but a sophisticated calibration is required to compensate higher order eddy currents.

We therefore suggest a solution where the higher order terms remain constant during the scan, but the linear shim terms are switched between 2 predefined sets when switching between the navigator and spectroscopy sub-sequences. We furthermore use fat-selective navigators to ensure the navigators are not corrupted by saturation slices from the spectroscopy, which is expected to be worse at 7T compared to 3T due to the longer T₁ relaxation times. Together this enables motion correction for single voxel MRS using higher order shimming at 7T.

Methods

All experiments were performed on two 7T MRI systems (Achieva, Philips, Best, Netherlands) and performed in accordance to the local ethical guidelines.

B₀ mapping was performed with whole brain 3D-B­­₀ maps (20 slices, 3.75x3.75x3.75mm; 240x240x116mm FOV, TR/TE/ΔTE=5.2/2.5/1ms). The brain was segmented and analyzed with first through third order shimming using non-linear constrained fitting on the following different approaches:

1. BRAIN shim; Whole brain shimming (BRAIN shim)
2. ROI shim; region of interest shimming, using 5 mm extension on the MRS voxel
3. DYN shim; a cost-function weighting shim between whole brain and ROI with dynamic linear shimming and a shared set of static higher order terms.

Simulations of the shim performance of the different approaches was performed on B₀ maps from 10 subjects with each 6 voxel locations throughout the brain.

Prospective motion correction was performed using the iMOCO framework [3]. A semi-LASER voxel (size 20x20x20mm³, TE/TR=36/4500, 16 averages, VAPOR water suppression) was placed in the arterial cingulate cortex (Figure 4). B₀ drift was tracked by readouts of the central k-space line, in a slice through the voxel. For motion tracking, a fat-selective navigator [4] was acquired in the pause after each spectroscopy readout, and the positions, angulation and B₀ were updated before the next repetition of the spectroscopy sequence (Figure 1). The navigator was acquired with a 1-degree binomial 1-3-3-1 excitation pulse, resolution = 7x7x7mm, FOV = 256x256x105mm³, TE/TR = 1.5/5ms, partial Fourier = 0.75x0.75, SENSE = 3x1.5.

Results

Figure 2 shows the result of the simulations for the different shim approaches. For the whole brain homogeneity, a DYN shim is clearly a large improvement over a ROI shim. For the voxel homogeneity, a DYN shim leads to a comparable shim as a ROI shim, while a BRAIN shim is clearly much worse. This is also demonstrated by Figure 3. With the ROI shim, the B₀ field over the navigator volume was compromised, and the fat-selection failed in parts of the brain. This was improved with the DYN shim, yielding a standard deviation of the shimmed B₀ map slightly worse than the BRAIN shim, but sufficient to ensure fat-selection over the brain. For the voxel, similar linewidths were observed with the ROI shim as with the DYN shim, while the BRAIN shim clearly yielded a broader linewidth.

Figure 4 shows the result of an experiment with motion. With motion and no correction all peaks experienced broadening after motion, the water suppression was reduced and the CRLB of the major metabolites doubled. With motion and correction, the spectral quality was similar to the experiment without motion (slightly larger line width after motion, similar water suppression, similar CRLB’s), and the metabolites are clearly resolved.

Discussion

We expect the dynamic shim strategy to also benefit EPI based navigators that are prone to distortions, such as vNAVs [1]. After motion, a degradation of the line-width can occur, even though the voxel position is updated. We will further investigate real-time updating of the shim after motion.

The shimming approach for both local and global homogeneity was here used to improve global homogeneity in the brain for a motion navigator, but might also be employed in other cases where both local and global homogeneity are important, for example with water and lipid suppression.

Conclusion

Acknowledgements

References

1. Hess AT, Tisdall MD, Andronesi OC, Meintjes EM, van der Kouwe AJW. Real-time motion and B0 corrected single voxel spectroscopy using volumetric navigators. Magn Reson Med. 2011 Aug;66(2):314–23.

2. Marsman A, Boer VO, Andersen M, Petersen ET. Real-time frequency and motion corrected Hadamard encoded spectral editing (CHASE). In: Proc Intl Soc Mag Reson Med. 2017. p. 5493.

3. Andersen, Boer, Marsman, Petersen, “A Generalized Prospective Motion Correction Framework for Improved Spectroscopy, Structural and Angiographic Imaging”, ISMRM 2017 #3934

4. D. Gallichan, J. P. Marques, and R. Gruetter, “Retrospective correction of involuntary microscopic head movement using highly accelerated fat image navigators (3D FatNavs) at 7T,” Magn. Reson. Med., vol. 75, no. 3, pp. 1030–1039, 2016.