Hypoxia-ischemic encephalopathy in full-term neonate: correlation proton MR spectroscopy with MR imaging

European Journal of Radiology 45 (2003) 91
/98 www.elsevier.com/locate/ejrad

Hypoxia-ischemic encephalopathy in full-term neonate: correlation proton MR spectroscopy with MR imaging Guoguang Fan a,*, Zhenhua Wu a, Liying Chen a, Qiyong Guo a, Binbin Ye a, Jian Mao b a

Department of Radiology, #2 Hospital/China Medical University, No. 36 Sanhao St., Heping dist, Shenyang, Liaoning 110004, People’s Republic of China b Department of Pediatrics, #2 Hospital/China Medical University, No. 36 Sanhao St., Heping dist, Shenyang, Liaoning 110004, People’s Republic of China Received 7 November 2001; received in revised form 11 January 2002; accepted 14 January 2002

Abstract Introduction: To evaluate 1H Magnetic Resonance Spectroscopy (1HMRS) in the diagnosis of hypoxia-ischemic encephalopathy (HIE) of full-term neonates correlated with Magnetic Resonance Imaging (MRI). Materials and methods: Thirty-eight cases of fullterm neonates diagnosed as HIE clinically were selected to perform MRI and 1HMRS examination. The ages ranged from 7 to 17 days, with median age of 8.2 days. In which, 26 cases were followed up and/or MRI reexamined at 6 months of age or later. Eight healthy neonates, with no evidence of birth asphyxia, also underwent 1HMRS for comparison. SE sequences were used for routine MR examination; point resolved spectroscopy sequence was required for 1HMRS. The metabolites in the spectra includes: N acetylaspartate (NAA), choline compounds (CHO), creatine compounds (CR), myo-inositol (MI), lactate (LAC), glutamate and glutamine (Glu
/Gln). Results: The peaks of NAA were fall in two cases; the peaks of LAC, which were elevated, appeared as typical double-peaks appearance in 26 cases; the peaks of Glu
/Gln, which were also elevated, appeared as zigzag appearance in nine cases. The peaks of CR were decreased in 11 cases, while those of MI were increased in seven cases. Mild type of lesions was present on MRI in 12 cases whose LAC/CR ratio lower than 0.5; mild and moderate types of lesions were present in 15 cases whose LAC/ CR ratio between 0.5 and 1.5. Whereas, nine cases of severe lesions and two cases of moderate lesions were present on MRI in 11 cases whose LAC/CR ratio greater than 1.5. Twenty-six of 38 cases were followed up and/or MRI reexamined after 6 months, in which, sequelae were present in 12 cases. Among them, eight cases of sequelae in nine cases whose LAC/CR ratio greater than 1.5 were present (account for 88.89%). Conclusion: 1HMRS plays an important role to diagnose and predict outcome of HIE. # 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Brain, diseases; Brain, metabolism; Hypoxia-ischemic encephalopathy; Neonate; Brain, MR imaging; Proton magnetic resonance spectroscopy

1. Introduction Hypoxia-ischemic encephalopathy (HIE) remains a major cause of perinatal and neonatal mortality, and of permanent neurodevelopmental disability [1,2]. Early detection is crucial for interventions aimed at preventing or reversing ongoing injury. Nowadays, Treatment of

* Corresponding author. Tel.: /86-24-2592-8185; fax: /86-242392-9902 E-mail address: [email protected] (G. Fan).

HIE is frustrated by the lack of accurate predictors of long-term neurologic outcome [3]. Proton magnetic resonance spectroscopy (1HMRS) enables us to measure the various metabolite ratios in vivo in normal and pathologic conditions. 1HMRS may be a helpful method to detect abnormalities of metabolism even when normal structures are present on magnetic resonance (MR) imaging studies [4
/7]. Therefore, combination of MR studies and 1HMRS may make it possible to evaluate severity and help predicting outcome of HIE. In our study, we compared the efficacy of MR imaging and 1HMRS in the diagnosis of HIE in 38

0720-048X/02/$ - see front matter # 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 7 2 0 - 0 4 8 X ( 0 2 ) 0 0 0 2 1 - 9

92

G. Fan et al. / European Journal of Radiology 45 (2003) 91
/98

cases. Meanwhile, 26 cases were followed up and/or MRI reexamined at 6 months of age or later to evaluate proton MRS in the prediction of long-term neurologic outcome of HIE.

