Childs Nerv Syst (2007) 23:305–314

DOI 10.1007/s00381-006-0221-5

ORIGINAL PAPER

Proton magnetic resonance spectroscopy in childhood

brainstem lesions
L. Porto & E. Hattingen & U. Pilatus & M. Kieslich &
B. Yan & D. Schwabe & F. E. Zanella & H. Lanfermann

Received: 8 February 2006 / Revised: 10 May 2006 / Published online: 16 September 2006
# Springer-Verlag 2006

Abstract Results Diagnosis based on localized proton spectroscopy

Background Diagnosis of brainstem lesions in children included brainstem glioma, brainstem encephalitis, demye-
based on magnetic resonance imaging alone is a challeng- lination, dysmyelination secondary to neurofibromatosis
ing problem. Magnetic resonance spectroscopy (MRS) is a type 1 (NF 1), and possible infection or radiation necrosis.
noninvasive technique for spatial characterization of bio- In all but one patient, diagnosis was confirmed by biopsy or
chemical markers in tissues and gives information regard- by clinical follow-up.
ing cell membrane proliferation, neuronal damage, and Conclusions This small sample of patients suggests that
energy metabolism. MRS is important in the differential diagnosis between
Methods We measured the concentrations of biochemical proliferative and nonproliferative lesions in patients without
markers in five children with brainstem lesions and neurofibromatosis. Unfortunately, in cases of NF 1, MRS
evaluated their potential diagnostic significance. Images can have a rather misdiagnosis role.
and spectra were acquired on a 1.5-T imager. The
concentrations of N-acetylaspartate, tetramethylamines Keywords Brainstem lesions . 1H MR spectroscopy .
(e.g., choline), creatine, phosphocreatine, lactate, and lipids Magnetic resonance imaging . Children . Brainstem glioma .
were measured within lesions located at the brainstem using Brainstem encephalitis . Neurofibromatosis type 1
Point-resolved spectroscopy sequences.

L. Porto : E. Hattingen : U. Pilatus : B. Yan : F. E. Zanella :

H. Lanfermann Introduction
Neuroradiology Department,
Klinikum der Johann Wolfgang Goethe-Universität,
The major diagnostic possibilities of brainstem lesions in
Frankfurt am Main, Germany
children are brainstem glioma, encephalitis (viral, autoim-
M. Kieslich mune, or protozoan), demyelination, dysmyelination sec-
Neuropediatric Department, ondary to neurofibromatosis type 1 (NF1), Langerhans’ cell
Klinikum der Johann Wolfgang Goethe-Universität,
histiocytosis, hamartomas, resolving hematoma, vascular
Frankfurt am Main, Germany
malformations, and tuberculoma [1]. The advent of mag-
D. Schwabe netic resonance imaging (MRI) has simplified the diagnosis
Pediatric Hematology/Oncology Department, of brainstem lesions in that it allows differentiation of
Klinikum der Johann Wolfgang Goethe-Universität,
brainstem astrocytomas from vascular malformations and
Frankfurt am Main, Germany
subacute hemorrhage. However, it is not possible to
L. Porto (*) definitively differentiate encephalitis or a tuberculoma from
Institut für Neuroradiologie, brainstem tumor on MRI. Magnetic resonance spectroscopy
Klinikum der Johann Wolfgang Goethe-Universität,
(MRS) offers additional information on the biochemical
Schleusenweg 2-16,
60528 Frankfurt am Main, Germany makeup of brainstem lesions and holds promise in
e-mail: [email protected] improving the diagnostic accuracy of MRI.
306 Childs Nerv Syst (2007) 23:305–314

