Xinqiang Yan1,2, John C. Gore1,2,3, and William A. Grissom1,2,3
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, 2Radiology, Vanderbilt University, Nashville, TN, United States, 3Biomedical Engineering, Vanderbilt University, Nashville, TN, United States
Synopsis
Dipole
and microstrip coils produce different and somewhat complementary B1 patterns
and hybrid E-field distributions. Based this observation, we developed a
16-channel transmit/receive array for 7T head imaging by interleaving dipole
and microstrip elements. Mutual coupling among any elements is <-14 dB
without including any other decoupling. Compared with 8-channel microstrip-only
and dipole-only arrays, the proposed 16-ch dipole+microstrip array has a higher
SNR gain and lower g-factor. No
decoupling treatment is needed for the mixed dipole and microstrip array, so it
can be used as a flexible transceiver array at ultrahigh field.Introduction
Loop,
microstrip
1, 2 and dipole
3, 4 transmit/receive arrays
are the most common RF coil configurations at 7T and higher. Dipole and microstrip
coils produce different and somewhat complementary B
1 patterns
3 and hybrid
E-field distributions. Based this observation, we developed a 16-channel transmit/receive
array for 7T head imaging by interleaving dipole and microstrip elements. Mutual
coupling among any elements is <-14 dB without including any other decoupling.
Compared with 8-channel microstrip-only and dipole-only arrays, the proposed
16-ch dipole+microstrip array has a higher SNR gain and lower g-factor.
Methods
Coil
Construction: A 16-channel array with 8 dipole and 8 microstrip elements was
constructed on a cylindrical acrylic tube (diameter 24.1 cm and length 25.4
cm), as shown in Fig. 1. The dipole elements and microstrip elements were placed
alternately. Each microstrip element was a typical half-wavelength resonator with
a Teflon substrate (18×4×1.27 cm3) 1. The width of the strip
conductors and grounds were 1 cm and 4 cm, respectively. One end of each microstrip
resonator was terminated with a 3.3 pF capacitor and the other end was
terminated with trimmer capacitors for tuning and matching. Each dipole was electrically
shortened by two lumped inductors and matched by a trimmer capacitor. The width
and length of the dipole conductors were 0.75 cm and 21 cm, respectively. All
dipole and microstrip elements were used for transmit and receive, and were
tuned to 298 MHz and matched to 50 ohm (S11 better than -25 dB).
Floated bazooka baluns were used for all elements to avoid “cable resonance”.
Bench
test and MR experiments: The S-parameters
of the 16-ch array loaded with a cylindrical water phantom (diameter 15 cm and height
20 cm) were measured with an Agilent 5071C network analyzer. GRE images of the
water phantom were obtained with a human 7T Philips Achieva scanner (Philips
Healthcare, Cleveland, Ohio, USA). MR images were also acquired using each
channel individually. The parameters of the GRE sequence were: FA=250,
TR/TE=500/10ms, FOV=180×180mm2, matrix=192×192, thickness=5mm. The SNR
on root-sum-of-squares (RSOS)-combined GRE images was calculated as: signal/std(noise)*0.655.
G-factor maps were calculated using Pruessmann’s
method 5. B1+ profiles of each channel were measured with the DREAM method 6.
Results and Discussions
Fig. 2A shows the S
11
and S
21 plots of adjacent dipole-microstrip, dipole-dipole and microstrip-microstrip
elements. Fig. 2B shows the S-parameter matrix of the 16-ch array loaded with the
water phantom. The worst isolation between any two elements was better than -14
dB, indicating excellent decoupling. In our experience, the coupling between most
channels will in practice be lower than -20 dB in heavier loading cases, e.g., with
a human head. Fig. 3 shows the GRE images
and B
1+ maps (both magnitude and phase) of the individual channels. As expected
from the S-parameter measurements, each channel produces quite distinct images
and B
1 profiles. Due to the relatively confined nature of EM fields from the
microstrips, their B
1 fields changed little when the dipole elements were placed
adjacent to them. The symmetry and deep penetration of the dipole elements’ B
1
fields were also maintained when they were placed next to the microstrip
elements, as shown in Fig. 3. However, their imaging coverage at surface areas
was slightly limited by the presence of the microstrip elements. Fig. 4 and
Fig. 5 show the calculated SNR and g-factor maps in a central slice of the
imaged phantom. Both SNR and g-factor are improved using the new hybrid design.
Conclusion
Interleaving
dipole and microstrip elements provides a higher overall SNR and a better
g-factor when compared to dipole-only or microstrip-only arrays. Note that the mixed
array does not need preamp decoupling and can also be used as a transmit-only
array. The diverse B
1+ field patterns generated by this configuration may be
advantageous for RF shimming. No decoupling treatment is needed for the mixed dipole
and microstrip array, so it can be used as a flexible transceiver array at
ultrahigh field.
Acknowledgements
This
work was supported by NIH R01 EB016695.References
[1] G. Adriany, et al, MRM, 53: 434-445 (2005). [2] B. Wu, et al, IEEE TMI. 31: 183-91 (2012). [3] A.J.
Raaijmakers, et al, MRM. 66: 1488-1497 (2011). [4] G. C. Wiggins, et al, ISMRM.
541 (2012). [5] K. P. Pruessmann, et al, MRM. 42: 952-962 (1999). [6] K. Nehrke,
et al, MRM 68:1517–26, (2012).