2. Materials and methods 2.1. Patient selection and clinical data collection Thirty-eight full-term babies with a mean birth weight of 3470 g (range 2680
/4565 g) and a mean gestational age of 40 weeks (range 38
/41 weeks), who were suspected of having suffered birth asphyxia, were examined with MR imaging and 1HMRS. All neonates participated in this study, which was approved by the institutional review board of our hospital. The 16 boys and 22 girls were 7
/17 days, with median age of 8.2 days during MR examination. No babies studied had evidence of congenital malformation or inherited metabolic disorder. Fetal distress had been suspected in 33 neonates, the need for resuscitation after birth in 13 neonates and 17 had been delivered after obstetric intervention. Among them, 10 neonates had history of cyanosis after delivery; eight with umbilical stalk encircling their necks during delivery. An apgar score of six or below at 5 min was present in 11 neonates. Among them, nine neonates with mild HIE (Sarnat stage 1: hyperalertness and hyperexcitability); 13 with moderate HIE (Sarnat stage 2: lethargy, hypotonia, and suppressed primitive reflexes); 16 with severe HIE (Sarnat stage 3: stupor, flaccidity, and absent primitive reflexes) according to clinical signs and the presence of a history of asphyxia. Meanwhile, 26 cases were followed up and/or MRI reexamined at 6 months of age or later. In which, sequelae were present in 10 of 18 cases of MR imaging reexamination; whereas three neonates suffered from psychomotor retardation in eight cases of follow-up examination. Eight healthy full-term neonates, with the informed consent of their parents, also underwent 1RMRS and MR imaging for comparison. Their median gestational age was 39 weeks (range 38
/41), and their median birth weight was 3600 (range 2850
/5050). The three boys and five girls were 6
/15 days, with median age of 8.4 days during MR examination. 2.2. MR imaging and proton MR spectroscopy techniques The localized MR spectroscopy was performed in conjunction with MR imaging of the brain on 2.0-T superconductive MR system (GE Elscint Prestige) with a circular polarized head coil. Informed consent was obtained and personnel from the intensive care unit

during the MR imaging and proton MRS examinations monitored the neonates. Sedation with chloral hydrate (dose, 50
/100 mg/kg) was provided as needed. Routine MR imaging examination included axial spin-echo T1weighted imaging and axial fast spin-echo (FSE) T2weighted imaging (TR ms/TE ms /620,3600/12,96; one signal acquired; 3
/5-mm-thick sections) and sagittal spin-echo T1-weighted imaging (620/12; two signal acquired; 5-mm-thick sections). Meanwhile, gradientecho sequence (TR ms/TE ms /100/5; one signal acquired; 3-mm-thick sections; total examination time, 15
/20 s) was acquired in all neonates suffered from asphyxia to rule out hemorrhages before routine MR examination performed. MR axial images were used to define optimal placement of the region of interest in the 1 HMRS examination. Localized water-suppressed proton spectra were obtained with use of a point resolved spectroscopy (PRESS) sequence (1500/35.45, 1020 data points, 200 signals acquired). Proton spectra were acquired in an 8-cm3 region of interest placed in the basal ganglia and thalamic areas, respectively. Localized shimming, phases correction, water suppression calibration and scan acquisition to eliminate artifacts caused by eddy currents were performed prior to acquisition of the spectra. Total study time averaged 30
/40 min. Major metabolites detected included N -acetyl compounds, primarily N -acetylaspartate (NAA); choline compounds (CHO); creatine compounds (CR); myoinositol (MI); lactate (LAC); glutamate and glutamine (Glu
/Gln). Spectral postprocessing included phase correct to reduce the noise level; baseline correction to eliminate any drift in the baseline; peak calibrations and spectrum plotting. Relative concentrations of metabolites were related to peak-area, and expressed with reference to CR. Spectra localized to the basal ganglia and thalami were analyzed from each neonate, and the results were averaged to give an overall result for each subject. 2.3. Data analysis and statistics 2.3.1. MR imaging grade All MR images were reviewed by two neuroradiologists without knowledge of the 1HMRS results. Each HIE patient’s MR imaging was graded into three degrees purposely, correlated with clinical symptoms and/or follow-up results (Grade I/minor brain abnormalities; Grade II /moderate brain abnormalities; Grade III /major brain abnormalities). Minor brain abnormalities were defined as streak-like hyperintensities on T1-weighted images along the gyri within the cortex or subcortical white matter that corresponds to vascular boundary zones; slight subarachnoid or subdural hemorrhage near the confluence of the sinuses and subtentorial region. Moderate brain abnormalities were

