Linear (“dipole-like”) currents are becoming increasingly dominant in ultimate SNR considerations at higher magnetic-field strengths [1]. Here, we design and implement a 10-channel array consisting of fractionated dipole antenna (fDA) elements for body imaging at 10.5T and investigate its electro-magnetic field performance in comparison with a fDA array at 7.0T [2] using numerical simulations and experimental studies on phantoms as human imaging is not yet approved at 10.5T but will be performed subsequent to pending FDA and IRB approval.

The 7.0T elements were replicas of the 30cm-long fDA elements developed by Raaijmakers et al. [2] (Figure 1.a). The 10.5T fDA elements were designed using the approach in Ref. [2], resulting in 21cm-long elements with 1cm-wide conductors and two 26mm-wide meanders on each pole (Figure 1.b). The 10.5T elements were matched to 50Ω using a lattice-balun network. fDA arrays at both field-strengths consisted of 10-elements housed in flexible fabric covers with a ~9cm center-to-center distance between neighboring elements (10.5T array, Figure 1.c). Six-elements were placed anteriorly and remaining four-elements were placed posteriorly.

The fDA arrays were modeled around the pelvis of an anatomically correct human model (Duke, Virtual Family [3]), with the elements placed at identical locations at both field strengths. Conductors of the fDA elements were placed 2cm off the surface of the skin, and the meshing size of the tissues of the human model were ≤1mm. EM-field distributions of each fDA element were computed using an FDTD solver in SEMCAD X software (SPEAG, Zürich, Switzerland) at 297.2MHz and 447MHz. EM-fields were imported to Matlab (Mathworks, Natick, MA) and phase-only shimming was applied to investigate shim-dependent B1+ and SAR distributions. Peak local 10g-averaged SAR was calculated using virtual observation points (VOP) [4]. Worst-case 10g-averaged local SAR was determined by computing the highest SAR in VOP that can be reached using a phase-only shim. B1+ transmit efficiency normalized to unit total coil power and B1+ SAR efficiency normalized to peak 10g SAR were calculated. Uniformity of the field distribution was assessed by calculating the coefficient of variation (CV) of the B1+ magnitude inside the prostate.

A demonstration of the fDA array was conducted on a whole-body Magnetom 10.5T scanner (Siemens Healthcare, Erlangen, Germany) equipped with sixteen 2kW power amplifiers, in a torso-sized phantom. Phase-only B1+ efficiency shimming was performed on a 4cm²-square anterior to an inner tube along the central slice, the approximate position of a prostate. The B1+ distribution inside the phantom was calculated experimentally using the actual flip-angle method [5] and numerically using SEMCAD.

Numerically computed B1+ transmit efficiency is plotted against the CV inside the prostate for the 10-channel fDA arrays (Figure 2.a). The 7.0T array provided a more uniform B1+ distribution however the 10.5T array provided 2.3% higher B1+ efficiency. Peak 10g local SAR was the limiting factor for RF safety at both field strengths for the investigated shim settings. B1+ SAR efficiency within the prostate normalized to peak 10g SAR is plotted in Figure 2.b. Transmit performance metrics of the peak B1+ efficiency shims at 7.0T and 10.5T (Shim A), and efficiency-homogeneity tradeoff shim at 10.5T (Shim B) are listed in Table 1. B1+ distributions along an axial-slice intersecting the prostate are shown in Figure 3 (corresponding shims are annotated in Figure 2). The 7.0T array provided 2.0% and 11.8% higher B1+ per unit peak 10g SAR compared to Shim A and B at 10.5T, respectively. Worst-case 10g-averaged SAR with phase-only shimming was 31.5% higher at 10.5T (Table 1).

Simulated and measured B1+ distributions inside the phantom at 10.5T are shown in Figure 4.a-b, respectively, with the square annotations showing the B1+ shimming region. Simulated B1+ distributions demonstrate good agreement with the experimental acquisitions.

Prior to receiving approval to run in vivo studies, we have designed, evaluated and tested the performance of a fDA array for body imaging at 10.5T on a phantom and a numerical human model. Comparisons of the two arrays were performed in the prostate due to its central location inside the body and small size where basic shimming methods are generally suitable at 7.0T [6]. Simulations comparing similar array configurations at 7.0T and 10.5T show promising results at 10.5T with only slight decrease in B1+ SAR efficiency. Large SAR increases (often predicted to scale quadratically with field strength) were not observed. Transmit homogeneity with simple RF management strategies (i.e. phase-only static shimming) decreased considerably at 10.5T even in the small target region studied. However this relative degradation in B1+ homogeneity at higher frequencies can be rectified with dynamic RF management strategies.