Sung-Min Sohn1, J. Thomas Vaughan1, Michael Garwood1, and Djaudat Idiyatullin1
1Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States
Synopsis
This is the
first demonstration of in vivo human MR imaging with simultaneous transmit and
receive using continuous mode SWIFT at 4T. Due
to a large RF power difference between Tx and Rx working at the same frequency,
the difficulties to obtain the high and stable Tx/Rx isolation, and the
sensitivity of the Tx/Rx isolation to the loading conditions, in vivo images using the simultaneous RF
pulse transmission and signal acquisition have not been reported. This work
proposed the simultaneous Tx/Rx system with highly minimized effects from
variation of coil loading, which allowed us to acquire the first in vivo images with continuous SWIFT at
4T. Purpose
To obtain reliable in vivo human head images with the simultaneous transmit and
receive system using SWIFT (SWeep Imaging with Fourier Transformation) in
continuous mode 1,2.
Introduction
The development of methods to allow simultaneous RF excitation and signal
acquisition has potential to significantly benefit MR imaging, particularly for
reducing RF power requirements and for imaging fast relaxing spins. Ideally
such capability would provide an optimal usage of RF transmitter peak power (lowest
peak power and low SAR3) and optimal signal sampling (high SNR). However,
the practical realization of such a system is challenging due to a large RF
power difference between Tx and Rx channels working at the same frequency and subsequently
the difficulties to obtain the necessary Tx/Rx isolation. The isolation required
to avoid saturation of the receiver can be achieved by compensation of leakage
signal by slightly mismatched RF coils1,2 or by destructive interference decoupling
using the extra Tx coil4.
Both methods allowed the acquisition of ex
vivo images. However, in vivo applications
using these methods have not been reported because the methods are sensitive to
variation of the coil loading conditions during MRI examinations of living
subjects. Here, we propose the simultaneous Tx/Rx system with highly minimized
effects from variation of coil loading, which allowed us to acquire the first in vivo images with continuous SWIFT at
4T.
Methods
The proposed system applied general techniques
used in wireless communication (Fig.1) to cancel out the mutual
coupled Tx leakage signal5-7. The system uses sampled and
manipulated Tx signal (SC(G,P)) to compensate the leakage (Sleak)
in Rx signal8. The critical part in our system is the RF power
sampling block, which extracts the signal Sc being proportional to S*Tx signal.
Thus, variation in loading conditions affecting the S*Tx signal is reflected also
in compensating signal, and therefore, the system minimizes the effect from the
variations of coil loading. The gain and phase of the sampled Tx signal are
optimized in the coil-independent active module to maximize the Tx/Rx isolation
without deformation of the quality of the coil.
Considering the
above-mentioned aspects, the passive (RF power sampling) and active feedback module
have been designed and fabricated as shown in fig.2. The passive module uses two
couplers with the high directivity and some passive components to adjust proportionality
to the reflected signal9. The active module consists of a power
combiner, gain controller (RVA-3000, Mini-circuits, NY), phase controller
(SPHSA-251, Mini-circuit, NY), and a low noise amplifier (ZX60-P103LN,
Mini-circuit, NY). The passive module was optimized for each tested coils (TEM,
birdcage, and loop) and the active module is remotely controlled from the MR
control room.
Images were acquired using a MRI
scanner equipped with a 4T (human)
magnet and a Varian DirectDrive
console. Acquisition parameters: sweep frequency span = 32.5 kHz, 128000
views (spokes in k-space), 256 complex points per view, the diameter of field
of view was 44 cm, isotropic resolution of 1.7 mm, and total acquisition time
10 min. The residual leakage signal was subtracted before image processing
using the algorithm described previously2.
Results
The isolation versus loading variation is
shown in Fig.3 (top), which was
obtained with a TEM volume coil connected to the conventional hybrid coupler or
the high directivity coupler. This result shows the load-insensitivity of the
power sampling block (Fig.1) in this work. Fig.3 (bottom) represents an example
of the measured isolation performances with and without the proposed compensating
system for 4T application (fLarmor = 169.3 MHz) with a loop coil
(radius: 9 cm). Stable isolation less than -60 db is obtained and it is enough
to keep the signal below the saturation threshold of the receiver system. Fig.4
represents MR images acquired with the proposed system and the continuous SWIFT
using the ultra-low RF peak power of 50
mW.
Discussion
and Conclusion
This work describes the
first MR imaging of the human head with a simultaneous Tx/Rx feedback system
and continuous SWIFT. Stable and high isolation have been accomplished with
insensitivity to loading variations taking
into account many factors: coil’s
properties, mutual coupling factors, transmission method (transceiver or
transmit/receive separate mode), the number of coil channel, etc. In our
opinion, this simultaneous Tx/Rx feedback system is more straightforward as
compared to other known methods, for example, the method using frequency
modulation10. To improve the reliability of the system, automatic isolation control will be applied for the
next step. Due to RF power
distribution in time, the continuous SWIFT uses low RF peak power. Therefore, such
system might be useful for the next-generation integrated and portable MR scanners.
Acknowledgements
This study was supported by NIH grants:
P41 EB015894, S10 RR023730, S10 RR027290, and R24-MH105998-01.References
[1] D. Idiyatullin, et al., 19th ISMRM. p.382, 2011. [2] D. Idiyatullin, et al., J Magn Reson 220, (2012). [3] D.
Idiyatullin, et al., J Magn Reson 193, (2008). [4] A.
Özen, et al., Magnetic Resonance Materials in Physics, Biology and Medicine (2015). [5]
D. Bharadia, et al., SIGCOMM Comput. Commun. Rev. 43, (2013). [6] Choul-Young
K, et al., IEEE MTT-S (2006). [7] Min-su K, et al., IEEE
Trans. On Industrial Electronics (2014). [8] Sung-Min Sohn, et al., IEEE MTT-S (2014). [9]
Scott M, et al., US9097769 B2. [10]
D.O.Brunner, et al. 19th ISMRM. p.625, 2011.