G. Fan et al. / European Journal of Radiology 45 (2003) 91
/98

defined as symmetrical hyperintensity surrounded by local edema in the deep white matter of the frontal lobe just in front of the anterior horn; streak-like hyperintensities on T1-weighted image along the wall of lateral ventricle, which represent venous congestion or hemorrhage [8], although sometimes combined with appearance of Grade I. Major brain abnormalities were defined as spotty hyperintensities on T1-weighted image in the area of basal ganglia and/or thalami and/or brain stem combined with poor myelination of posterior limbs of internal capsule (these areas appeared characteristic low signal intensity that normally was high signal intensity with fully myelination after birth, myelinating state of other structures was difficult to determine at this age on routine MR imaging); diffuse edema within the white matter; intraventricular hemorrhage; periventricular leukomalacia; necrosis in the subcortical area, although sometimes combined with appearance of Grades I and II.

1

2.3.2. HMRS grouping According to difference of the mean values of LAC/ CR, correlated with MR imaging and clinical symptoms or follow-up results, HIE neonates were divided into three groups: group 1 was the value of LAC/CR lesser than 0.5; group 2 was the value between 0.5 and 1.5; group 3 was the value greater than 1.5.

2.3.3. Statistical analysis All data of 1HMRS obtained in both healthy control group and HIE neonates’ group were summarized as the mean9/standard deviation (SD). We used a t -test and t?test to determine the mean levels among healthy control group and different groups of HIE. Only values of P B/ 0.05 were considered significant.

3. Results

3.1. 1HMRS manifestation of normal full-term neonates (eight cases) In eight cases of normal healthy full-term neonate, the appearance of 1HMRS was similar in shape. The peaks of CHO, NAA and CR were elevated compared to those of HIE neonates, and those of MI were slightly decreased, whereas, peaks of Glu
/Gln and LAC were flattened or undetectable. The amplitudes in high-to-low order were as follows: CHO, NAA, CR, MI, Glu
/Gln and LAC (Fig. 1). The metabolite peak-area ratios were given in Table 1.

93

Fig. 1. 1HMRS of the basal ganglia in normal neonate

3.2. 1HMRS manifestation of full-term neonates suffered from HIE (38 cases) The peaks of NAA were cliffy and sharp in shape in 36 of all 38 cases, while those were reduced in height relative to the other major peaks and flattened in two cases, indicating probably impairment of neuron (Fig. 2) [9,10]. The peaks of LAC, which were elevated in height, appeared as typical double-peaks appearance in 26 cases, indicating excessive accumulation of lactate in the brain. The peaks of Glu
/Gln, which were also elevated, appeared as zigzag appearance in nine cases (Fig. 3); the peaks of CR were decreased in 11 cases, while those of MI were increased in seven cases. Table 1 gives the results for the peak-area ratios NAA/CR, CHO/CR, MI/CR, Glu
/Gln/CR, and LAC/ CR for neonates suffered from HIE, with results from the control group. The levels of NAA, MI, Glu
/Gln, and LAC relative to CR were significantly higher in HIE group compared to control group (P B/0.01), while no statistic difference of levels of CHO relative to CR was present in these two groups. Table 2 shows the lactate levels relative to CR in control group and different HIE groups, the level of lactate relative to CR was significant difference among control group and different HIE groups (P B/0.05). Nine cases of grade I and two cases of grade II were present on MRI in 12 cases whose LAC/CR ratio lower than 0.5 except one case with normal appearance (Fig. 4a and b). Eleven cases of grade II, two cases of grade I and one case of grade III were present in 15 cases whose LAC/CR ratio between 0.5 and 1.5 except one case with normal appearance (Fig. 5a and b). While eight cases of grade III, two cases of grade II and one case of grade I were present on MRI in 11 cases whose LAC/CR ratio greater than 1.5 (Fig. 6a and b, Table 3). The coincidence rate for MRI grading in different 1HMRS group was 75, 73.33, and 63.64%, respectively. The abnormal

