Synopsis

The quantification of metabolite signals in the ¹H-spectrum of the human brain in vivo is complicated by an extensive spectral overlap of resonances, due to the restricted chemical shift dispersion of the ¹H-nucleus. For instance, the myo-Inositol peaks observable in the ¹H-spectrum at 7T are close the water resonance and overlap with that of glycine. In contrast, the high resolution of the ¹³C-MR spectrum aids in the interpretation of the highly sensitive but poorly resolved ¹H-MR spectrum. In this study, we demonstrate the advantage of ¹³C-MRS at 7T by measuring high resolved myo-Inositol C₁-C₆ resonances in the human brain.

Introduction

In vivo ¹³C-MRS at ultra-high field (i.e. ≥ 7T) benefits significantly from improved signal-to-noise ratio (SNR) and spectral resolution for the resolution of overlapped metabolite signals. Precisely, the large chemical shift dispersion of the ¹³C-nucleus makes in vivo ¹³C-MRS potentially useful for quantitative measurements of metabolites that are difficult to resolve in the ¹H-spectrum, such as myo-Inositol [1]. An active transport through the blood-brain barrier maintains myo-Inositol in the brain at much higher concentration than in blood, predominantly in central cerebellum and occipital cortex [2]. However, whilst owning six equivalent CH protons, quantitative measurements of myo-Inositol by in vivo ¹H-MRS up to 7T have been limited to two resonances (H_1,3 and H_4,6) [3], due to the restricted chemical shift dispersion of the ¹H-nucleus. Because of the inherent low sensitivity of the ¹³C-nucleus and its low natural abundance, in vivo ¹³C-MRS is typically performed using surface coils, as their high filling factor results in increased SNR, if using adiabatic pulses [4]. In addition, optimal sensitivity and spectral resolution of ¹³C-MRS requires efficient ¹H-decoupling, which can be achieved with low transmit power using a ¹H-quadrature RF coil [5]. In this context, the aim of this study was to explore the advantage of ¹³C-MRS at 7T in terms of sensitivity gain and improved spectral resolution, and to investigate its potential feasibility towards quantitative measurements of myo-Inositol C₁-C₆ in the human brain in vivo at 7T.

Methods

All measurements were performed on a 7T human scanner (Siemens Erlangen/Germany) using a home-built ¹³C-linear/¹H-quadrature RF surface coil. A pulse-acquire sequence for sensitivity-enhanced ¹³C-MRS using bi-level ¹H-decoupling was developed. Uniform ¹³C-excitation was achieved at 7T using two phase-inverted adiabatic half-passage (AHP) pulses of 2ms applied in alternate scans [6] within 650Hz to cover the ~3.2ppm of myo-Inositol C₁-C₆. Bi-level ¹H-decoupling consisted of a series of equidistant WALTZ-cycles interleaved with time delays during ¹³C-relaxation time for the generation of low-power NOE, and WALTZ-16 [7] for broadband ¹H-decoupling during ¹³C-signal acquisition. The duration of the main WALTZ-cycle 90°-pulse was fixed to 1.5ms to cover the ~1ppm of myo-Inositol H₁-H₆. The ¹³C-MR relaxation properties of myo-Inositol C₁-C₆ at 7T were evaluated in vitro on a phantom containing 150mM of myo-Inositol. In particular, the NOE enhancement was measured by applying successively a series of WALTZ-cycles until the steady-state was reached. In addition, the spin-lattice (¹³C-T₁) and spin-spin (¹³C-T₂) relaxations times were measured using an optimized adiabatic inversion recovery and Hahn spin-echo [8] sequences. All in vitro ¹³C-MRS measurements were performed using the centre of the ¹³C-spectrum at the C_1,3 resonance (i.e. 73.3ppm), while high spectral resolution was achieved using the appropriate acquisition time and decoupling duration based on the T₂* calculated from the natural line-width of the myo-Inositol C₁-C₆ resonances at 7T. Natural abundance ¹³C-MR spectra were acquired in vivo from the occipital cortex of two healthy volunteers who signed informed consent of the local ethics committee. First and second order shims were adjusted using FAST(EST)MAP [9].

Results and discussion

In vitro ¹H-decoupled ¹³C-MRS of myo-Inositol at 7T revealed high spectral resolution including two degenerate resonances at 73.3ppm (C₁ and C₃) and 71.9ppm (C₄ and C₆) and two single resonances at 75.1ppm (C₅) and 73.1ppm (C₂) (Figure1 left). In addition, improved sensitivity was implied using bi-level ¹H-decoupling by a maximal NOE enhancement factor of 3, achieved by increasing the TR (Figure1 right). Besides, the ¹³C-T₁ of distinct myo-Inositol carbons were found to be rather similar to each other (Figure2 left), while the ¹³C-T₂ were slightly disparate (Figure2 right). The natural line-width of all myo-Inositol carbons measured 7.5Hz, yielding to 42ms T₂*. High spectral resolution was achieved by adjusting the acquisition time and decoupling duration to the FID (~215ms). The un-localized ¹³C-MR spectra acquired from the occipital cortex of the human brain in vivo at 7T revealed high resolved myo-Inositol C₁-C₆ resonances (Figure3 top), while the proximity of the dominating glycerol C₂ peak (69.5ppm) induces a distortion in the baseline of the spectrum, which may complicate myo-Inositol quantification. A localized ¹³C-MR spectrum was acquired from the same subject and the two glycerol lipid peaks were completely suppressed, suggesting possible extension into localized myo-Inositol by increasing the scan time. In addition, by placing the excitation pulse downfield, several resonances were observed, which we tentatively assigned to Glutamate-C₂ (55.6ppm), Glutamine-C₂ (55.1ppm), Creatine-C₄ and Choline (54.7ppm), N-acetyl-aspartate NAA-C₂ (54ppm) and Aspartate-C₂ (53.2ppm) (Figure 3 bottom). Simultaneous detection of myo-Inositol, metabolites from 53-56ppm above-mentioned, and the low-concentrated Taurine-C₁ (48.4ppm) was possible by placing the excitation pulse in the middle of the spectrum in conjunction with bi-level broadband ¹H-decoupling (Figure4).

Conclusion

We conclude that ¹³C-MRS at 7T has the potential to provide high resolved myo-Inositol C₁-C₆ in the human brain in vivo, and that it could become the method of choice for investigating non-region specific brain diseases. The reduced baseline distortion using localization suggests possible extension into reliable quantification of myo-Inositol in the human brain in vivo at 7T. Although the measurement has been performed in the occipital cortex, it could be extended into the cerebellum for functional studies using a double-quadrature ¹³C/¹H surface coil [10] that provides high sensitivity and better coverage of this brain region.