Purpose: To determine relaxation times of metabolite signals in proton magnetic resonance (MR) spectra of immature brain, which allow a correction of relaxation that is necessary for a quantitative evaluation of spectra acquired with long TE. Proton MR spectra acquired with long TE allow a better definition of metabolites as N-acetyl aspartate (NAA) and lactate especially in children.
Materials and Methods:Relaxation times were determined in the basal ganglia of 84 prematurely born infants at a postconceptional age of 37.8 Ϯ 2.2 (mean Ϯ SD) weeks. Metabolite resonances were investigated using the doublespin-echo volume selection method (PRESS) at 1.5 T. T1 was determined from intensity ratios of signals obtained with TRs of 1884 and 6000 msec, measured at 3 TEs (25 msec, 136 msec, 272 msec). T2 was determined from signal intensity ratios obtained with TEs of 136 msec and 272 msec, measured at 2 TR. Taking only long TEs reduced baseline distortions by macromolecules and lipids. For myo-inositol (MI), an apparent T2 for short TE was determined from the ratio of signals obtained with TE ϭ 25 msec and 136 msec. Intensities were determined by fitting a Lorentzian to the resonance, and by integration.Results: Relaxation times were as follows: trimethylaminecontaining compounds (Cho): T1 ϭ 1217 msec/T2 ϭ 273 msec; total creatine (Cr) at 3.9 ppm: 1010 msec/111 msec; Cr at 3.0 ppm: 1388 msec/224 msec; NAA: 1171 msec/499 msec; Lac: 1820 msec/1022 msec; MI: 1336 msec/173 msec; apparent T2 at short TE: 68 msec.Conclusion: T1 and T2 in the basal ganglia of premature infants do not differ much from previously published data from basal ganglia of older children and adults. T2 of Cho was lower than previous values. T2 of Cr at 3.9 ppm and Lac have been measured under different conditions before, and present values differ from these data. LOCALIZED IN VIVO nuclear magnetic resonance (NMR) spectroscopy allows investigation of specific metabolites in situ, and extends morphologic information provided by imaging modalities. Combined magnetic resonance imaging (MRI) and spectroscopy (MRS) has developed to a stage that enables routine studies of the neonatal brain in a clinical environment (1-4). For 1 H spectroscopy, protocols for absolute quantitation of metabolite concentrations have been proposed that utilize the tissue water signal as the internal concentration reference (5-7). In general, the determination of metabolite concentrations from MR spectra requires correction for relaxation or saturation. These effects can be minimized using short TE and long TR, but with common volume selection methods, there is a remaining error due to signal decay during the nonzero TE. Shorter TR may be necessary to allow sufficient averaging when measuring small volumes. In short TE, spectra signal area determination of metabolites is complicated-especially in small infants-due to macromolecules distorting the baseline, a high glutamine/ glutamate signal partially masking the small N-acetyl aspartate (NAA) signal, and lipids masking lactate (Lac...