A Systematic Comparison of Different Dipole Antenna Array Configurations for Knee Imaging Applications at 7T

Introduction

In recent years, different dipole antenna designs have been developed and employed for imaging applications of various regions of interest. Many studies involving an effort to control the dimensions of a conventional dipole antenna and tune a shorter dipole at Ultra-high-field strength, such as 7T, gave excellent solutions, such as incorporating meander elements in different sections of the conventional dipole design^1–3. Our previous studies on the dipole antenna yielded two promising solutions for coating the conventional dipole with high dielectric constant material to tune the shorter dipoles at 7T^4–8. Each of the dipole designs comes with its advantages and disadvantages. Thus, it is necessary to evaluate each design’s performance in an array configuration and to find a way to choose an appropriate way to determine which design to choose for a specific application. To facilitate this demand, we came up with a fair comparison structure where we evaluated 4,8, and 16-channel array configurations formed by each design explicitly modeled for the knee imaging application. Further, we systematically compared the performance of each array configuration based on the B1⁺ fields, SAR maps, and inter-element isolation to understand which design performs better in a specific case.

Method

We modeled 4,8, and 16-channel array configurations from the following dipole designs: End Meander dipole, Complete Meander dipole, Fractionated Dipole, Fully Dielectric Material Coated (FDMC) Dipole, Discretely Dielectric Material Coated (DDMC) Dipole. The length of each dipole design was kept the same at 25 cm to provide a fair comparison and was placed circularly around the human-knee-shaped phantom, which can be seen in Fig1. Due to broader meanders on the fractionated dipole, the design could not be arranged in a 16-channel configuration; hence, the 16-channel case for fractionated dipole was not evaluated. All designs were numerically simulated using COMSOL Multiphysics software. The phantom used in simulations is identical to the human knee shape and was assigned the following material properties: relative permittivity ε_r: 48, electrical conductivity σ:0.537 S/m, and density ρ: 985 Kg/m³. The dimension used for each dipole design is shown in Fig1. Every channel was fed via the lumped port of 1V, and phase was specified for every port based on channel placement. Further, each channel was tuned at 300MHz by selecting appropriate parameters based on the design and matched by using a matching shunt capacitor connected to the lumped port. We evaluated the frequency response of each array by computing the frequency domain study at 300MHz, where all the channels were fed together, and calculated B1⁺ & SAR maps in the phantom and the inter-element isolation.

Figure 1.

The 4,8, and 16-channel array configurations for the Discretely Dielectric Material Coated (DDMC) Dipole, Fully Dielectric Material Coated (FDMC) Dipole, Complete Meander Dipole, End Meander Dipole, and Fractionated Dipole, respectively. Each figure shows the dimensions for each channel used in the respective configuration. 16-channel configuration for Fractionated Dipole shows the overlaps caused by broader meanders and the reason prohibiting the evaluation of the same.

Results

Fig2. shows the sagittal B1⁺ slices in the phantom produced by the 4,8 and 16-channel configuration of each dipole design expressed in μT/√W. The DDMC dipole produced the highest B1⁺ values of 5.18 μT/√W among others in a 16-channel array configuration. In the 8-channel array configuration, Fractionated dipole design produced the highest B1⁺ values of 4.15 μT/√W among others. Finally, End Meander, Fractionated, and FDMC dipole produced very similar B1⁺ values of 2.65,2.61, and 2.61 μT/√W, respectively in the 4-channel array configuration. Further, Fig3. shows the sagittal SAR maps in the phantom by all the array designs expressed in dB. In the 16-channel array configurations, the Complete Meander dipole produced the highest SAR values of −19.1 dB, among others. The End Meander dipole produced the highest SAR values of −22.9 dB in 8-channel array configurations. Finally, DDMC dipole produced the highest SAR values of −22.9 dB, among others. Finally, Fig4. shows the S-parameters matrix for each array design. All the array designs had very similar inter-element coupling values; the data can be seen in Fig5. In addition, Fig5. shows all the data values for peak B1⁺ and peak SAR values produced by each array design.

Figure 2.

Sagittal B1⁺ maps of each design’s 16,8, and 4-channel array configuration in the human-knee-shaped phantom expressed in μT/√W, respectively.

Figure 3.

Sagittal SAR maps produced by each dipole design’s 16,8, and 4-channel array configurations in the human-knee-phantom expressed in dB.

Figure 4.

S-parameters noise matrix for each array configuration expressed in dB.

Figure 5.

The table shows each array configuration’s Peak B1⁺, Peak SAR, and isolation values.

Discussion/Conclusion

In this study, we modeled and simulated 14 array systems formed by each dipole design and compared the performance of each array system based on the peak B1⁺, peak SAR, and inter-element isolation. Based on our comparison, we found that for 16-channel array configurations, the DDMC dipole design outperformed other designs by producing the highest B1⁺ values and lowest SAR values than the rest of the designs, making it a better choice for 16-channel configurations. For 8-channel array configurations, the Fractionated dipole produced the highest B1⁺ values than the rest but also produced the highest SAR values, which may impose a problem. Thus, the second-best combination of values was generated by the FDMC dipole and can be preferred as the alternative to Fractionated Dipole in 8-channel array configurations. Finally, for the 4-channel array configurations, the End Meander, FDMC, and Fractionated dipole produced the best combination of values, which can be used for 4-channel array configurations. Finally, FDMC provided better isolation in the 16 and 8-channel array configurations, and DDMC provided better isolation in the 4-channel array. We also found that the wider size of the fractionated dipole restricts the design from being incorporated in the 16 or more channel arrays for knee imaging. Based on the current setup, only the FDMC and DDMC can be arranged in array configurations for knee imaging with more than 16 channels.

Synopsis.

In this study, we investigate different multi-channel array configurations formed by various dipole antenna types and compare the array configurations in B1+ field maps, SAR maps, and inter-element isolation for knee imaging at the ultrahigh field of 7T.