Quantification of glutathione transverse relaxation time T using echo time extension with variable refocusing selectivity and symmetry in the human brain at 7 Tesla.

Glutathione (GSH) is an endogenous antioxidant implicated in numerous biological processes, including those associated with multiple sclerosis, aging, and cancer. Spectral editing techniques have greatly facilitated the acquisition of glutathione signal in living humans via proton magnetic resonance spectroscopy, but signal quantification at 7 Tesla is still hampered by uncertainty about the glutathione transverse decay rate T relative to those of commonly employed quantitative references like N-acetyl aspartate (NAA), total creatine, or water. While the T of uncoupled singlets can be derived in a straightforward manner from exponential signal decay as a function of echo time, similar estimation of signal decay in GSH is complicated by a spin system that involves both weak and strong J-couplings as well as resonances that overlap those of several other metabolites and macromolecules. Here, we extend a previously published method for quantifying the T of GABA, a weakly coupled system, to quantify T of the strongly coupled spin system glutathione in the human brain at 7 Tesla. Using full density matrix simulation of glutathione signal behavior, we selected an array of eight optimized echo times between 72 and 322 ms for glutathione signal acquisition by J-difference editing (JDE). We varied the selectivity and symmetry parameters of the inversion pulses used for echo time extension to further optimize the intensity, simplicity, and distinctiveness of glutathione signals at chosen echo times. Pairs of selective adiabatic inversion pulses replaced nonselective pulses at three extended echo times, and symmetry of the time intervals between the two extension pulses was adjusted at one extended echo time to compensate for J-modulation, thereby resulting in appreciable signal-to-noise ratio and quantifiable signal shapes at all measured points. Glutathione signal across all echo times fit smooth monoexponential curves over ten scans of occipital cortex voxels in nine subjects. The T of glutathione was calculated to be 145.0 ± 20.1 ms (mean ± standard deviation); this result was robust within one standard deviation to changes in metabolite fitting baseline corrections and removal of individual data points on the signal decay curve. The measured T of NAA (222.1 ± 24.7 ms) and total creatine (153.0 ± 19.9 ms) were both higher than that calculated for GSH. Apparent glutathione concentration quantified relative to both reference metabolites increased by up to 32% and 6%, respectively, upon correction with calculated T values, emphasizing the importance of considering T relaxation differences in the spectroscopic measurement of these metabolites, especially at longer echo times.