Introduction

Radiofrequency (RF) magnetic field inhomogeneity is a significant concern for imaging of large fields-of-view at 3T MRI. The wavelength of the RF field gets shortened to ~26 cm when measured in the body. It leads to the RF magnetic field’s constructive or destructive interference that implies local areas with a brighter and darker signal.¹ High-permittivity dielectric pads², consisting of mixed ceramic powders and heavy water, have been shown to increase the RF magnetic field in low-transmit efficiency regions at high and ultra-high magnetic fields. However, dielectric pads have several drawbacks: change of the dielectric constant with time, bio-incompatibility of some constituents, and weight of up to several kilograms.³ Recently, an artificial dielectric metamaterial that mitigates those disadvantages has been shown feasible for imaging of parietal lobes at 7T.^4,5 Here, we demonstrate via simulations and in vivo study the ultralight and thin metasurface to improve abdominal imaging at 3T and compare its performance with the conventional dielectric pad.

Methods

The metasurface is shown in Figure 1. The metasurface is organized as a two-dimensional periodic stricture of cross copper strips (18 μm thick) ended with square copper patches (with sizes of 5.76×5.76 mm²). The neighboring crosses are printed on the opposite sides of a flexible dielectric substrate (Dupont Pyralyx AP8515R with a thickness of 25 μm and ɛ = 3.4), and their patches produce parallel-plate capacitors (C = 40 pF). The overall metasurface dimensions are L₁×L₁ = 28×28 cm². The period and the capacitance of the metasurface were chosen to achieve the same phase velocity of the wave propagating along the surface and having an electric field polarized along with the patches as for the conventional dielectric pad⁶ (dimensions of 28×28×1.5 cm³, ɛ_r = 300, and σ = 0.4 S/m). Those conditions correspond to excitation by a body coil of the structure positioned over the human body’s abdominal region. In other words, the metasurface provides the same phase delay as the conventional pad. All numerical simulations were performed in CST Microwave Studio. The transmit RF field’s distribution inside the voxel model (“Gustav” with BMI = 22.3 kg/m²) and a two-port high pass shielded body birdcage coil tuned to 128 MHz was simulated for three cases: (1) without any pad (i.e., reference case), with (2) conventional dielectric pad and (3) metasurface placed on the abdomen. The body coil diameter was 70 cm, length – 49 cm. The figure-of-merit of field homogeneity is the coefficient of variation C_v defined as the standard deviation of B₁⁺ over its mean value in the region-of-interest (ROI).⁷ The ROI is shown by the dashed black line (Figure 2). The SAR (10g averaged) and B₁⁺ maps were calculated for 1W of accepted power. Experimental data were acquired with a Siemens Magnetom Trio 3T whole-body MRI scanner using a T2-weighted HASTE sequence. The receive coil used for the acquisition was Body Matrix Coil (Siemens). The high-permittivity dielectric pad (left panel in Figure 1) was based on CaTiO₃ ceramic powder mixed with heavy water. The dielectric pad and the metasurface had the same geometric and material parameters as in the simulations. A male volunteer with BMI = 28.6 kg/m² was investigated.

Results

Figure 2 shows simulated B₁⁺ maps for three cases: (1) without any pad (i.e., reference case), with (2) conventional dielectric pad, and (3) metasurface placed on the abdomen. One can observe the B₁⁺-field inhomogeneity in the ROI for the reference case (Figure 2a). The addition of the pad and the metasurface improves the B₁⁺-field in ROI (Figure 2b,c), which is confirmed by the decrease in figure-of-merit by ~6% for both cases as compared to the reference case. Figure 3 illustrates simulated local SAR in the same cross-section where the field B₁⁺ was shown. Note that the addition of the pad or the metasurface does not significantly increase local SAR in ROI. Maximum SAR was located in the right arm, and it was 0.11 W/kg for all three cases. Experimentally obtained MR images of a volunteer in the axial plane are shown in Figure 4 and Figure 5. First, the birdcage coil was used to transmit and Body Matrix Coil to receive (Figure 4), whereas, in Figure 5, we used the birdcage coil for both transmit and receive modes. In Figure 4a, one can see a dark region in the abdomen’s top area. The dielectric pad (Figure 4b) or the metasurface (Figure 4c) brings brightness into the dark region. The same effect is observed in Figure 5, where the reference case’s dark region is even larger.

Discussion and conclusion

The metasurface has been designed to improve the homogeneity of clinical MRI 3T abdominal imaging. Its effect is similar to the high-permittivity dielectric pad. However, the metasurface is based on polyimide material, that is ultralight, cheap, and flexible, marking the proposed metasurface’s advantages. By adjusting the sizes of the printed parallel-plate capacitors in the metasurface unit cells, one can achieve equivalence to different values of permittivity of conventional pads also used in other imaging applications. For instance, a similar metasurface can be implemented to solve other imaging problems caused by the dielectric artifact (i.e., fetal, heart imaging) and at the other static magnetic field strength.

Acknowledgements

This work was supported by the Russian Science Foundation (Grant No. 18-79-10167). The authors thank Dr. ir. Wyger Brink for assistance with dielectric pad fabrication.