Johanna Dorst1, Loreen Ruhm1, Nikolai Avdievich1, Wolfgang Bogner2, and Anke Henning1,3
1High-Field MR Center, Max Planck Institute for Biological Cybernetics, Tübingen, Germany, 2High-Field MR Center, Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Austria, 3Advanced Imaging Research Center, UT Southwestern Medical Center, Dallas, TX, United States
Synopsis
For
phosphorus single voxel MR spectroscopy, ISIS is often the method of choice,
but it requires an 8 steps encoding scheme and post-acquisition signal
combination which makes it prone to motion. Therefore, a SPECIAL-semiLASER
sequence for 31P MRS was implemented in this study, which only needs
two encoding steps for 3D localization. Spectra from the human brain acquired
at 9.4T with both sequences are compared in terms of absolute SNR and SNR
efficiency. Absolute SNR as well as SNR efficiency is lower for
SPECIAL-semiLASER, but calculated concentrations are comparable.
Introduction
For
phosphorus single voxel MR Spectroscopy, ISIS1 is a commonly used
localization technique. This technique offers good localization with high SNR
due to short echo times, but is susceptible to stochastic sources of variation
since 8 scans are required to obtain one three-dimensional spectrum2.
Therefore, in this study, SPECIAL-semiLASER3,4 was implemented and
tested for 31P MRS. SPECIAL-semiLASER requires a longer minimum echo
time, but needs only 2 scans for localization. For both sequences, spectra were
acquired from healthy resting human brains at 9.4T, and compared in terms of
absolute SNR and SNR efficiency.Methods
All
measurements were performed on a 9.4T whole-body MRI scanner (Siemens) using a
home-built double-tuned 31P/1H human head array5,
in which the entire power was only applied to the three bottom 31P
coil elements.
In
this study, the two multi-shot sequences ISIS (TE: 0.425ms) and SPECIAL-semiLASER
(TE: 17ms) were compared (Fig. 1). Both sequences use GOIA-W(16,4) pulses6
(5ms, gradient amplitude 25mT/m, inversion bandwidth 12.9kHz for 3cm slices and
21.4kHz for 5cm slices7) for inversion. For excitation, ISIS uses a frequency
selective rectangular pulse (0.25ms, 4.5kHz) and SPECIAL-semiLASER a slice-selective
Hamming-windowed sinc pulse (1.5ms, TBWP 6.5). No phase cycling or crusher
gradients were implemented.
Absolute
SNR was compared with a repetition time of 25s (≥5xT18)
to guarantee full relaxation. 64 averages were acquired for one volunteer (28
years). For SNR efficiency comparison, 8 healthy volunteers (28 ± 2.8 years, 4 female,
4 male) were recruited. The measurement time was 10 minutes for both sequences,
which resulted in TRISIS: 5s and 120 averages, and, due to SAR
limitations, in TRSPECIAL-semiLASER: 6s and 100 averages. Volumes of
interest of 5x5x5cm3 and 3x3x3cm3 were chosen in the
occipital lobe.
Raw data were processed (averaging, truncation, SVD coil combination based on
PCr9, zero-order phase correction, no filtering) by an in-house
written MATLAB software. SNR was calculated in the frequency domain as ratio
between the metabolite peak heights and the spectral noise between 15 ppm and
30 ppm. For quantification, data were fitted using an in-house written MATLAB
implementation of AMARES10. The model function was optimized in a
way to match the Voigt line shape. Obtained concentrations were corrected for T18
and T2 relaxation (recently measured at our institute).Results
Figure 2 shows spectra obtained
for the absolute SNR measurement in a 5x5x5cm3 voxel (inset) for a)
ISIS and b) SPECIAL-semiLASER localization. SNR and FWHM values are summarized
in Table 1. For phosphomonoesters and –diesters and inorganic phosphate, SNR of
SPECIAL-semiLASER is on average 35% lower, for PCr 45% lower, and for ATP 60%
lower in comparison to ISIS.
Spectra from the SNR efficiency
measurements and corresponding AMARES fits are shown in Figure 3 for voxel
sizes of 5x5x5cm3 and 3x3x3cm3 in the upper and lower
row, respectively, and the two sequences ISIS (left column) and
SPECIAL-semiLASER (right column). SNR of phosphomonoesters, Pi and
phosphodiesters of SPECIAL-semiLASER spectra is 30% lower than in ISIS spectra,
PCr is 20% lower, and ATP is 65% lower. AMARES quantification results for ISIS
(a) and SPECIAL-semiLASER (b) are reported in Figure 4. Concentrations are
given as ratio to γ-ATP. The results are corrected for relaxation, but not for pulse profile
distortions.Discussion
The amount of reduction in
absolute SNR for SPECIAL-semiLASER in comparison to ISIS corresponds to T2
relaxation losses11,12 due to a longer minimum TE in
SPECIAL-semiLASER. The linewidths are comparable for all metabolites except PC.
