In vivo ³¹P MRS provides an important neuroimaging tool for studying high-energy phosphate metabolisms, neuroenergetics and NAD redox state ^[1-2]. However, its sensitivity is still limited for addressing challenging scientific questions even at ultrahigh field. This study was to investigate the efficacy and utility of ultra-high dielectric constant (uHDC) technology for significantly improving SNR and reducing RF power requirement for in vivo ³¹P MRS at ultrahigh field of 7T. The demands of ultrahigh field have been growing rapidly due to its substantial improvement of the spatial/temporal/spectral resolution with increased signal-to-noise ratio (SNR), which are essential for studying and understanding brain function, connectivity and neuroenergetics. Despite the advantages, increased RF power deposition or specific absorption rate (SAR) has become a major safety concern and technical hurdle at ultrahigh field. As one of engineering solutions, the introduction of high dielectric constant (HDC) material in MR field has shown some benefits for improving both the transmit efficiency (|B₁⁺|) and receive sensitivity (|B₁^-|) ^[3-5]. However, further development is still needed to maximize the B₁ efficiency and the performance of the HDC material. Thus, our research aim was to 1) develop and optimize the ultrahigh HDC (uHDC) materials based on computer simulation and advanced material science for MR imaging application, and 2) to quantitatively investigate the B₁ efficiency and sensitivity gain of the optimally designed uHDC material for ³¹P MRS application at 7T based on both phantom and in vivo validation.

Design of uHDC: First, optimal dielectric constant of permittivity for the target operation frequency (120.3MHz for ³¹P at 7T) with ³¹P volume coil configuration was determined by numerical simulation of the B₁ fields using xFDTD (REMCOM, USA). Subsequently, a monolithic block made of lead zirconium titanate (PZT) (TRS, State College, PA, USA) was used to reach the ultrahigh permittivity constant (ε_r).
³¹P MRS: All ³¹P MRS measurements were performed at 7.0T/90 cm bore human scanner (Siemens/Magnex) with ³¹P-¹H double-tuned RF volume coil; ¹H channel for anatomic imaging and B₀ shimming, and ³¹P channel for acquiring ³¹P MRS. ³¹P MRS data were collected on phantom (a rectangular bottle filled with inorganic phosphate (Pi: 50 mM) and NaCl (50 mM), and gadolinium contrast agent for shortening the T₁ value of Pi to 300 ms) and on the leg muscle in vivo from a healthy volunteer. All ³¹P measurements for the phantom study were made using a 3D chemical shift imaging (CSI) with Fourier Series Window (FSW) technique ^[6] (TR= 1s, FA= 90°, phase encode= 9×9×7, bandwidth= 5 kHz, FOV= 15×15×20 cm³, and hard excitation pulse width = 750 μs). In vivo ³¹P CSI data were acquired with slightly different parameters of TR= 1.5s with an Ernst flip angle and pulse width of 500 μs. Finally, spatial maps of relative B₁⁺ (i.e., inversely proportional to the RF transmit voltage to reaching a 90° RF pulse flip angle) and B₁^- (proportional to the maximum signal at 90° flip angle) were calculated using fitting algorithm for determining the integrals of phosphorous creatine (PCr) signal for in vivo data and Pi signal for phantom data and their dependence on the variable RF pulse voltages.

Based on the simulation results, we employed four uHDC pads with dielectric constant, ε_eff ≈ 1000, surrounding the phantom or human leg for ³¹P MRS studies at 7T. Both RF transmission field (B₁⁺) and reception field (B₁^-) for acquiring 3D ³¹P CSI at 7T demonstrate large improvements with the uHDC pads for Pi phantom (Fig. 1) and human leg muscle in vivo (Fig. 2). Strikingly, the improvements of B₁ in both phantom and in vivo cases with the uHDC pads increased the SNRs for more than 200% in some tissue region of interest (see Fig. 3) and reduced RF power requirement (more than 200% voltage reduction) in comparison to that of no uHDC pads. Moreover, Figure 3 also shows localized ³¹P MR spectrum by a small RF surface coil (diameter = 5cm) placed close to the targeted muscle. Comparing the spectra from the head volume coil, using uHDC pads could provide equivalent or even slightly better sensitivity. These results provide experimental evidence that uHDC materials improved B₁ efficiency, leading to SNR enhancements and SAR reduction. Based on the field-dependent SNR of the PCr metabolite ^[7], a 200% SNR gain with the uHDC technique observed at 7T is equivalent to a similar performance and SNR expected at > 15T with no added concerns for SAR.We have demonstrated experimentally that uHDC materials significantly improved both |B₁⁺| and |B₁^-| for in vivo ³¹P MRS. Therefore, utilizing uHDC materials could be an important and cost-effective engineering solution for overcoming high specific absorption rate (SAR) and significantly gaining SNR at ultrahigh fields, which will provide enormous benefits for in vivo human applications including brain research.