We performed MRS in all the children in this series with 3.24 ppm (mainly choline, phosphocholine, and glycero-
brainstem lesions and endeavored to establish whether MR phosphocholine, abbreviated as Cho), creatine at 3.05 ppm
spectroscopy is helpful in differentiating major brainstem (Cr, originating from creatine and phosphocreatine),
pathologies, particularly encephalitis or demyelinating N-acetylaspartate (NAA), and N-acetylaspartate-glutamate
lesions from brainstem tumor. (NAAG) at 2.03 and 2.05 ppm, plus seven exponentially
damped sinusoids that accounted for all other resonances
present in the spectrum (e.g., lactate or lipid at 1.3 ppm), to
Materials and methods optimize the fitting of the resonances of interest. Since no
clear distinction between NAA and NAAG can be
The MR images and clinical records of five children with achieved, both metabolite intensities were added up to
brainstem lesions were reviewed. MR imaging and MRS tNAA. Intensities of Cho, tCr, and NAA were quantified as
were performed. ratios between the lesion and the cerebellum. In addition,
MRS was performed on a 1.5-T whole-body MR-scanner the ratio between NAA and total creatine (NAA/tCr) and
(Magnetom Vision, Siemens, Germany) using a standard the ratio between choline and total creatine (CHO/tCr)
circular polarized head coil. For contrast-enhancing masses, within the lesion were calculated and used for interpretation
postcontrast T1-weighted images were obtained (Gadolin- of the data.
ium-DPTA, TR=237 ms, TE=4.8 ms, pulse angle=80). For Spectroscopic inclusion criteria were the following: (1)
noncontrast-enhancing tumors, T2-weighted images were A maximum partial volume effect of 10% was accepted (a
obtained (TR=7,400, TE=114 ms, pulse angle=160°). maximum of 10% of nonmass area within the VOI,
Single-voxel MRS was performed using the Point-resolved determined volumetrically). (2) Local field inhomogeneities
spectroscopy (PRESS) volume selection with an echo time were accepted with a line broadening of less than 10 Hz
(TE) of 135 ms and a repetition time (TR) of 1,500 ms. full-width at half maximum. (3) No artifacts due to
Suppression of water signal was achieved by frequency- movement of the patient. (4) A sufficient reference
selective chemical shift-selective (CHESS) pulses. Sizes of spectrum was obtained using identical acquisition parame-
volumes of interest (VOI) ranged from 2.2 to 12.7 ml and ters from the cerebellum of normal appearance. (5) There
were positioned within the lesion core. The volume was had to be sufficient water suppression and no extravoxel
adapted to the size of the lesion. The minimum extension of lipide contamination to avoid inaccurate quantification of
each side was at least 1 cm in the different directions. The the resonances for choline, lipide, lactate, and NAA.
reference spectrum was obtained using identical acquisition There was no overlap between the control spectra and
parameters from the cerebellum of normal appearance. the brainstem abnormality spectra.
Depending on the size of the selected VOIs, 128-256
acquisitions were added, with total acquisition times
ranging from 3.25 to 6.50 min per spectrum. Before Results
acquisition, global and localized shimming was con-
ducted on the nonsuppressed water signal, followed by Data on clinical information and lesion location as well as
adjustment of the frequency-selective pulses for optional data on imaging findings and outcome are presented in
water suppression. The total time required for the MR- Tables 1 and 2 (see Figs. 1, 2, 3, 4 and 5).
spectroscopic examination ranged from 30 to 40 min.
Initially, data were analyzed on the spectrometer console
using the postprocessing tools provided by the manufac- Discussion
turer. For quantitative analysis, data were processed off-
line on a LINUX workstation using the jMRUI software Childhood brainstem lesions are difficult to evaluate with
(Version 2.1). The residual water signal was removed MRI. Importantly, this is an area of the brain where biopsy
from the measured free induction decay (FID) by means is dangerous. In contrast to MRI and CT, MRS provides
of time-domain Hankel Lanczos singular value decompo- completely different information regarding neural integrity,
sition (HLSVD) filtering [18] as implemented in the cell proliferation, cell degradation, energy metabolism, and
MRUI package [14, 18–20]. necrotic transformation of brain or tumor [12, 13].
The resulting pure metabolite FIDs were then analyzed It is generally recognized that NAA is a marker of
with AMARES, a nonlinear least square fitting algorithm functional neurons and their appendages. When brain tissue
operating in the time domain [21]. The time-domain model is damaged or replaced by any destructive, degenerative, or
function was composed of four exponentially decaying infiltrative process, NAA is markedly reduced [5, 10, 11,
sinusoids with identical damping, corresponding to peaks in 23]. Since Cr-bound phosphates are a substrate of the ATP/
the frequency domain assigned to trimethylamines at ADP cycle, Cr is considered an indicator of energy
Childs Nerv Syst (2007) 23:305–314 307