G. Fan et al. / European Journal of Radiology 45 (2003) 91
/98

94

Table 1 Metabolite peak-area ratios in the HIE group and control group (x9SD)

Control group HIE group

Cases

NAA/CR

CHO/CR

MI/CR

Glu
/Gln/CR

LAC/CR

8 38

1.8890.62 4.2293.56

1.5190.35 1.6691.22

0.6290.13 1.3790.84

0.5590.31 1.3590.95

0.2390.29 3.3592.86

5.401 P B 0.01

1.921 P  0.05

3.740 P B 0.01

3.986 P B 0.01

4.105 P B 0.01

t ? value P value

Fig. 2. 1HMRS of the basal ganglia in HIE neonate. NAA peak and CR peak are decreased, MI peak and LAC peak are increased.

Fig. 4. (a) 1HMRS of the basal ganglia in HIE neonate. LAC peak is increased, LAC/CR is lower than 0.5 in patient 1. (b) Representative T1-weighted spin-echo image (620/12) shows minor brain abnormalities in patient 1. Fig. 3. 1HMRS of the basal ganglia in HIE neonate. The peaks of Glu
/Gln are increased, appeared as zigzag appearance.

Table 2 Lactate peak-area ratios in the HIE groups and control group (x9SD)

Control group LAC/CRB 0.5 0.5B LAC/CRB 1.5 LAC/CR 1.5

Cases

LAC/CR

t value

P value

8 12 15 11

0.2390.29 0.4790.22 1.2390.39 3.9391.91


/ 2.326 3.624 4.947


/ P B 0.05 P B 0.01 P B 0.01

changes of peak amplitude for major metabolites except LAC in different 1HMRS groups were showed in Table 3. Comparison of the results of different 1HMRS groups reveals considerable changes in NAA, CR, MI, and Glu
/Gln occurred with the increase of peak-area ratio of LAC/CR. Meanwhile, correlation of 1HMRS grouping with clinical Sarnat stage, seven cases of Sarnat stage 1 and five cases of Sarnat stage 2 were present in 1HMRS group 1; eight cases of Sarnat stage 2, two cases of Samat stage l, and five cases of Sarnat stage 3 in group

G. Fan et al. / European Journal of Radiology 45 (2003) 91
/98

95

Fig. 5. (a) 1HMRS of the basal ganglia in HIE neonate. LAC peak is increased, LAC/CR is between 0.5 and 1.5 in patient 2. (b) Representative T1-weighted spin-echo image (620/12) shows moderate brain abnormalities in patient 2.

Fig. 6. (a) 1HMRS of the basal ganglia in HIE neonate. LAC peak is increased, LAC/CR is greater than 1.5 in patient 3. (b) Representative T1-weighted spin-echo image (620/12) shows major brain abnormalities in patient 3.

2; while the remaining cases of Sarnat stage 3 were present in 1HMRS group 3. Twenty-six of 38 cases were followed up and/or MRI reexamined after 6 months (18 cases */MRI reexamination; eight cases */follow-up), in which, sequelae were present in 13 cases (Table 4). Among them, eight cases of sequelae in nine cases whose LAC/CR ratio greater than 1.5 were present (account for 88.89%). In the group whose LAC/CR ratio greater than 1.5, periventricular leukomalasia was present during MRI reexamination after 6 months in one case with normal appearance on MRI at neonatal stage.

metabolite concentrations in the brain and play an important role in the determination of metabolic disorder, intracellular acidosis, and impairment of neuron which are caused by HIE [12,13]. The characteristic appearance of 1HMRS in neonates suffered from HIE is the rise of LAC peak, appearing as double peak at 1.3 ppm. The steep rise in LAC relative to the other metabolites is presumably largely due to the excessive production of lactic acid in cerebral tissue, which attributable to reduced oxygen supply resulting in decreased oxidative phosphorylation, causing enhanced glycolysis [14
/17]. Creatine (CR) was chosen as the metabolite of reference in 1HMRS because it is not thought to change significantly after hypoxia-ischemia [13
/18]. Therefore, we select peak-area ratio of LAC/ CR as the marker of lactate levels in the brain. The lactate level in the brain, probably rising immediately after the hypoxia-ischemic insult, may finally fall if the accumulation of lactate not exceeding definite levels within 2
/3 weeks after the insult. The reason for that may be the removal of lactic acid from the brain either by local metabolism or by transport such as Na /H

