Chang-Hoon Choi1, Airat Galiamov1, Suk-Min Hong1, Jörg Felder1, Wieland A. Worthoff1, and N. Jon Shah1,2,3,4
1INM-4, Forschungszentrum Juelich, Juelich, Germany, 2INM-11, Forschungszentrum Juelich, Juelich, Germany, 3JARA-BRAIN-Translational Medicine, Aachen, Germany, 4Department of Neurology, RWTH Aachen University, Aachen, Germany
Synopsis
X-nuclei
MR offers unique access to important metabolic information in tissues. Multi-tuned
coils are required for the X-nuclei measurements, but designing a
well-performing coil is challenging. In this study, we present our novel design
and performance evaluation of an interchangeable, twin, double-tuned, head coil
array for 1H/23Na MR imaging and 1H/31P
MR spectroscopy at 7 T. The outer proton array was built using an alternatingly positioned 4-channel dipole antenna and a 4-channel microstrip transmission
line array to improve decoupling. The inner 8-channel X-nuclei loop arrays,
orthogonal to the 1H, were designed identically to enable conventionally switching between 23Na and 31P.
Introduction
Investigating
X-nuclei (non-proton nuclei), e.g. sodium-23 (23Na) and
phosphorus-31 (31P) has been of great interest recently with
increasing the availability of ultra-high field systems. X-nuclei offer unique access
to complementary and important biochemical and metabolic information in tissues,
and alterations in these nuclei are strongly related to a variety of
pathological and neurodegenerative conditions.1,2 However, studies using
X-nuclei suffer from their intrinsically lower natural abundance and MR
sensitivity compared to those of the protons, making any improvement in signal-to-noise
ratio (SNR) desirable. B0 shimming with X-nuclei is also
problematic; therefore, the concurrent acquisition of proton imaging is
beneficial. The nuclear Overhauser effect (NOE) can also be used to boost SNR
of, for example, 31P metabolites, and to increase spectral fitting
accuracy.3 Multi-resonant RF coils are, therefore, likely to be used
to attain the requirements, but designing a well-performing multi-tuned coil is
challenging and a number of design approaches have been introduced previously.4-6
In this work, we present our novel design and performance evaluation of an interchangeable,
twin, double-tuned, head coil for 1H/23Na MR imaging and 1H/31P
MR spectroscopy at 7 T. Material and methods
As shown in Figure 1, our transmit/receive
coil assembly consisted of a proton antenna array and an X-nuclei coil array which
included either an 8-channel 23Na or a 31P loop array. The
proton array was located in an outer shell (diameter/length = 300/220 mm) and the
X-nuclei coil array (diameter/length = 260/170 mm) was placed inside the proton
array. The proton array included 4-channel dipole antennas and 4-channel microstrip
transmission lines (MTLs) - which were selected as they offer benefits in B1
penetration and symmetric excitation at ultra-high field.7,8 The
dipole and MTLs were placed alternatively one next to another in order to
improve decoupling.9 Both the 23Na and 31P
array were individually tuned, but the design was identical, and both X-nuclei
coil arrays could share the outer 1H array, and were exchangeable. In
order to avoid any loss, the dipole or MTL element was located across the
middle of each X-nuclei loop. This concept was chosen since the loop and either
the dipole or the MTL are penetrated perpendicularly as they are geometrically
isolated relating to 1H elements without requirement10 of
lossy decoupling components, e.g. traps. The eight X-nuclei loops were
capacitively decoupled between next neighbours.11
In order to evaluate the
performance of the proposed coil/antenna array, we first measured the S-parameters
of each nucleus, and then carried out MR imaging and spectroscopy on a 7 T
Siemens Terra. In the MR experiments, standard MR sequences were used, and the
parameters were for 1H: 2D FLASH, TR = 8.6 ms, TE = 3.69 ms,
Averages = 2, resolution = 0.5 × 0.5 × 5 mm3, acquisition time = 18.6 seconds, for 23Na:
3D FLASH, TR = 25 ms, TE = 3.63 ms, 32 averages, resolution = 4 × 4 × 4 mm3, acquisition time = 10:29 minutes and for 31P:
3D CSI, TR = 2000 ms, TE = 0.35 ms, 4 averages, weighted phase encoding scheme,
voxel-of-interest (VOI) = 25 × 25 × 25 mm3, with and without applying NOE, acquisition
time = 6 minutes. A 100 mM NaCl and 100 mM KH2PO4 doped
water phantom was used for 23Na and 31P measurements. As
an initial step, all proton and X-nuclei arrays channels were tested individually. Results
Figure 2 displays an S-parameter
matrix of 1H, 23Na and 31P arrays measured on
the bench using a vector network analyser. It can be seen that the decoupling, particularly
in the proton channel, improved significantly (better than -20 dB) between adjacent
elements. Without the proposed decoupling schemes or circuits, the value was, as
expected, quite poor (around -7 dB).
