Natalia Gudino1, Stephen J Dodd1, Steve Li2, and Jeff H Duyn1
1LFMI, NINDS, National Institutes of Health, Bethesda, MD, United States, 2MIB, NIMH, National Institutes of Health, Bethesda, MD, United States
Synopsis
Optically controlled on-coil
amplifiers have been presented for the practical implementation of pTx system at
ultra-high field MRI. We present a new prototype that can transmit power to excite 1H and X-nuclei. We show bench and MRI data
with a dual-tuned on-coil amplifier implementation for 1H and 31P
excitation at 7T. We expect this technology can allow a simpler and more versatile
implementation of a multinuclear multichannel hardware compared to the traditional
multinuclear setup based on 50 Ω broadband voltage mode amplification.
Introduction
Human multinuclear
studies at high field have the potential to become powerful diagnostic tools. Dual-tuned
multichannel excitation and reception have been proposed to increased SNR and B1+
homogeneity in multinuclear imaging1,2. To simplify the
implementation of multinuclear multichannel hardware, we present an optically
controlled dual-tuned on-coil current-mode amplifier for 1H and X-nuclei
excitation. On-coil amplification has been proposed for the practical implementation
of 1H parallel transmit (pTx) systems3-5. We expect this new prototype will allow a
more flexible multinuclear multichannel Tx hardware by eliminating
the need of multiple cable traps, matching networks and coaxial connections needed
at each resonance frequency.Methods
The dual-tuned
amplifier received the optical carrier signals through a broadband optical RX interface
designed previously to control on-coil amplifiers for 1H excitation4,5, the input circuitry for the preamplification and Current-Mode Class-D
(CMCD) amplification stage6 were designed to maximize the gate-source
voltage to fully switched ON the FETs in both stages at 1H and X-nuclei frequencies
through dual-resonance LC networks (Fig. 1). For this implementation the
networks were tuned for 1H (297.2 MHz) and 31P (120.3
MHz) excitation. Initial network component values were selected based on SPICE
simulations (Altium Designer). A path for harmonics currents, generated from
the switch-mode amplification, was provided by an output filter. The dual-resonance amplifier was connected directly (not 50 Ω impedance matching) to a dual-tuned loop. Proton and phosphorus carrier signals were connected to the input ports of a hybrid combiner the output of which was connected to the RF signal input of the optical interface box 4,5. Carrier signals were transmitted optically (not simultaneously)
through a single fiber to the dual-tuned amplifier. Coil current amplitude and
harmonic content was measured for both excitation frequencies with a calibrated
probe coupled to the Tx coil and connected to a high-speed oscilloscope (2.5
GHz Infinium, Keysight Technologies). The performance of the dual-tuned on-coil
amplifier was compared to a previous 1H-tuned on-coil prototype. A
preliminary multinuclear MR experiment was performed in a 7 T MRI scanner
(Siemens, Erlangen). Carrier signals from the scanner control were
connected to the in-house optically interface located in the scanner
electronics room. Multinuclear excitation was performed with the new amplifier
and Tx coil loaded with a 31P rich solution (50 mM potassium phosphate).
Proton and phosphorous signals were detected with surface loops tuned to the corresponding
frequencies (120.3 MHz and 297.2 MHz). The Rx coils were connected to a 31P
/1H interface box (Quality Electrodynamics, Ohio) plugged to the
patient table in the 7 T scanner. A 1H
localization image was acquired (5
ms TE, 20 ms TR, 192 x 192 matrix size and 8 mm slice thickness) after
which spectroscopy data was acquired with a Chemical Shift Imaging (CSI) sequence (0.5 us RF hard
pulse, 3 s TR, 16 x 16 matrix size, 200 mm x 200 mm FOV, 10 average).Results
Figure 2 shows simulated
frequency response of the dual-resonance LC network (CMCD gate circuit) by
sweeping a single component value at a time. Figure 3 shows the tuning
of the dual-tuned, not 50 Ω matched,
loop and corresponding load impedance for both frequencies. Coil current vs bias
voltage of the amplifier power stage (VDD) for both nuclei is shown
in Figure 4. In this setup, maximum power delivered to the coil was 46
W and 81 W for 31P and 1H respectively. Total harmonic
distortion (THD) values, estimated from the FFT of the coil current at both
frequencies, were THD ~ 1.2 % and THD~4.8 % for the 1H and 31P
current respectively. No degradation in performance was observed for 1H
excitation as shown from the power (delivered to the coil) measurement performed
with the dual-tuned amplifier and a single-tuned 1H prototype (Fig. 4). Figure
5 shows the result of the multi-resonant MRI experiment, 31P
peak (left), CSI image (center) and metabolite image (right) were
successfully obtained with this preliminary setup.Discussion
The dual-tuned
on-coil Tx setup allows to have multinuclear excitation without the need
of additional matching networks, cable traps, and coaxial connections.
Therefore, this approach can simplify the implementation of a multinuclear multichannel
setup compared to those built based on 50 Ω broadband
voltage amplifiers typically found in MRI systems with multinuclear
capability. We presented preliminary data for 31P and 1H excitation,
however the dual-tuned amplifier can be tuned to excite other X-nuclei (23Na,
13C, 19F etc.). In addition, the design of the gate circuit
could be extended to allow more than 2 nuclei excitation. In this work we
performed 31P and 1H excitation through a single dual-tuned
Tx coil, however the design can allow the combination of the amplifier with
nested coils as presented elsewhere to improve efficiency2. Finally, in a 1H
pTx system built with this technology each channel can be simply “transformed” by
the control to excite a different nucleus as necessary for different
applications. Therefore, we consider this technology will allow the
implementation of a more versatile Tx hardware to be able to excite both 1H and X-nuclei.Acknowledgements
References
1- Ianniello
C, et al. Magn
Reson Med. 2019 82(4):1566-1575. 2- Brown R, et al. Neuroimage. 2016 124(Pt A):602-611. 3- Gudino N, et al. Magn Reson Med. 2013 70:276-89. 4- Gudino N, et al. Magn
Reson Med. 2016 76:340-9. 5- Gudino N, et al. Magn Reson Med. 2018
79(5):2833-2841. 6- Kobayashi
H, et al. IEEE Trans Microwave Theory Tech 2001;49:2480–2485.