metabolism. In addition, Cr is believed to be relatively indicates hypoxic conditions and hypermetabolic glucose
constant in various metabolic conditions, it has often been consumption.
used in previous studies as an internal standard for Although as many as 43% of patients with NF 1 will
semiquantitative evaluation of metabolic changes of other show hamartomas by MR imaging, fewer than 10% of the
brain metabolites. However, the use of a ratio will result in lesions show growth on serial examinations [4]. Of note,
loss of metabolic information since a ratio does not change there appears to be no correlation between the number and
when both the numerator and the denominator alter equally, location of these lesions with the clinical status of the
nor does a ratio reveal the direction of the changes. patients. Therefore, it is important to differentiate hamarto-
Furthermore, working with only a metabolite ratio will mas from low-grade astrocytomas. Obviously, early detec-
not take into account the known regional and age- tion of astrocytomas is desirable. Jones et al. [8] reported
dependent differences in metabolite signals. Therefore, we that focal brain lesions in patients with NF 1 could be
have calculated the ratios from the same metabolite (e.g., separated into two groups, one demonstrating only slight
NAA) measured in the lesion core and the normal metabolite ratio changes relative to normal brain and the
appearing cerebellum as an internal standard. other group showing significant increase in choline and
Cho resonances originate mainly from intermediates of decrease in NAA. In this latter group of patients, regression
phospholipid metabolism such as phosphocholine and of a previously identified lesion was noted. The significant
glycerophosphocholine. Both metabolites play an important increase in Cho/Cr ratios and in the Cho levels, as in our
role in the structure and function of cell membranes. patient 1, is unexpected, as an increase in Cho is often seen
Consequently, increased Cho can be seen in processes with in the presence of rapidly dividing cells and Cho is
elevated cell-membrane turnover, such as in proliferating increased in all primary and secondary brain tumors. It is
tumors and in the developing brain. Lipids and lactate are therefore perhaps surprising to encounter this change in
physiologically not detectable in healthy brain. Studies what are considered to be benign, asymptomatic lesions.
have shown that the amount of lipids detected by Wang et al. [22] postulated that Cho elevations reflect
spectroscopy correlates well with the degree of micro- and increased myelin turnover in areas of intramyelinic edema,
macro-necrosis seen on histology [24]. Brain lactate is which is followed by neuropil injury (reduced NAA). This
produced in conditions of anaerobic glycolysis. It is a result assumption is in agreement with their report of elevated
of a mismatch between glycolysis and oxygen supply and Cho and relatively preserved NAA within unidentified

Table 1 Clinical information and lesion location

No. Age/ Diagnosis Symptoms No. of Location

gender lesions

1 4 NF 1 Positive family history and 2 1. Brainstem mass with extension

years visual failure to the right cerebellar peduncle
old/F
2. Optic chiasm/hypothalamus
astrocytoma
2 8 BS glioma, confirmed by Fever, vomiting, dysphagia, slurred 1 Brainstem mass with extension
years open biopsy speech, and disequilibrium to the cerebellum
old/M
3 5 Presumed BS encephalitis. Left-sided horizontal and vertical 1 Hyperintensity in the tegmentum
years CSF was unremarkable. nystagmus, double vision, and of the pons and mesencephalon
old/M Oligoclonal bands and disequilibrium after coryzal
mitochondrial tests illness
were negative.
4 11 Positive oligoclonal bands Right-sided weakness and Multiple Lesions in the white matter, in the
years confirming the diagnosis disequilibrium; bilateral left basal ganglia, on the left side
old/F of MS cerebellar signs of the brainstem, and in middle
cerebellar peduncle
5 15 Presumed infection or Acute left hemiparesis and ataxia, 2 Stable small cystic changes in the
years radiation which progressed in the next 48 h; previous region of the glandula pinealis;
old/F history of low-grade pineocytoma new abnormality in the right
and radiation therapy upper pons

