Bassem Henin1, Ewald Weber1, and Stuart Crozier1
1ITEE, The University of Queensland, Brisbane, Australia
Synopsis
Magnetic resonance
imaging (MRI) is a very important non-invasive modality for the diagnosis of
knee anatomy and pathology. The use of 7 Tesla systems for knee imaging can
provide higher signal-to-noise ratios (SNR) than lower field systems. In this
work, a novel 8-channel transceive open knee coil for dynamic musculoskeletal
(MSK) MR imaging of the knee is simulated and constructed. The open design
concept will facilitate the diagnostic and functional assessment of knee
injuries and pathology, and allow for imaging of a moving knee.Target audience
Radiofrequency (RF) coil engineers, MR
scientists with an interest in knee imaging at high fields.
Introduction
Magnetic resonance
imaging (MRI) is a very important non-invasive modality for the diagnosis of
knee anatomy and pathology. The use of 7 Tesla systems for knee imaging can
provide higher signal-to-noise ratios (SNR) than lower field systems. In this
work, a novel 8-channel transceive open knee coil for dynamic musculoskeletal
(MSK) MR imaging of the knee is simulated and constructed. The open design
concept will facilitate the diagnostic and functional assessment of knee
injuries and pathology, and allow for imaging of a moving knee. Installing an
RF shield around the coil reduces radiation losses and
unwanted coupling with the patient bed as well as shielding the other leg of
the patient.
Methods
Figure1 shows the single loop structure. It
consists of two multilayered blades in Z direction connected by two horizontal PCB’s
with additional tuning/matching capacitors. The size of the gaps on the multilayered
structure controls the capacitance of the blade. The tilted angle of the
multilayered structure is optimized to achieve the optimum field pattern at the
resonance frequency (297 MHz). This design of the blades could improve the RF
penetration depth and regulate the mutual coupling between the array elements
[1-2]. Figure 2 shows the structure of
the prototype transceive coil array. The coil consists of 8 loop-elements
(placed in two rows) positioned on a cylindrical section of 180mm in diameter and
a length of 200mm in z-direction. The RF shield was placed inside the coil
housing 30mm radially away from the array elements. Figure 3 shows the
decoupling technique for different individual coils. In order to decouple loops
which are in the same raw or column, small inductive decoupling loops are
formed in the corner or the edge of each loop to decouple from the adjacent
loops. Inductively coupled 8-shape loops are positioned between the coil
elements and the RF-shield to decouple the cross loops (for example loop #1 and
#6).
Results and Discussion
The coil parameters are optimized using the EM
full wave simulator Ansoft HFSS. The tilted angle and the gaps in the blades
are optimized for the optimum field pattern and the best decoupling between the
loops. Also, the shape and size of the decoupling loops are optimized to
achieve decoupling between the loops better than -20dB. The optimization runs
with the presence of an RF shield and a homogeneous knee phantom with the
following parameters: radius = 85mm, height = 200mm, εr=48.6 and
σ=0.6. The optimized array shows decoupling better than -20dB, and port
matching to 50Ω better than -20 dB as well. Figure 4 shows the magnetic field
distribution in two different symmetrical planes for both the loaded and the
unloaded coil. The first is a horizontal plane cut in the centre of the coil
(between the two rows), and the second is a symmetrical vertical plane cut
(plane A-A in Fig.2). The field pattern shows good homogeneity in the region
closer to the inner coil loops (where the knee will be positioned), while it
tails off in the open region. The first prototype of the knee coil has been
fabricated and assembled as shown in Fig.5. Bench testing for the
fabricated coil shows a good agreement with the simulated results.
Conclusion
This work presented the design and fabrication
of a new 8-channel transceive open knee coil. Both the simulation and the bench
experimental results show excellent agreement.
Acknowledgements
No acknowledgement found.References
[1] Weber et al, IEEE EMBS, pp. 2039-2042, 2008.
[2] Weber et al, ISMRM, pp.
151, 2008.