4. Discussion The occurrence of HIE includes a series of complicated pathological and biochemical processes [11]. The predilection site and severity of HIE are correlated with the duration of hypoxia-ischemia and variation of cerebral metabolites levels. Proton MRS, which is based on the phenomena of chemical shift, can reflect the main

G. Fan et al. / European Journal of Radiology 45 (2003) 91
/98

96

Table 3 Comparison of 1HMRS grouping and MR imaging grade 1

HMRS

MR imaging grade (cases) 1

HMRS grouping

LAC/CRB 0.5 0.5B LAC/CRB 1.5 LAC/CRB 1.5 Note :

%

, fall of the peak;

Cases

Normal

I

II

12 15 11

1 1 1

9 2 2 11 1 2

III

Coincidence rate (%)

NAA%

CR%

MI'

Glu
/Gln'

0 1 7

75 73.33 63.64

0 0 2

1 3 7

1 2 4

1 3 5

'

, rise of the peak.

LAC/CR, which appeared as direct sign after hypoxia-ischemic insult [22,23], was sensitive indicator for estimation of outcome and severity of HIE. Our study indicated that most of outcome predicted by LAC/CR coincided with the result of MRI (account for 71.05%) and clinical symptoms (account for 68.42%) in total 38 cases of HIE, whereas 1HMRS was more sensitive than MRI in 21.05% of cases. In one case whose LAC/CR ratio greater than 1.5, handicap was present after 6 months though normal appearance on MRI at neonatal stage. Therefore, the change of 1HMRS may be present earlier than that of image in hypoxia-ischemic infants and might provide early information concerning the severity of the asphyxia. 1 H spectra were localized to the basal ganglia because previous studies using MRI have identified it as vulnerable to hypoxia-ischemic injury in neonates [24,25]. Spectra from this region are also less susceptible to interference arising from extracranial fat tissue [26], so we selected the basal ganglia for localized 1HMRS to increase the probability of disclosing abnormalities in lactic acid and other metabolites. Significant difference in other cerebral metabolites between hypoxia-ischemic neonates and controls were also found. Considerable changes in NAA, CR, MI, and Glu
/Gln occurred with the increase of peak-area ratio of LAC/CR. The fall of NAA peak and the decrease of NAA/CR ratios were found in two cases in our study. However, the abnormal changes in NAA were more prominent at 3 months of age, probably because neuronal necrosis or dysfunction regarding NAA synthesis not yet occurred in neonatal phase [27,28]. Nevertheless, the fall of NAA indicates the future handicap for such neonates may occur.

exchange [19]. If the lactates persist increasing to a definite level, it may result in poor outcome due to energy failure and irreversible insult of neuron [19
/21]. To our knowledge, it has not yet been established about concrete value of LAC/CR resulting in irreversible cerebral insult due to different MR devices and techniques as well as variable changes of lactate level in the asphyxial neonatal brain [12
/21]. We selected 1
/2 weeks old of neonate to study in order to observe actual lactate levels after perinatal asphyxia because relatively stable lactate levels after acute phase, which approaching actual levels may be present during this period [21,22]. Our study showed significant difference was present between control group and HIE group as well as among control group and different HIE groups whose LAC/CR were different. In the group whose LAC/CR greater than 1.5, most of severe lesions tended to be present on MRI (63.64%) and high probability was present for abnormal metabolites levels such as NAA, Glu
/Gln, MI, CR on 1HMRS. In addition, severe clinical symptoms and signs (Sarnat stage 3) were likely present in group whose LAC/CR greater than 1.5 (100%), which confirmed that LAC/CR was a sensitive indicator for the outcome. Meanwhile, about 88.89% of infants suffered from HIE had handicap after 6 months. On the contrary, slight and moderate lesions were present on MRI, and only 45.45% of handicap occurred in the group whose LAC/CR lower than 1.5. According to the difference of LAC/CR, we divided neonates suffered from HIE into three groups. In which, severe lesions and poor outcome were most likely present in the group whose LAC/CR greater than 1.5. Table 4 The outcome of MRI reexamination or follow-up in 26 cases of HIE 1