BS brain stem, MS multiple sclerosis, NF 1 neurofibromatosis type 1

308 Childs Nerv Syst (2007) 23:305–314

Table 2 Imaging findings and outcome

No. MR: Brainstem abnormalities MRS findings Outcome after treatment

1 1. Hyper mass in the brain stem with 1. Decreased levels of NAA and Cr while The patient after 18 months without
extension to the right cerebellar peduncle Cho was only moderately elevated (Fig. 1c). aggressive treatment is doing well with
and mass effect on the floor of the fourth MRS findings were suggestive of no increase in the brain stem mass.
ventricle “hamartoma,” but such values of metabolites
could be also suggestive for low-grade
astrocytoma.
2. MRI control after 5 months showed 2. The follow-up MRS showed decreased
increased mass effect of the BS lesion. levels of NAA while Cho levels showed a
steep increase (approx 290% compared to
the normal appearing parenchyma),
suggestive of high-grade glioma (Fig. 1e).
3. Later MRS control, after 14 months
without any specific therapy, showed
reduction of the Cho levels, but it remained
within high levels (approx 200% compared
to the normal appearing parenchyma).
2 Abnormal signal and enhancement within the Decreased levels of NAA with elevated Cho Died
brain stem and cerebellum (Fig. 2a and b) and Cr, compatible with high-grade glioma
(Fig. 2c)
3 Isolated prolongation of T2 signal in the MRS at the time of MRI control showed a The child was treated with
tegmentum of the pons and mesencephalon normal spectrum without increased lactate or corticosteroid and showed clinical
(Fig. 3a and b) without enhancement after choline (Fig. 3c). improvement. After 4 years, there was
contrast. MRI 1 week later showed complete no further clinical relapse.
regression of the prolonged T2 signal.
4 Multiple hyper lesions in the white matter, in Decreased level of NAA (Fig. 4d). Cho was Improved with therapy
the left basal ganglia, on the left side of the only mildly elevated. MRI and MRS
brainstem, and in the middle cerebellar suggested demyelination.
peduncle (Fig. 4a–c).
5 Round area of signal abnormality in the right MRS (Fig. 5d) showed decreased level of Improved
upper pons, which was hyper on DWI NAA and Cr. Cho was not elevated. A lipid
(Fig. 5c). The lesion showed a ring-like peak and moderate elevation of lactate were
enhancement after contrast (Fig. 5b). On the seen. The diagnosis was presumed to be
T2, 6 months later, the lesion was mildly either infection or radiation. Importantly, a
hypo, suggesting small areas of hemorrhage proliferative process was ruled out by reason
with occult vascular malformation (Fig. 5d). that MRS showed an absence of choline
MR with DWI findings was suggestive of peak.
pontine infarction, but neither infection nor
radiation necrosis could not be ruled out.

Iso indicates on MR isointense; hypo, hypointense; and hyper, hyperintense

BS brain stem, DWI diffusion-weighted imaging, MS multiple sclerosis, MR/MRI magnetic resonance imaging, MRS magnetic resonance
spectroscopy, NF 1 neurofibromatosis type 1, T1 T1-weighted images, T2 T2-weighted images

neurofibromatosis objects (UNOs) in younger patients patients with NF 1 deserve further investigation with
(<10 years) and normal Cho in older subjects. temporal analyses at MRS.
With the knowledge that high Cho levels within UNOs Several former studies [5, 10, 11, 23] observed a
can be seen in younger patients with NF, an MRS control common feature of primary and metastatic brain tumor
after 14 months without any aggressive therapy was done, with increased Cho and decreased NAA. In the case of
which showed moderate but significant decrease in the Cho patient 2, the initial diagnosis was acute infectious or
levels without change in the morphology or size of the autoimmune encephalitis. MRI showed abnormal signal in
brainstem mass. We presume that the brainstem lesion in the pons and cerebellar hemispheres (Fig. 2a and b). MRS
patient 1 was in fact a result of an increased myelin (Fig. 2c) showed significantly decreased levels of NAA.
turnover in areas of intramyelinic edema, with subsequent Cho (249%) was markedly elevated. Due to diminished
spontaneous reduction of the Cho levels. Brain lesions in NAA level and relatively high Cr level, the NAA/Cr ratio
Childs Nerv Syst (2007) 23:305–314 309

