Joseph Corea1, Balthazar P. Lechene1, Thomas Grafendorfer2, Fraser Robb3, Ana Claudia Arias1, and Michael Lustig1
1UC Berkeley, Berkeley, CA, United States, 2GE Healthcare, Stanford, CA, United States, 3GE Healthcare, Aurora, OH, United States
Synopsis
Extremely
thin, lightweight, and flexible receive arrays can be achieved by the use of
printed electronics. Coil arrays printed layer-by-layer from solution have
shown potential to deliver a comfortable customized fit for many patients.
However, relatively low SNR and poor mechanical robustness prevented these devices from
performing to their full potential. Here we offer SNR within 3% of a
traditionally made coil by using high quality polymeric films as dielectric
layers in capacitors, high conductivity inks, and a mechanically robust
fabrication processes using fewer printed layers and stronger connections.
Using these techniques shoulder and elbow images of a volunteer were obtained. Target Audience
MR engineering, RF coil designers, and clinicians.
Introduction
In our previous work, we have
shown that printed thin, lightweight, and flexible receive coil arrays conform
well to the human body [1-2]. In this work, we present a mechanically robust
receive coil array comprised of high quality printed elements. Flexible arrays
fabricated with our process are lightweight, thin, and easily specialized to
fit different patients and areas of the body with increased comfort. We are
able to achieve this by moving all non-printed components off-coil,
eliminating the need for complicated attachment methods and only requiring one
robust connection between printed and non-printed sections (Fig 1). To create such a device,
alternative coil circuit designs, fabrication, and testing were developed and
characterized.
Methods
and Results
I. Materials Selection and Testing
Our coil was comprised of two printed conductive traces that sandwiched a plastic substrate creating the coil element with capacitors (Fig 2). Many common engineering plastics can be used for a printed flexible receive coil, however, it is non-trivial to develop a process in which conductive ink can be printed on a new flexible substrate. To quickly characterize the plastic films, a test rig was made using copper in the shape of the final coil that clamped films between them to create a coil structure (Fig. 3a). Next, the structure was placed between two probes separated by 20 cm to measure the unloaded Q (Quality Factor away from a conductive sample) [3]. From unloaded Q, we quickly identified polytetrafluoroethylene (PTFE), polyimide (PI), polyethylene-napthalate (PEN), polyether ether ketone (PEEK), polyetherimde (PEI) and polyethylene terapthalate (PET) as films that were strong candidates to create printed coils. Three different types of conductive inks were used to print coils on the selected films (Creative Materials 118-09, Inktek PA-101 and Dupont 5064H). We then measured the unloaded Q of the printed coils to characterize performance. From the results of the testing, Dupont 5064H was chosen for further coil experiments because of the high unloaded Q (54-100) it displayed relative to the other inks (Fig. 3b). SNR experiments were performed on a 3T (Siemens Trio) clinical scanner with a NaCl/NiCl2 doped phantom (conductivity 0.66 S/m) to verify bench top testing (Fig. 3c). Coil elements were 8.7 cm in diameter and tuned to 123.3 MHz. From these experiments, it was seen that all plastic films behaved within 92-97% of a control coil made of non-printed copper and capacitors.
II. Circuit Design and Testing
To create lightweight and thin coils, all non-printed capacitors, inductors, diodes, and preamplifiers were moved off-coil via RG-316 coaxial cable. This created a design (schematic in Fig 4a) that had a detachable coil that mounted on the patient’s body and interfaced with the scanner through a gateway. To characterize the effects of having a remote Q-spoiling and sensing board, 9 cm test coils were created with and without this length of line and imaged on a 3T scanner on a NaCl/NiCl2 doped phantom (conductivity 0.66 S/m). SNR from images (Fig. 4b) indicate a marginal performance loss (3%). However, a well conforming purpose built array would make up this small loss compared to poorly fitting generic or misused array [4].
III. Array Design and Testing
To demonstrate the potential applications of a thin and lightweight printed coil array, a 6-channel coil array was created to image the shoulder and elbow of an adult volunteer. Coils were printed on the mechanically stable polyether ether ketone (PEEK) film using Dupont 5064H silver ink. Coils were then encapsulated in 50 microns of Teflon (PTFE) creating a non-flammable and easy to clean barrier with high dielectric breakdown strength. T2 weighted turbo spin echo scans (TR = 3000 ms/TE = 42 ms) were performed on a volunteer on different areas showing how this arrangement can be implemented (Fig. 5). The full 6-channel array used to image our volunteer weighed only 80 grams while common surface arrays can weigh several kilograms.
Conclusions and Outlook
Our printing technique
combined with robust packaging and high quality films can produce a lightweight
and mechanically robust printed receive coil array that can deliver high image
quality for specialized areas of the body.
Acknowledgements
The
authors would like to recognize the contributions of Anita Flynn (UC Berkeley)
for her research support. Additionally, James Tropp (GE)
also made extremely helpful contributions in coil design. This work was made possible
with funding from GE, NIH R21 EB015628, NIH R01
EB019241, the Hellman Fellowship, the Okawa Fellowship, the Bakar Fellowship, and the
Sloan Research Fellowship.References
[1] J. Corea et. al, Screen Printed Flexible 2-Channel
Receive Coil Array, ISMRM
2012 [2] J. Corea et. al, Design
and Implementation of Flexible Printed Receive Coils Arrays ISMRM 2013 [3] Hayes, C. E.
& Axel, L. Noise Performance of Surface Coils for Magnetic-Resonance
Imaging at 1.5-T. Med Phys 12, 604-607 (1985). [4] Wright, S.M, Full-wave
analysis of planar radiofrequency coils and coil arrays with assumed current
distribution . Con in Mag Res
15, 2-14 (2002) [5] Roemer, P. B et. al,
The NMR phased array. Magn Reson Med 16, 192-225 (1990).