Benson Yang1, Fred Tam1, Pei-Shan Wei1, Clare E McElcheran2, and Simon J Graham1,3
1Physical Sciences, Sunnybrook Research Institute, Toronto, ON, Canada, 2Baylis Medical, Missisauga, ON, Canada, 3Medical Biophysics, University of Toronto, Toronto, ON, Canada
Synopsis
Interest
in parallel transmission (pTx) continues to grow with many research groups
investigating methods to increase channel count and applications on commercial
MRI systems. It can be challenging, however, to integrate pTx hardware onto
existing systems without disrupting normal operation. The present work
successfully interposes a four-channel pTx system on an existing 3 T Siemens
Prisma system and performs validation to demonstrate: (1) four-channel
radiofrequency (RF) shimming; and (2) reduced RF heating in an electrically
conductive implant.
Introduction
Interest in parallel transmission (pTx) continues to
grow with many research groups investigating methods to increase channel count and
applications on commercial MRI systems. It can be challenging, however, to
integrate pTx hardware onto existing systems without disrupting normal
operation. The present work successfully interposes a four-channel pTx system
on an existing 3 T Siemens Prisma system. PTx system validation was performed
to demonstrate: (1) four-channel radiofrequency (RF) shimming; and (2) reduced
RF heating in an electrically conductive implant.Methods
The
four-channel prototype pTx system has a peak output power of 4 kW and
integrates with the Siemens MRI system as shown in Fig 1. A custom four-channel
adaptor module (Stark Contrast, Erlangen, GER) equipped with a Tim4G plug connects
to the MRI system and provides access to the existing transmit/receive (TxRx)
pathway. The adaptor module also has connections for a four-channel TxRx coil.
The existing RF transmit signal (on the adaptor module) is connected to a 40 dB
directional coupler (C11082-13, Werlatone, Patterson, NY); the existing RF
signal power is dissipated through a load and the small-signal output is further
attenuated by 30 dB (8322, Bird Electronics, Cleveland, OH) before being input
into the RF modulator (CGP-128-4C, CPC, Hauppauge, NY). The RF modulator splits
the input into four independent signals, with adjustable amplitude and phase
via a web interface, that each connect to a separate 60 dB RF power amplifier
(RFPA) (BT01000-AlphaSA, Tomco Technologies, Stepney, AUS) and feed back into
the adaptor module for TxRx coil transmission. The RFPAs are blanked by
converting the optical signal from the MRI system to an RFPA-compatible
transistor-transistor logic signal.
Prototype pTx system verification included an RF shim
test and an RF heating test. For the RF shim test, the phases on the RF
modulator were adjusted to maintain the optimal 90⁰ offset between
neighbouring channels during fast low angle shot (FLASH) MRI of a uniform cylindrical
phantom (TR/TE/FA=100 ms/5 ms/20⁰). An image was acquired with sub-optimal RF shim for
comparison (50⁰ phase shift to
Channel 1). For the RF heating test, an initial experiment was conducted
to verify previous simulation studies that calculated optimal amplitude and
phase settings to minimize heating in targeted regions of electrically
conductive implants1,2. Turbo spin echo (TSE) scans (TR/TE/FA = 516
ms/6.7 ms/150⁰, 3m 45s scan time) were acquired of a head-shaped gel
phantom with an inserted fibre-optic temperature sensor (Opsens Inc., Quebec
City, QC) placed at the exposed tip of implanted insulted copper wire approximating
a realistic lead trajectory for a deep brain stimulation device. Amplitude and
phase adjustments were applied (based on simulation) to achieve pTx operation
in two modes: (1) for localized heating; and (2) for suppressed heating
(scientific and technical details are reported in another submission).Results
Fig. 2 shows images for (a) optimal and (b)
sub-optimal RF shim settings. In the optimal condition, the mean and standard
deviation of the MRI signal were 91 and 14, respectively, in the central region-of-interest
(ROI, green outline). Image uniformity deteriorated with the sub-optimal RF
shim such that the mean decreased by 36% and the standard deviation increased
by 40% in the ROI. During the RF heating
tests, imaging in mode 1 raised the temperature near the tip of the wire by 2.6⁰C (from 22.8⁰C to 25.4⁰C), whereas mode 2 produced only 1⁰C elevation (from 22.9⁰C to 23.9⁰C). Fig. 3 summarizes the RF
heating test results and the pTx modulation settings (accounting for phase
variation at the four-channel TxRx coil) to maintain 90⁰ between neighbouring
channels for mode 1 and optimized amplitude and phase settings for mode 2.Discussion
The optimal RF shim image (Fig. 2a) agrees reasonably
well with previous pTx results3,4. As expected, sub-optimal shim
setting on channel 1 degraded signal strength and image uniformity.
Furthermore, the pTx platform demonstrated a ~62% reduction in RF heating that
is consistent with previous experimental results in the laboratory with
different implant geometries2Conclusion and Future Work
A prototype four-channel pTx platform has been integrated
successfully with a 3 T Siemens Prisma MRI system. The modular system design
offers the flexibility and ease to increase channel count and replace hardware
components. Future work will include more advanced experiments and hardware
optimization.Acknowledgements
1. Canada Foundation for Innovation
2. Natural Sciences and Engineering Research Council of Canada
References
[1]
McElcheran et al., “Parallel Transmission for Heating
Reduction in Realistic Deep Brain Stimulation Lead Trajectories” ISMRM (2017)
[2] McElcheran
et al., “Parallel Radiofrequency Transmission at 3 Tesla to Improve Safety in
Bilateral Implanted Wires in a Heterogeneous Model” MRM (2017)
[3] Ullmann
et al., “Experimental Analysis of Parallel Excitation Using Dedicated Coil
Setups and Simultaneous RF Transmission on Multiple Channels”, MRM (2005)
[4] Stang et
al., “An Extensible Transmit Array System using Vector Modulation and
Measurement”, ISMRM (2008)