The PC signal seems broadened when detected with SPECIAL-semiLASER, but it is
difficult to detect due to low SNR which may render the FWHM estimate
inaccurate. Another error source might be an overlap with 2,3-diphosphoglycerate
from blood12.
SNR efficiency of the two
sequences is not only limited by T2 losses, but also by SAR, and
therefore by longer TR and less averages per unit time for SPECIAL-semiLASER. Frequency
resolution in both sequences enables the quantification of 8 metabolites. The β-ATP
peak is strongly attenuated in ISIS spectra and not visible in
SPECIAL-semiLASER due to limited excitation bandwidths.
Even though a direct comparison of quantification results to
other studies is not trivial since we report concentrations not corrected for pulse
profile distortions,
concentrations measured with ISIS are similar to previously calculated
concentrations measured at 7T with a non-localizing FID sequence13. For
tNAD, our measured concentrations are much higher than that previous 7T data
propose13. Concentrations from SPECIAL-semiLASER spectra are
comparable to ones obtained from ISIS, but show a wider dispersion since
spectra suffer from lower SNR and fit reproducibility. PCr concentration is
remarkably higher than in ISIS spectra, which might be due to a worse
localization efficiency, and, therefore possible muscle contamination.
We did not determine significant differences between different voxel sizes.Conclusion
This
study presents a comparison of ISIS and SPECIAL-semiLASER localization for 31P
MRS in the human brain at 9.4T. Though SPECIAL-semiLASER requires longer TE and
TR, and therefore SNR is lower in comparison to ISIS, good spectral quality can
be obtained. Fitting results show good reproducibility. Measured metabolite
concentrations are comparable for both sequences, except a higher PCr
concentration in SPECIAL-semiLASER.Acknowledgements
Funding
by the European Union (ERC Starting Grant, SYNAPLAST MR, Grant Number: 679927) and
by the Cancer Prevention and Research Institute of Texas (CPRIT) (Grant Number:
RR180056) is gratefully acknowledged.References
1Bogner W, Chmelik M, Andronesi OC, et al.
In Vivo 31P Spectroscopy by Fully Adiabatic Extended Image Selected In
Vivo Spectroscopy: A Comparison Between 3T and 7T. Magn Reson Med.
2011;66(4):923-930.
2Keevil SF. Spatial localization in nuclear
magnetic resonance spectroscopy. Phys Med Biol. 2006;51:R579-R636.
3Steinseifer IK, Mekle R, Gruetter R, et
al. Implementation of GOIA-Wurst Pulse in a SPECIAL Localization
Sequence at 7 T. In Proceedings of the Annual Meeting of ISMRM, Montreal,
Canada, 2011. Abstract 1431.
4Fuchs A, Luttje M, Boesiger P, et al. SPECIAL
Semi-LASER with Lipid Artifact Compensation for 1H MRS at 7T. Magn Reson Med.
2013;69:603-612.
5Avdievich NI, Ruhm L, Dorst J, et al. Double-Tuned
31P/1H Human Head Array with High Performance at Both Frequencies for Chemical
Shift Spectroscopic Imaging (CSI) at 9.4T. Proc. Intl. Soc. Mag. Reson Med.
2019;26:3994.
6Andronesi OC, Ramadan S, Ratai E, et al.
Spectroscopic imaging with improved gradient modulated constant adiabaticity
pulses on high-field clinical scanners. J Magn Reson. 2010;203(2):283-293.
7Chmelík M, Just Kurková I, Gruber S, et al.
Fully adiabatic 31P 2D-CSI with reduced chemical shift displacement error at 7T
– GOIA-1D-ISIS/2D-CSI. Magn Reson Med. 2013;69(5):1233-1244.
8Pohmann R, Raju S, Scheffler K. T1 values of
phosphorus metabolites in the human visual cortex at 9.4T. Proc. Intl. Soc.
Mag. Reson Med. 2018;26:3994.
9Bydder M, Hamilton G, Yokoo T, et al. Optimal
phased-array combination for spectroscopy. Magn Reson Imaging.
2008;26(6):847-850.
10Vanhamme L, van den Boogaart A, van Huffel S.
Improved Method for Accurate and Efficient Quantification of MRS Data with USE
of Prior Knowledge. J Magn Reson.
1997;129:35-43.
11Lei H, Zhu X-H, Zhang X-L, et al. In
Vivo 31P Magnetic Resonance Spectroscopy of the Human Brain at 7T: An Initial
Experience. Magn Reson Med. 2003;49:199-205.
12Van der Kemp WJM, Klomp DWJ, Wijnen JP. 31P T2
of Phosphomonoesters, Phosphodiesters, and Inorganic Phasphate in the Human
Brain at 7T. Magn Reson Med. 2018;80:29-35.
13Ren J, Sherry AD, Malloy CR. 31P-MRS of healthy
human brain: ATP synthesis, metabolite concentrations, pH, and T1 relaxation
times. NMR Biomed. 2015;28:1455-1462.
14Nassirpour S, Chang P, Henning A. High and
ultra-high resolution metabolite mapping of the human brain using 1H FID MRSI
at 9.4T. 2018;168:211-212.