4097

Evaluation of a Four-Ring Birdcage Head Coil for ¹H Imaging and ¹³C Spectroscopy at 3 Tesla

Joseph Murphy-Boesch¹, Kazu Yamamoto², Ananda Kumar³, Peter L. Choyke⁴, and Murali Krishna Cherukuri²
¹NINDS, National Institutes of Health, Bethesda, MD, United States, ²NCI, National Institutes of Health, Bethesda, MD, United States, ³Lambda-Z Technologies, Baltimore, MD, United States, ⁴NCI, NIH, Bethesda, MD, United States

Synopsis

Dissolution DNP with Hyperpolarized ¹³C requires receive coils optimized for sensitivity and a transmit coil capable of delivering strong B₁ fields for the carbon frequency. Here, a dual-tuned, four-ring birdcage head coil is described that can produce a transmit field of >90 microTesla and serve as a receive coil for an MRI system with few receivers. Hyperpolarized ¹³C spectra localized using 2D-CSI were obtained in 6 seconds from a human pancreatic tumor xenograph grown in the flank of a nude mouse. SNR of ¹³C spectra obtained from a tumor volume of <1 ml was suitable to measure the lactate/pyruvate ratio.

Introduction

Dissolution dynamic nuclear polarization (DNP) techniques¹ have made it possible to acquire sensitive ¹³C metabolic images from prostate², and brain^3,4. The low gyromagnetic ratio of ¹³C requires strong transmit B₁ fields to excite ¹³C spins. For head imaging, nested birdcage designs with 8 struts per layer are often used⁵. The four-ring birdcage, however, utilizes a single layer of conductors to provide superior axial B₁ homogeneity at both frequencies⁶. Here, a 16-leg four-ring birdcage is described that provides excellent baseline performance for ¹H imaging and ¹³C metabolic imaging at 3 Tesla. To provide a measure of the ¹³C sensitivity that might be obtained from a tumor in human brain, localized ¹³C spectra were acquired from a small (1 ml) human pancreatic tumor MiaPaCa-2 xenograft grown on the flank of the hind leg of a nude mouse.

Methods

A dual-tuned, four-ring birdcage coil⁶, was designed as a transmit/receive coil for ¹H imaging and ¹³C spectroscopy of the head. The coil, shown in Fig. 1a, was constructed with 16 legs, mounted concentrically inside an RF shield, and configured to operate in quadrature mode at both frequencies. The dimensions and ring spacings for the coil, shown in Fig. 1b, were chosen to be proportionally the same as for a dual tuned body coil reported earlier⁷. Four saddle-type inductive couplers were mounted over outer meshes of the coil as elements of a four-port coupling network for the ¹H frequency. For the ¹³C frequency, two quadrature feed cables were routed over the RF shield to matching points on the inner legs of the coil. Traps for the ¹H frequency were inserted in series with these connections to improve isolation. Tuning of the coil was performed on the bench using a head-shaped phantom (Phantom Labs, Niskayuna, NY) filled with PVP and saline (ε = 52, σ = 0.55). The RF currents and fields of the coil were simulated at both frequencies using FEKO software (Altair Hyperworks, Troy, MI). For the ¹H field simulations, Thevenin sources were connected across four outer mesh capacitors. In Fig. 3, gradient echo ¹H images were acquired from the head phantom on a Philips Ingenia Elition 3T scanner, and ¹H sensitivity was compared with that of the standard head coil. To evaluate ¹³C sensitivity, 2-D CSI spectra were acquired from an 18 cm spherical phantom filled with 100% ethylene glycol and processed using R5.6.0 software on the scanner. Finally, localized hyperpolarized ¹³C spectra were acquired from a pancreatic tumor xenograft grown in the flank of a nude mouse. Bolus 96 mM [1-¹³C]pyruvate, which was generated by SPINLAB (GE Healthcare, Milwaukee, Wisconsin), was intravenously injected to the animal 10 seconds after dissolution, and 2-D localized CSI spectra were acquired from the entire mouse 30 sec after injection.

Results

Tuning of the four-ring birdcage for ¹H and ¹³C was straight-forward, since none of the high-order modes of the birdcage structures were overlapping. ¹H port to ¹³C port isolations were greater than -45 dB for both frequencies. RF field simulations for each frequency are shown in Figs. 2b and 2d for coronal views through the coil. In sagittal views,¹H images of the head phantom show good uniformity, with image intensity dominated by the dielectric properties of the phantom. ¹H sensitivities measured at the center of a silicon oil phantom were 169 for the four-ring birdcage and 173 for the standard head. For the ¹³C frequency, simulations in Fig. 2b showed good coverage over the brain. Bench measurements of a loaded coil using a calibrated loop yielded a B₁ of 1.68 µT/ Watt^1/2. The maximum available transmit power for ¹³C was 3150 Watts, yielding a maximum B₁ for the coil of 94 µT. This was confirmed with tip angle measurements of the phantom. In Fig. 4, 2-D localized ¹³C spectra acquired from the ¹³C phantom show baseline resolved triplets from ethylene glycol. Spectroscopic images in from the center and side peaks showed good uniformity, with image intensities stronger near the coil. Nude mice were injected with hyperpolarized [1-¹³C]pyruvate, and 2-D CSI spectra were acquired in 6 seconds in a 7 x 14 array positioned over the animal (Fig. 5). Spectra shown were obtained from 5mm x 5mm x 15mm voxels from the brain, liver, and tumor located in the hind leg.

