Introduction

At UHF, the use of transmit/receive (Tx/Rx) coil arrays is advantageous to mitigate B₁⁺ inhomogeneities, e.g. with pTx techniques, and to perform accelerated imaging. Form-fitted¹ and flexible coils^2–6 for 7 T MRI have become one of the hot topics in RF coil engineering. Coaxial coils^7,8 are especially interesting as they are lightweight, robust against inter-element coupling^6,7 and can be ideally form-fitted to the anatomy of interest, even with large inter-subject variability. In this work, we fabricate flexible coaxial coils (CCs), as described in a previous study^9,10, with one or multiple gaps in the inner and outer conductor and four different loop diameters, auto-resonant at the target resonance frequency for 7 T ¹H-MRI. On the bench and in MRI, we examine the CCs’ robustness upon bending, their transmit efficiency and SNR performance as compared to standard coils (SCs) made of copper wire.

Methods

Coil design and fabrication:
Depending on the desired biomedical application, the coil diameter has to be adapted for optimal MR signal Tx and Rx¹¹: 4cm (skin, hand, wrist), 7cm (elbow, ankle, breast, head), 10cm (breast, head, knee) and 15cm (abdomen). While a single-turn, single-gap coaxial coil self-resonant at ≈300MHz would be restricted to a maximum loop diameter of ≈4cm^5,10, introducing multiple gaps in the inner coaxial cable conductor and cable shield allows for larger CC diameters whilst maintaining the desired self-resonance frequency. Using an equivalent circuit scheme⁹, we calculated CC solutions listed in Table 1 with up to 3 gaps and two different cable types that yielded the closest possible self-resonant coil geometries to the target resonance frequency (f_Larmor=297.2MHz). The modifiable parameters were the number of cable gaps (n_g) and turns (n_t), and the coaxial cable type. For comparison purposes, SCs with the same diameters and segmented each ≈λ/10 were fabricated of 1mm copper wire. All fabricated coils including basic interfacing circuitry at the coil port for tuning and matching are shown in Fig.1a and b.

Bench and MRI measurements:
A 5 or 25 l container and holder set-up with a balloon placed on a Teflon sheet shown in Fig.1c were used as phantom loads for bench and MRI measurements, all filled with saline solution (1.6g NaCl/l deionized water, DC conductivity ≈0.2 S/m) and doped with 1ml/l Gd-based contrast agent. Bench and MRI tests were performed with the 4/7/10 cm coils directly placed on/under the phantom, whereas 15 cm coils were placed at 1cm distance. On the bench, SCs and CCs were measured in flat position and CCs in half-bent and bent position. On the bench, the loaded resonance frequency (f_l) and the matching level (in dB) were determined with S₁₁ measurements on a vector network analyzer (E5071C, Agilent, Santa Clara, CA, USA). A decoupled double loop probe¹² was used to measure the loaded Q factor. Operating the coils in Tx/Rx mode, MR experiments were carried out on a SIEMENS Magnetom 7 T system for flat CCs and SCs as well as bent CCs. B₁⁺ maps normalized to the input power were calculated from flip angle distributions, measured using rectangular pulses of 500μs and adjusted pulse amplitudes per coil diameter to achieve a 90° flip angle in a ROI at ≈d₀/4 distance from the coil (25, 63, 130 and 200 V for the 4, 7, 10 and 15 cm coils, respectively). SNR maps were calculated from acquired GRE MR images in the central transversal slice and SNR was evaluated in a circular ROI underneath the coil center (see Fig. 3).

Results and discussion

Bench measurement results in Table 2 show that the stronger the CC bending, the more f_l shifts and the matching level at 297.2MHz degrades as the coils were initially tuned and matched in flat configuration; also different phantom sizes were used in bent measurements. Given a specific application, coils would realistically be tuned and matched to a suitable load and the average bending. B₁⁺ efficiency presented in Fig.2 shows good comparability of SCs and CCs of 4/7 cm diameters. The poor matching level of the 4 cm CC in bent position explains the lower B₁⁺ than in flat position. The 7/10/15 cm bent CC measurements indicate robustness and show an increase in B₁⁺/√P when form-fitted to the phantom. SNR maps cropped to the respective coil size are presented in Fig.3. In a circular ROI in flat coil arrangement, depending on the coil size, CCs have comparable or lower SNR than SCs (-8%, -2%, -27%, -36% for d₀ = 4, 7, 10, and 15 cm). However, form-fitting the CCs to the phantom compared to the SC in flat position results in a considerable SNR gain (+51%, +46%, +240% and +287% for d₀= 4, 7, 10 and 15 cm).

Conclusion

We demonstrate flexible multi-gap coaxial coils with 4-15 cm loop diameter made of commercial coaxial cables as Tx/Rx elements for 7 T MRI. In flat configuration, depending on the coil size, CCs show comparable to decreased transmit efficiency and SNR performance as compared to SCs made of copper wire. However, when form-fitted to a phantom, a significant SNR gain as compared to flat SCs is observed for all coil sizes. In future work, we will investigate array configurations and perform SAR simulations.

Acknowledgements

This project was funded by the Austrian/French FWF/ANR grant, Nr. I-3618, “BRACOIL“, and Austrian/French OeAD WTZ grant FR 03/2018.