Body imaging at 7T is challenging due to wavelength effects, power constraints, and SAR sensitivity, especially in deep tissue such as the prostate [1]. Hybrid coil arrays, consisting of both radiative antennas and classical resonators, have been shown to increase imaging performance at 7T [2,3,4]. Here we investigate this design concept for body imaging. A hybrid array was optimized in simulation, candidates for both transmit and receive were dipole, loop, and birdcage arrays. The optimal array was build and initial images were acquired.

Imaging performance of a coil array for body imaging of the 50 percentile man [5] and smaller was optimized in EM simulations (HFSS, Ansoft), choosing an elliptical phantom (38x29cm, length=60cm) with average body properties (ɛ_r=32, σ=0.4S/m) as an imaging target. Coil conductors were placed ≥3cm from the subject for load stability, 2cm beneath an elliptical shield (48x40cm, 40cm length) used for cable stabilization and patient safety. Type, size, and lumped element distribution were subject to optimization. The 8ch array candidates were a dipole, a loop, and a birdcage array. The dipole array was optimized for inductor length and location, assuming half wavelength resonance; the loop array was optimized for length and width of individual elements; and the 8-port driven birdcage was optimized for length as well as lumped element capacitance and location. Distance between lumped elements was ≥5cm, and individual elements were evenly distributed around the coil. The 8ch transmit coil configuration was optimized first, based on the trade-off between power and local SAR sensitivity in the central voxel. Load stability was verified on the optimal model. Receive coil performance was then optimized for an 8ch receive coil, added to the chosen transmit array, aiming to maximize SNR for the central voxel. Co-simulation was used to optimize lumped element distribution, and all elements were matched to 50Ω.A dipole array (30cm elements length, inductors 5cm from the feed point) was found to have optimal transmit performance, with reasonable trade-off between power and SAR sensitivity [Figure1]. A birdcage array (20cm length, 4 pF end-ring capacitance, 20pF leg capacitance) produced optimal SNR for the central voxel [Figure2].A shield, made of sixteen 30cm long PCB shield elements attached to a 3D printed holder, acted as structure both for interior coil conductors and the exterior DC and RF cabling [Figure 3]. The top and bottom sections of the array are separable for patient accessibility. The shield and the birdcage coil are connected with socket connectors at the split. Conductors are routed at a feed point behind the shield for both arrays, with a two wire line connected to a lattice balun for dipole elements, and a preamp board including detuning for the birdcage elements (Siemens, Erlangen, Germany). Both the preamplifier board and the lattice balun were mounted outside the shield elements. Special care was taken in constructing the lattice baluns used to match the dipole elements, to avoid cable currents. Baluns were characterized in a separate 3 port measurement fixture [Figure3c], to ensure a 180 degree phase shift and equal amplitude at the differential port. The shield was slotted and shorted with 470pF SMD capacitors, to avoid eddy currents at gradient frequencies. The eight birdcage channels were routed through an ODU. Transmit coils were matched to a body sized phantom with tissue equivalent properties, and connected to an in house built T/R interface box. Passive detuning circuits were placed in the end-rings in addition to the active detuning in the feed port of the birdcage coil, to ensure invisibility from the dipole array during transmit and to maintain transmit efficiency. Power matching and pre-amplifier input match was performed using a body sized phantom (23x33cm length=43cm) filled with tissue simulating liquid. Initial imaging experiments were performed on a 7T system (Siemens, Erlangen, Germany), with low power localizers used to show expected field distributions in quadrature drive [Figure 4].An extensive optimization of different coil designs was performed for abdominal imaging at 7T, and the optimal coil array was constructed. Preliminary results showed that imaging with an 8ch dipole transmit, 16ch dipole/birdcage receive coil is possible. However, additional active detuning of the birdcage coil, on-coil TR switches, and pre-amplifiers for the dipole elements are still forthcoming, so coil performance is not yet optimal. Currently, the long cables used in the transmit path degrade the usable input power to the coil as well as the receive signal for the dipole elements, and a solution to this will be explored in future work.