Introduction

Transmission line resonators¹ have been studied for various applications^2,3. Lately, a new concept using coaxial cables with one gap of the inner and the outer conductor was introduced^4,5. These self-resonant coils have good geometrical decoupling characteristics over an extended overlap range. In combination with thin coaxial cables, this design is well suited for applications where coil flexibility is advantageous, i.e. especially when anatomical inter-subject variability is large or the shape of the region of interest is changing during the measurement. Here, an ultra-flexible 3-channel receive-only one-gap coaxial coil array with compact interfaces is presented.

Methods

Three one-gap coaxial coils with diameter d_coil = 80 mm were made from very thin coaxial cable (Siemens Rx cable, Stark Contrast, Erlangen, Germany) with an inner conductor diameter of 0.2 mm, an outer conductor diameter of 0.9 mm. The interface contains tuning, active detuning, matching as well as preamplifier decoupling combined in one compact layout (15x45x10 mm³) (Fig. 1). To obtain the optimal overlap d_opt, S₂₁ was measured on the network analyzer (E5061B, Keysight Technologies, Santa Rosa, CA, USA) while the center-to-center distance between two coils was varied between 30 and 160 mm. The three coaxial coils were then centered at the corners of an equilateral triangle with side length d_opt (Fig. 2a). To keep the geometry roughly unaltered upon bending, the coils were woven into a wide-meshed textile tissue. This mesh was sewn onto an additional 2 mm foam padding and an outer layer of medical synthetic leather for stability, component protection, and electrical insulation (Fig 2b, top textile layers not shown). To evaluate robustness w.r.t bending, the S-parameter matrix was measured in flat and bent position on a torso phantom filled with tissue-equivalent gel³. The noise correlation matrix was calculated from MR noise-only data with the array in flat configuration on the phantom. To demonstrate the flexibility of the array, two T₂-weighted double echo 3D MR scans were performed on a pineapple with the array either positioned laterally (Fig. 4a) with T_R = 15.98 ms, T_E = 5.35 ms, α = 25° and on the bottom of the fruit (not shown) using T_R = 14.97 ms, T_E = 4.85 ms, α = 25°.

Results and Discussion

A picture of the implemented array is shown in Fig. 2b. The optimal overlap was determined to be d_opt ≈ 62 mm (Fig. 3a), which is similar to the optimal overlap described by Zhang et al., i.e. d_opt ≈ 0.78 d_coil. The S-parameter matrices for the flat and bent setup are very similar (Fig. 3b,c). Noise correlation values < 4% were obtained, as demonstrated in Fig. 3d. The pineapple was nicely depicted in the MR scans (Fig. 4b,c). The total weight of one coil element including interface and preamplifier is about 12 g, excluding the textile enclosure. With the specific weight of the 4-layer textile used (≈ 2 kg/m²), adding array cabling (≈ 0.5 m per channel, ≈ 3.2 g/m), cable traps (≈ 5 g/channel), and additional components for fixation on the patient (≈ 200 g), we can extrapolate a total weight of ≈ 1 kg for a 32-channel version of such an array, making it a comfortably wearable device for patients.

Conclusion

An ultra-flexible and light-weight 3-channel receive-only coaxial coil array for 3T MRI with compact interfaces implemented on flexible textile is presented. The robustness of matching and decoupling upon bending was demonstrated, rendering such coil elements promising candidates for larger flexible arrays for applications, where anatomical inter-subject variability is large, e.g. in breast MRI.

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.