Traveling wave MR using an array of regular RF resonators
Xinqiang Yan1,2 and Xiaoliang Zhang3

1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, 2Radiology, Vanderbilt University, Nashville, TN, United States, 3Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, United States

Synopsis

In this study, we investigate the feasibility of using regular microstrip resonators as RF array elements for traveling wave parallel imaging. In the proposed microstrip array, electromagnetic decoupling between the array elements is sufficient for the practical use. Additionally, geometric factors and diverse B1 fields from individual array elements can be obtained in a relatively large area in the magnet bore. Furthermore, in non-accelerated imaging applications, this decoupled multi-channel traveling wave method could improve sensitivity of traveling wave MRI, which is currently a main issue for traveling wave MRI.

Purpose

It has been demonstrated that parallel imaging can be realized in traveling wave MRI [1] by using multiple patch antennas on the whole body 7T MR [2]. The physical size of the patch antennas is relatively big, particularly when using them as RF array elements. This potentially limits the number of channels. In this study, we investigate the feasibility of using relatively small microstrip resonators [3, 4] as RF array elements for traveling wave parallel imaging. This multi-channel array method can be also used as a way to alleviating the low SNR problem of the traveling wave MR. In the proposed design, all the microstrip resonators are arranged along the circumference of the magnet bore. This arrangement would provide more diverse B1 distribution from each array element in the imaging area and help to gain parallel imaging performance. It also provides easy access for patients to the magnet.

Methods

A copper cylinder bore with a diameter of 63 cm and a length of 160 cm was used as waveguide, as shown in Fig. 1. Microstrip array with 8 elements was placed at one end of the bore. Each element is a typical quarter-wavelength microstrip line resonator, with one end terminated with tuning and matching capacitors and the other end shorted. The widths of ground and stripe conductor are 2.54 cm and 0.635 cm, respectively. Compared with traditional half-wavelength resonators [3], quarter-wavelength resonators have a reduced dimension and alternate magnetic/electric field distributions, which ultimately decrease the coupling between adjacent elements. Simulations were performed using a full-wave electromagnetic software (HFSS, ANSYS, Canonsburg, PA, USA). S-parameter performance, noise correlation matrix and B1 field distribution were evaluated in the simulation. G-factors for 1D SENSE were calculated by the method in Pruessmann et al [5].

Results

Fig. 2A shows the reflection coefficients (S11) of each element and transmission coefficients (S21) between any two elements. It is clear from Fig. 2A that all microstrip elements were well matched to 50 ohm, with S11 better than -30 dB. This ensures the accuracy of the simulation. S21 is about -29 dB between adjacent elements and better than -40 dB between non-adjacent elements, indicating excellent decoupling performance. This excellent decoupling performance has also been validated by the noise correlation matrix (Fig. 2B).

Fig. 3A shows that sagittal B1+ field in empty bore when all eight elements was driven with circular polarization (CP) mode. It is clear from Fig. 3A that B1 distribution along z-direction was quite homogeneous, illustrating “traveling wave” behavior. Fig. 3B and 3C show the transverse B1 profiles of eight elements when fed individually. The axial planes in Fig. 3B and Fig. 3C are 50 cm and 100 cm away from the microstrip array, respectively. Each element shows different B1 profile, making it possible to perform parallel transmission and parallel imaging.

Fig. 4 shows the axial g-factor maps on a water phantom. These axial planes are 40 cm to 120 cm away from the microstrip array. Average g-factor values are marked in red color in Fig. 4. For most of the slices, the average g-factor at can be better than 1.1 when R=2. Fig. 5 shows the SNR maps from all eight channels (top row) and a single channel (bottom row). SNR was calculated by the method in Kellman et al [6]. Due to the excellent decoupling performance (Fig. 2A and B), SNR can be significantly improved by using all eight channels simultaneously.

Conclusions

In the proposed microstrip array traveling wave imaging, geometric factors and diverse B1 fields from individual array elements can be obtained in a relatively large area in the magnet bore. Additionally, electromagnetic decoupling between the microstrip elements is sufficient for the practical use. This result shows that it is possible to use microstrip resonators as RF array elements for traveling wave parallel imaging. The proposed method provides a way to increasing channel number in traveling wave parallel imaging using multi-channel RF arrays. In non-accelerated imaging applications, this decoupled multi-channel traveling wave method may be a way to improve sensitivity of traveling wave MRI, which is currently a main issue for traveling wave MRI.

Acknowledgements

This work was supported by NIH R01 EB016695.

References

[1] Brunner DO, et al. Nature 2009; 457:994-9. [2] Pang Y, et al. MRM 2012; 67(4): 965–978. [3] Adriany G, et al, MRM 2005; 53:434–445. [4] Zhang X, et al, MRM 2001; 46: 443–450. [5] Pruessmann KP, et al. MRM. 1999; 42 (5):952-62. [6] Kellman P, et al. MRM. 2005; 54(6): 1439–1447.

Figures

Figure 1 3-D Models of copper waveguide and 8-channel Microstrip line array in traveling wave MRI.

Figure 2 Simulated S-parameter plots (A) and noise correlation matrix (B) of 8-channel quarter-wavelength microstrip line array in traveling wave MRI.

Figure 3 B1+ field in empty bore. A: Sagittal B1+ field in circular polarization (CP) mode. B and C: Individual B1+ field in different transverse planes (50 cm and 100 cm away from the coil array).

Figure 4 G-factor maps for 1D SENSE on the phantom, FOV=40X40 cm2. These g-factor maps were obtained at different transverse planes (55 cm to 115 cm away from the coil array).

Figure 5 Normalized SNR maps from all eight channels (top row) and a single channel (bottom row) of the quarter-wavelength microstrip array. These SNR maps were obtained at different transverse planes (60 cm to 130 cm away from the coil array). The 8-channel array demonstrated an average of 2.6-fold SNR gain over the 1-channel coil at the central areas. The SNR was calculated by Kellman’s method [6].



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
2129