Discussion

The four-ring birdcage provides excellent coverage for the head with minor loss in ¹H sensitivity as compared with a single-tuned coil for an MRI system. The maximum transmit B₁ field achieved for the ¹³C frequency was 94 µT, more than double that of a “clam shell” coil provided with the ¹³C accessory. Because the ¹³C accessory provides only six receiver channels, a quadrature coil provides near optimal sensitivity as compared with a limited array. Here, the four-ring birdcage provided enough sensitivity from a <0.5 ml tumor voxel to resolve and quantify the lactate-to-pyruvate ratio.

Conclusion

The four-ring birdcage provides excellent ¹H image uniformity and excellent transmit B¹ capability for hyperpolarized ¹³C experiments. ¹³C sensitivity and coverage are adequate for metabolic imaging of the human brain. The four-ring birdcage provides near optimal performance as a quadrature coil and provides a competitive baseline with which to compare head arrays^3,4.

Acknowledgements

This research was supported (in part) by the Intramural Research Programs of the NIH, NINDS and NCI.

References

1. Ardenkjaer-Larsen J-H, Fridlund B, Gram A, et al. Increase in signal-to-noise ratio of >10,000 times in liquid-state NMR. PNAS 2003;100(18):10158.

2. Nelson S, Kurhanewicz J, Vigneron D, et al. Metabolic Imaging of Patients with Prostate Cancer Using Hyperpolarized [1-¹³C]Pyruvate. Sci Translational Med 2013;5(198):1.

3. Mareyam A, Carvajal L, Xu D, et al. 31-Channel brain array for hyperpolarized ¹³C imaging at 3T. Proc. ISMRM 2017;25:1225.

4. Ma J, Hashoian R, Sun C, et al. Development of 1H/13C RF Head Coil for Hyperpolarized 13C Imaging of Human Brain. Proc. ISMRM 2019;27:568.

5. Fitzsimmons J, Beck B, Brooker R. Double Resonant Quadrature Birdcage. Magn Reson Med;1993;30:107-114.

6. Murphy-Boesch J, Srinivasan R, Carvajal L, et al. Two Configurations of the Four-Ring Birdcage Coil for ¹H Imaging and ¹H -Decoupled ³¹P Spectroscopy of the Human Head. J Magn Reson 1994;103B:103-114.

7. Boskamp E, Xie Z, Taracila V, et al. A Dual-Tuned 70 cm Whole-Body Resonator for ¹³C and Proton MRI/MRS at 3T. Proc. ISMRM 2018;26:1714.

Figures

Figure 1. The 16-leg four-ring birdcage shown in (a) is surrounded by a shield, with guard rings at each end to confine the RF fields. The model coil with the dimensions used for simulations are shown in (b). The diameter of the spherical phantom used the model was 22 cm, with electrical parameters, ε= 78 and σ = 0.55.

Figure 2. EM simulations showing current density and transverse magnetic field in (a) and (b) for the proton mode of the coil, and in (c) and (d) for the ¹³C mode of the coil. The coil was loaded with a 22 cm diameter sphere shown in Fig.1b. The maximum SARs in the phantom for an input power of 100 Watts (CW – continuous wave) and averaged over a 10 gm volume and were 26.7 W/kg for the ¹H frequency and 21.2 W/kg for the ¹³C frequency.

Figure 3. ¹H images were acquired from the head-shaped phantom filled with a polymer dielectric material (PVP, ε = 52, σ = 0.55). Axial uniformity benefits from currents of the same magnitude flowing in all 16 legs of the coil. Sagittal and Coronal intensities benefit from B₁ flux concentrated at the two ends of the coil in the H-F direction (see Fig. 2b).

Figure 4. A two-dimensional chemical shift image was acquired from an 18 cm diameter spherical phantom containing 100% ethylene glycol, as shown in (a). Natural abundance ¹³C CSI spectra were acquired in (d) in a 10 x 10 two-dimensional array with 20 mm x 20 mm in-plane resolution,15 mm slice thickness, with tip angle = 30 deg and TR = 750 msec. The spectroscopic image in (b) is for the center peak; the image in (c) is for the right peak.

Figure 5. Hyperpolarized ¹³C spectra taken from brain, liver and tumor are shown in (a), (b) and (c), respectively. The ¹H image in (d) shows a 12 cm diameter bottle used to establish the shim settings. A high-resolution image of the mouse and its CSI grid is superimposed on this image to show the approximate location of the mouse. CSI data were acquired in a 7 x 14 array with 5 mm x 5 mm in-plane resolution,15 mm slice thickness, and a tip angle = 5 deg. The grid in the region of the tumor is shown in (e).

Proc. Intl. Soc. Mag. Reson. Med. 28 (2020)

4097