HMRS grouping

LAC/CRB 0.5 0.5B LAC/CRB l.5 LAC/CR 1.5

MRI reexamination

Follow-up examination

Cases

Normal

PVL

DM

EM

DCC

6 5 7

5 3 0

1 1 2

0 0 2

0 0 1

0 1 1

0 0 1

Cases

Normal

PMR

3 3 2

2 2 1

1 1 1

Notes : SAE, subarachoid effusion; PVL, periventricular leukomalasia; DM, delayed myelination; EM, encephalomalacia; DCC, dysgenesis of the corpus callosum; PMR, psychomotor retardation

G. Fan et al. / European Journal of Radiology 45 (2003) 91
/98

Glu
/Gln peak, resonating at 2.3
/2.5 ppm was detect in more of the HIE neonates than controls, and the difference was statistically significant because of the possible role of glutamate as an excitotoxin in hypoxiaischemic brain injury [19,29]. Accurate detection, however, was difficult due to the proximity of Glu
/Gln to the dominating NAA resonance in this region [29,30]. Our study also showed that three of all nine cases of HIE neonates with the rise of Glu
/Gln peak clinically suffered from seizures. We are not quite sure that excessive accumulation of glutamate may possibly be associated with seizure; nevertheless it may be interesting for future study. MI is always present in newborns due to the still ongoing myelination, the excessively elevated MI peak, compared to that of healthy neonates in seven cases, however, was possibly related to gliosis and poor myelination because MI has been labeled as a breakdown product of myelin [6,26]. The fall of CR peak in 11 cases seemed paradox with its internal reference for the other metabolite peaks. However, no decrease of peak-area of CR may support its cerebral constant content during asphyxial insult. The presence of fall in CR intensity was possibly related to failure of energy metabolism. In summary, increases in LAC/CR peak-area ratios, as well as abnormal changes of other metabolites such as reductions in NAA carry a poor prognosis in hypoxiaischemic neonates. 1HMRS and MRI can reveal important clinical information and provide useful prognostic information, which aid in the selection of infants who would benefit from neuronal rescue therapies. Our studies performed to determine the predictive value of localized proton MR spectroscopy for later neurodevelopmental outcome after HIE have shown promising results but need further evaluation on larger patient samples.

References [1] Cooper CE. In vivo measurements of mitochondrial function and cell death following hypoxic/ischaemic damage to the new-born brain. Biochem Soc Symp 1999;66:123
/40. [2] Rutherford MA, Pennock JM, Schwieso JE, et al. Hypoxic ischaemic encephalopathy: early magnetic resonance imaging findings and their evolution. Neuropediatrics 1995;26(4):183
/91. [3] Pu Y, Li QF, Zeng CM, et al. Increased detectability of alpha brain glutamate/glutamine in neonatal hypoxic-ischemic encephalopathy. AJNR Am J Neuroradiol 2000;21(1):203
/12. [4] Zhiyue W, Robert AZ, Rolf S. Proton MR spectroscopy of the brain: clinically useful information obtained in assessing CNS diseases in children. Am J Roentgenol 1996;167:191
/9. [5] Peden CJ, Rutherford MA, Sargentoni J, et al. Proton spectroscopy of the neonatal brain following hypoxic-ischemic injury. Dev Med Child Neurol 1993;35:502
/10. [6] Ross B, Michaelis T. Clinical application of magnetic resonance spectroscopy. Magn Reson (Quarterly) 1994;10:191
/247.