Figures 3 and 4 show the
proton/sodium images and the 8 × 8 phosphorus spectra (without NOE) overlaid on the associated
proton image acquired at all individual channels. This confirmed that all 24
channels were operating equitably. The SNR increase at the selected VOIs as a
result of NOE enhancement is also shown in Figure 4. Discussion
This study has successfully demonstrated
the novelty and benefits of using the multi-tuned coil array design, and shown
its feasibility for use in 1H/23Na MRI and 1H/31P
MRS. The inner, twin X-nuclei arrays can be flexibly switched between 23Na
and 31P, and potentially to any other nuclei of interest, e.g. 7Li,
13C or 17O. We found a negligible tuning variation of the
proton elements when changing from the 23Na to the 31P
array. The inner, 8-channel loop arrays can also be extended to 16-channels12,
or more, in the z-axis to improve SNR or to shorten the scan time. The proton
elements are independent and are well decoupled from each other, potentially improving
the parallel transmit performance. To further investigate this, several tasks
would also be required, for example, parallel imaging examination, additional
evaluation relating to NOE enhancement3 and implementation of a whitened
singular value decomposition13 spectra combination method. These will be studied in the future.Acknowledgements
No acknowledgement found.References
1. Shah NJ, et
al. Imaging of sodium in the brain: a brief review. NMR Biomed.
2016;29(2):162-174.
2. Bonvento G,
et al. Imaging and spectroscopic approaches to probe brain energy metabolism
dysregulation in neurogenerative diseases. J Cerebr Blood F Met.
2017;37(6):1927-1943.
3. Peeters TH,
et al. 3D 31P MR spectroscopic imaging of the human brain at 3T with a 31P
receive array: an assessment of 1H decoupling, T1 relaxation times, 1H-31P
nuclear Overhauser effects and NAD+. NMR Biomed. 2019.
4. van de Bank
B, et al. Optimized 31P MRS in the human brain at 7T with a dedicated RF coil
setup. NMR Biomed. 2015;28(11):1570-1578.
5. Brown R, et
al. A nested phosphorus and proton coil array for brain magnetic resonance
imaging and spectroscopy. Neuroimage. 2016;124:602-611.
6. Malzacher M,
et al. Feasibility study of a double resonant 8-channel 1H/8-channel 23Na
receive-only head coil at 3 tesla. Magn. Reson. Imaging. 2019;59:97-104.
7. Raaimakers
AJ, et al. Dipole antennas for untrahigh-field body imaging: a comparison with
loop coils. NMR Biomed. 2016;29(9):1122-1130.
8. Wu B, et al.
Shielded microstrip array for 7T human MR imaging. IEEE Trans Med Imaging.
2010;29(1):179-184.
9. Yan X, et al.
A mixed dipole and microstrip transmit/receive array. Proc. Int. Soc. Magn.
Reson. Med. 2016, pp.2140.
10. Hong S-M, et
al. Design of a quadrature 1H/31P coil using bent dipole antenna and
four-channel loop at 3T MRI. IEEE Trans Med Imaging. 2018;37(12):2613-2618.
11. von Morze C, et al. An
eight-channel, nonoverlapping phased array coil with capacitive decoupling for
parallel MRI at 3T. Concepts Magn Reson Part B. 2007;31B(1):37-43.
12. Steensma BR,
et al. An 8-channel Tx/Rx dipole array combined with 16 Rx loops for
high-resolution functional cardiac imaging at 7 T. MAGMA. 2018;31(1):7-18.
13. Rodgers CT,
Robson MD. Receive array magnetic resonance spectroscopy: whitened singular
value decomposition (WSVD) gives optimal Bayesian solution. Magn. Reson. Med.
2010;63:881-891.