Assessment of Normal Fetal Brain Maturation In Utero by Proton Magnetic Resonance Spectroscopy Nadine Girard, 1 * Sylviane Confort Gouny, 2 Ange `le Viola, 2 Yann Le Fur, 2 Patrick Viout, 2 Kathia Chaumoitre, 3 Claude D’Ercole, 4 Catherine Gire, 5 Dominique Figarella-Branger, 6 and Patrick J. Cozzone 2 Cerebral maturation in the normal human fetal brain was inves- tigated by in utero localized proton MR spectroscopy ( 1 H MRS). Fifty-eight subjects at 22–39 weeks of gestational age (GA) were explored. A combination of anterior body phased-array coils (four elements) and posterior spinal coils (two to three ele- ments) was used. Four sequences were performed (point-re- solved spectroscopy (PRESS) sequence with short and long TEs (30 and 135 ms), with and without water saturation). A significant reduction in myo-inositol (myo-Ins) and choline (Cho) levels, and an increase in N-acetylaspartate (NAA) and creatine (Cr) content were observed with progressing age. A new finding is the detection of NAA as early as 22 weeks of GA. This result is probably related to the fact that oligodendrocytes (whether mature or not) express NAA, as demonstrated by in vitro studies. Cho and myo-inositol were the predominant res- onances from 22 to 30 weeks and decreased gradually, proba- bly reflecting the variations in substrate needed for membrane synthesis and myelination. The normal MRS data for the second trimester of gestation (when fetal MRI is usually performed) reported here can help determine whether brain metabolism is altered or not, especially when subtle anatomic changes are observed on conventional images. Magn Reson Med 56: 768 –775, 2006. © 2006 Wiley-Liss, Inc. Key words: fetal brain; prenatal development; MRI; MR spec- troscopy; maturation The indications for fetal brain MRI have increased due to continuous improvements in MRI methods and instru- ments. As a matter of fact, MRI has become the method of choice to evaluate brain maturation and development (1,2), as well as abnormalities in those processes (1,3,4). The primary indications for fetal brain MRI are evaluations of central nervous system (CNS) malformations, ventricu- lar dilatation, and pregnancies at risk of fetal brain damage (1,3–5). The classic MR protocol includes single-shot T 2 - weighted images and T 1 -weighted images. Diffusion im- ages can also be used routinely to evaluate tissue micro- structure. In addition, proton MR spectroscopy ( 1 H MRS) has become an efficient tool for obtaining metabolic infor- mation from the human brain. MRS has been used as a powerful diagnostic tool in the pediatric population, espe- cially for detecting hypoxic encephalopathy, leukoen- cephalopathyies, and inborn errors of metabolism (6 – 8). MRS also provides metabolic information on the develop- ing brain (9 –11). However, little is known about metabolic changes that occur in the human brain during in utero development (12–15). Indeed, many of the available data were obtained ex utero from premature neonates, a situa- tion that does not reflect the physiological conditions of pregnancy. A second bias stems from the differences in acquisition techniques used to collect these data, since for ex utero data a head coil dedicated to brain studies is used, whereas in utero data are obtained with phased-array body and spinal coils (4). Kok et al. (12,13) demonstrated the possibility of inves- tigating in vivo the human fetal brain using 1 H MRS. They also proposed normative values for levels of fetal brain metabolites during the third trimester of pregnancy (ges- tational age (GA) = 30 – 41 weeks), and used two MRS acquisition sequences: stimulated-echo acquisition mode (STEAM) with a short TE of 20 ms, and point-resolved spectroscopy (PRESS) with a long TE of 135 ms. In the present study we attempted to establish normative meta- bolic values in a wider and earlier range of GAs (22–39 weeks) than that suggested by Kok et al. (12) (30 – 41 weeks), and used the PRESS sequence at short and long TEs (30 and 135 ms). MATERIALS AND METHODS Subjects A total of 69 examinations were performed, and 58 yielded good-quality MR spectra. In the remaining examinations the limited quality of spectra was due to maternal or fetal movements (e.g., a fetal head moving with the mother’s breathing) during the MR exploration, especially in cases of breech presentation. The present study thus consisted of 58 MR examinations performed on 58 subjects, since none of the subjects was investigated twice during the preg- nancy. MR data on normal fetal brains were obtained from subjects who ultimately displayed normal standard MR images and/or normal brain examination postnatally, and normal neurological outcome at 3 months of age. These 1 Service de Neuroradiologie, Assistance Publique-Ho ˆ pitaux de Marseille, Ho ˆ- pital la Timone, Universite ´ de la Me ´ diterrane ´ e, Marseille, France. 2 Centre de Re ´ sonance Magne ´ tique Biologique et Me ´ dicale, Unite ´ Mixte de Recherche (UMR) Centre National de la Recherche Scientifique 6612, Faculte ´ de Me ´ decine la Timone, Marseille, France. 3 Service de Radiologie, Assistance Publique-Ho ˆ pitaux de Marseille, Ho ˆ pital Nord, Marseille, France. 4 Service de Gyne ´ cologie Obste ´ trique, Assistance Publique-Ho ˆ pitaux de Mar- seille, Ho ˆ pital Nord, Universite ´ de la Me ´ diterrane ´ e, Marseille, France. 5 Service de Ne ´ onatologie, Assistance Publique-Ho ˆ pitaux de Marseille, Ho ˆ pi- tal Nord, Marseille, France. 6 Service de Neuropathologie, Assistance Publique-Ho ˆ pitaux de Marseille, Ho ˆ pital la Timone, Universite ´ de la Me ´ diterrane ´ e, Marseille, France. *Correspondence to: Pr. Nadine Girard, Service de Neuroradiologie Diagnos- tique et Interventionnelle, Ho ˆ pital la Timone, 264 rue Saint-Pierre, 13385 Mar- seille Cedex 05, France. E-mail: nadine.girard@ap-hm.fr Received 10 December 2004; revised 12 June 2006; accepted 15 June 2006. DOI 10.1002/mrm.21017 Published online 8 September 2006 in Wiley InterScience (www.interscience. wiley.com). Magnetic Resonance in Medicine 56:768 –775 (2006) © 2006 Wiley-Liss, Inc. 768