97

[7] Nakai T, Rhine WD, Enzmann DR, et al. A model for detecting early metabolic changes in neonatal asphyxia by 1H-MRS. J Magn Reson Imaging 1996;6(3):445
/52. [8] Barkovich AJ, Westmark K, Partridge C, et al. Perinatal asphyxia: MR findings in the first 10 days. AJNR Am J Neuroradiol 1995;16(3):427
/38. [9] Cady EB. Magnetic resonance spectroscopy in neonatal hypoxicischaemic insults. Childs Nerv Syst 2001;17(3):145
/9. [10] Falini A, Barkovich AJ, Calabrese Q, et al. Progressive brain failure after diffuse hypoxic ischemic brain MM: a serial MR and proton MR spectroscopic study. AJNR Am J Neuroradiol 1998;19(4):648
/52. [11] Sheth RD, Bodensteiner JB, Riggs JE, et al. Differential involvement of the brain in neonatal asphyxia: a pathogenic explanation. J Child Neurol 1995;10(6):464
/6. [12] Ordidge R, Thornton J, Clemence M, et al. NMR studies of hypoxic-ischaemic injury in neonatal brain using imaging and spectroscopy. Adv Exp Med Biol 1997;428:539
/44. [13] Huppi PS, Barnes PD. Magnetic resonance techniques in the evaluation of the newborn brain. Clin Perinatol 1997;24(3):693
/ 723. [14] Pavlakis SG, Kingsley PB, Harper R, et al. Correlation of basal ganglia magnetic resonance spectroscopy with Apgar score in perinatal asphyxia. Arch Neurol 1999;56(12):1476
/81. [15] Hanrahan JD, Sargentoni J, Azzopardi D, et al. Cerebral metabolism within 18 hours of birth asphysia: a proton magnetic resonance spectroscopy study. Pediatr Res 1997;39:584
/9. [16] Groenendaal F, Veenhoven RH, van der Grond J, et al. Cerebral lactate and N -acetylaspartate/choline ratios in asphyxiated fullterm neonate demonstrated in vivo using proton magnetic resonance spectroscopy. Pediatr Res 1994;35:148
/51. [17] Graham SH, Meyerhoff DJ, Bayne L, et al. Magnetic resonance spectroscopy of N -acetylaspartate in hypoxic-ischemic encephalopathy. Ann Neurol 1994;35(4):490
/4. [18] Cooper CE, Wyatt JS. NMR spectroscopy and imaging of the neonatal brain. Biochem Soc Trans 2000;28(2):121
/6. [19] Penrice J, Lorek A, Cady ED, et al. Proton magnetic resonance spectroscopy of the brain during acute hypoxia-ischemia and delayed cerebral energy failure in the newborn piglet. Pediatr Res 1997;41:795
/802. [20] Chateil JF, Quesson B, Brun M, et al. Pediatr localised proton magnetic resonance spectroscopy of the brain after perinatal hypoxia: a preliminary report. Radiology 1999;29(3):199
/205. [21] Barkovich AJ, Baranski K, Vigneron D, et al. Proton MR spectroscopy for the evaluation of brain injury in asphyxiated term neonates. AJNR Am J Neuroradiol 1999;20(8):1399
/405. [22] Penrice J, Cady EB, Lorek A, et al. Proton magnetic resonance spectroscopy of the brain in normal preterm and term infants and early changes after perinatal hypoxia-ischemia. Pediar Res 1996;40:6
/14. [23] Nakai T, Rhine WD, Enzmann DR, et al. A model for detecting early metabolic changes in neonatal asphyxia by 1H-MRS. J Magn Reson Imaging 1996;6(3):445
/52. [24] Groenendaal F, Roelants-Van Rijn AM, van der Grond J, et al. Glutamate in cerebral tissue of asphyxiated neonates during the first week of life demonstrated in vivo using proton magnetic resonance spectroscopy. Biol Neonate 2001;79(3
/4):254
/7. [25] Novotny E, Ashwal S, Shevell M. Proton magnetic resonance spectroscopy: an emerging technology in pediatric neurology research. Pediatr Res 1998;44(1):1
/10. [26] Wang Z, Zimmerman RA, Sauter R. Proton MR spectroscopy of the brain: clinically useful information obtained in assessing CNS diseases in children. AJR Am J Roentgenol 1996;167:191
/9. [27] Leth H, Toft PB, Peitersen B, et al. Use of brain levels to predict outcome after perinatal asphyxia. Acta Pediatr 1996;85:859
/64.

98

G. Fan et al. / European Journal of Radiology 45 (2003) 91
/98

[28] Maneru C, Junque C, Bargallo N, et al. (1)H-MR spectroscopy is sensitive to subtle effects of perinatal asphyxia, Neurology 2001 25;57(6):1115
/1118. [29] Provencber SW. Estimation of metabolite concentrations from localized in vivo proton NMR spectra. Magn Reson Med 1993;30:672
/9.

[30] van Cappellen AM, Heerschap A, Nijhuis JG, et al. Hypoxia, the subsequent systemic metabolic acidosis, and their relationship with cerebral metabolite concentrations: an in vivo study in fetal lambs with proton magnetic resonance spectroscopy. Am J Obstet Gynecol 1999;181(6):1537
/45.