Synopsis

Keywords: Non-Array RF Coils, Antennas & Waveguides, Skeletal, 31P/1H Dual-tuned, NIRS

In this work, we present a 3T ³¹P/¹H calf coil with two interleaved and co-centered birdcages in one layer to provide homogenous transmit fields and good SNR for both ³¹P and ¹H nuclei. In addition, the coil design allows for integration of a Near-infrared spectroscopy (NIRS) probe for simultaneous NIRS and MRI readouts during exercise. Simulation and phantom experimental results show that the coil provides homogeneous transmit and receive fields. In vivo experiment show that the coil provide good SNR for both ³¹P MRS and ¹H MRI and MRS

INTRODUCTION

³¹P MRS directly estimates the oxidative adenosine triphosphate (ATP) synthesis rate in skeletal muscle during exercise and recovery (1-3), while ¹H MRI and MRS can provide complimentary information regarding oxygen delivery (i.e. skeletal muscle perfusion), oxygen utilization (i.e. venous oxygen saturation), and differentiates intramyocellular from extramyocellular lipids. Therefore, it is desirable to have a ³¹P/¹H calf coil designed for providing both good ³¹P and ¹H transmit and receive performance. An interleaved layout of two phased arrays in one layer has been found to provide optimal transmit and receive performances for both nuclei at ultra-high field (4,5). However, at 3T and lower fields, the strong coupling between coil elements makes phased array designs difficult. Birdcage is an excellent alternative option for ³¹P/¹H calf coil design at 3T. In this case, birdcage provides homogenous transmit and receive fields for both nuclei, which are essential for image-based evaluations of perfusion or fat distribution as well as ¹H and ³¹P MRS. In this work, we thus present a ³¹P/¹H calf coil with two interleaved birdcages in one layer to provide homogenous transmit fields and good SNR for both ³¹P and ¹H. In addition, the coil design allows for integration of a NIRS probe for simultaneous NIRS and MRI readouts during exercise.

METHOD

Full-wave simulation: Microwave Studio in CST (Dassault Systèmes) is used to model an interleaved and co-centered ³¹P high-pass birdcage and ¹H low-pass birdcage, loaded with a 2mm isotropic Duke leg (Figure 1a) (6). The ³¹P birdcage’s diameter is 210mm, length 200mm, and copper width 15mm, while the ¹H birdcage’s diameter is 212mm, length 150mm, and copper width 8mm. 50ohm ports were placed at each gap on the birdcages, and co-simulation tool in CST was used for tuning and matching with two quadrature-driven ports symmetrically located 45° away from the bottom of the leg. B₁⁺ maps and SNR were calculated for ³¹P and ¹H. Coil prototyping: Figure 1b shows the CAD model of the coil housing, including a coil holder, an interface box, a NIRS probe holder, and a coil support. The birdcages are mounted inside the coil holder and have the same lengths and copper widths as specified for the simulation. A 0.78mm-thick FR4 block was used to separate the two birdcages at each conductive-overlap area (Figure 1c). After matching , each port was connected to a shielded ³¹P/¹H dual-tuned cable trap, and then connected to a TR-switch integrated on a preamplifier board. A ³¹P/¹H quadrature hybrid was used to drive the two ports of the ³¹P/¹H birdcage (Figure 1d). Figure 1e shows setup of the NIRS probe when working with the coil. The NIRS device is held in a NIRS insert (Figure 1f). Phantom experiments: B₁⁺ maps and SNR in a methyl phosphate phantom (cylinder-shaped, 2 g/L) were measured to characterize the spatial coverage and performance of the ³¹P and ¹H birdcages. For ³¹P, B₁⁺ map was calculated with a 2D CSI sequence using the double angle method; and the same 2D CSI sequence with/without RF excitation was used to calculate SNR voxel-by-voxel by dividing the peak absolute signal by the standard deviation of noise. For ¹H, SNR maps were reconstructed from two GRE sequences with/without RF excitation, and a Turbo Field Echo (TFE) sequence was used to acquire the B₁⁺ maps (7). In vivo Experiments: Human subject studies were performed on a Siemens Prisma 3T MRI system (Siemens Healthcare) with institutional review board approval. In vivo ¹H T1w images were acquired with a 3D MPRAGE sequence. An in vivo ¹H MR spectrum was acquired with PRESS sequence; and in vivo ³¹P spectra were acquired with the same 2D CSI sequence as described above.

RESULT AND DISCUSSION

The unloaded to loaded ratio is 8.32 for the ³¹P birdcage, and 10.5 for the ¹H birdcage. Table 1 lists the capacitance of each tuning capacitor in simulation, prototype, and theoretically calculated for a single birdcage (8). These values are very similar, demonstrating that the two birdcages do not strongly couple. The matching of one birdcage also does not change much with/without the other birdcage being present and resonant, as observed during assembly and workbench testing. Therefore, no LCC trap was implemented. Figure 2 shows the simulated SNR and B₁⁺ in the Duke leg and the experimental SNR and B₁⁺ in the phantom of the ¹H and ³¹P birdcages. In both simulation and experiment, SNR and B₁⁺ are slightly stronger in area close to the two ports for both ¹H and ³¹P, but homogenous in the center area. The ¹H MPRAGE images in Figure 3 show anatomical details of the calf, and a representative ¹H MR spectrum shows intramyocellular (IMCL) and extramyocellular lipids (EMCL). Figure 4 shows in vivo ³¹P MRSI data of the calf muscle in the transversal plane, and a representative ³¹P MR spectrum in one of the voxels. The ³¹P MRSI data show that the peak of ³¹P MR spectra in voxels in the center area are similar, indicating homogeneous ³¹P signal intensity.

CONCLUSION

The 3T ³¹P/¹H calf coil with two interleaved birdcages described herein provides homogeneous transmit and receive fields and good SNR for both ³¹P MRS and ¹H MRI and MRS.

Acknowledgements

This work was funded by Cancer Prevention and Research Institute of Texas (CPRIT) RR180056 and was performed under the rubric of the Advanced Imaging Research Center, UT Southwestern Medical Center.