Fig. 1 A 4-year-old girl with

visual failure. a Axial FLAIR
image shows a hyperintense
mass in the brain stem with
extension to the right cerebellar
peduncle and mass effect on the
floor of the fourth ventricle.
b Midsagittal T2-weighted im-
age shows the brain stem mass
and a second hyperintense mass
in the region of the optic chi-
asm/hypothalamus compatible
with optic chiasm/hypothalamus
astrocytoma. c, d Left spectrum:
brain stem mass lesion. Right
spectrum: normal cerebellum.
Values of NAA (2.0 ppm), Cr
(3.0 ppm), and Cho (3.2 ppm)
reflect signal peak integrals (ar-
bitrary units). Metabolite levels
relative to the contralateral nor-
mal cerebellum: NAA 53%; Cr
91%; Cho 117%. Levels of
NAA and Cr are decreased
while Cho is only moderately
elevated compatible with
hamartoma. e, f Control spec-
trum after 5 months. e Left
spectrum: brain stem mass le-
sion. f Right spectrum: normal
cerebellum. The left spectrum
shows NAA reduction to 30%
with significant increase in Cho
levels (approx 290%). The
NAA/Cr ratio is very low (0.3;
normal, 1.73), due to diminished
NAA and relatively high Cr
level. These findings were sug-
gestive of high-grade glioma,
probably grade III. Further con-
trol after 14 months without
therapy showed reduction of the
Cho levels
310 Childs Nerv Syst (2007) 23:305–314

Fig. 2 An 8-year-old boy with vomiting, slurred speech, and (3.2 ppm) reflect signal peak integrals (arbitrary units). Metabolite
disequilibrium. a Axial FLAIR image shows abnormal signal in levels relative to the contralateral normal cerebellum: NAA 31%; Cr
the pons and cerebellar hemispheres, with the left side more involved 163%; Cho 249%. No lactate peak. Levels of NAA are decreased
than the right. b Axial T1-weighted contrast enhanced MR image while Cho are significantly elevated compatible with high-grade
shows regions of enhancement in the cerebellum. c Spectrum of brain glioma. The NAA/Cr ratio is very low (0.32; normal, 1.73) due to
stem mass lesion. Values of NAA (2.0 ppm), Cr (3.0 ppm), and Cho diminished NAA and relatively high Cr level

Fig. 3 A 5-year-old boy presented with sudden onset of sixth nerve hyperintense T2-weighted MR imaging signal has fully regressed in
palsy and disequilibrium. a Sagittal T2-weighted image shows a 1 week (not shown). b FLAIR image obtained through the pons shows
diffuse area of hyperintensity in the posterior half of the pons with a well-defined area of hyperintensity in the tegmentum. c Normal
cranial extension to the tegmentum of the mesencephalon. The spectrum of brain stem lesion
Childs Nerv Syst (2007) 23:305–314 311

Fig. 4 An 11-year-old right-

handed girl with acute
right-sided hemiparesis and dis-
equilibrium. Oligoclonal bands
were positive in this patient
confirming the diagnosis of MS.
a, b Sagittal T2 images show
abnormal signals on the left side
of the brainstem and middle
cerebellar peduncle. c Axial T2-
weighted MR image shows the
hyperintense area on the left side
of the brainstem with extension
to the middle cerebellar pedun-
cle. d Spectrum of brain stem
mass lesion. Values of NAA
(2.0 ppm), Cr (3.0 ppm), and
Cho (3.2 ppm) reflect signal
peak integrals (arbitrary units).
Metabolite levels relative to the
contralateral normal cerebellum:
NAA 41%; Cho 118%. No
lactate peak. Levels of NAA are
decreased while Cho are mildly
elevated, which is compatible
with demyelinating disease

was low. The high Cho peak was suggestive of high-grade phase of demyelination in multiple sclerosis (MS) and of
astrocytoma. Primitive neuroectodermal tumor (PNET) was various leukodystrophies, choline levels are usually elevat-
included in the differential diagnosis. It has been reported ed [3]. The increased “choline” signal is believed to result
previously [12] that a significant decrease of Cr signal from increased levels of the myelin breakdown products—
intensity was present in PNET compared to low- and high- glycerophosphocholine and phosphocholine. In patient 3
grade astrocytomas. Therefore, the NAA/Cr ratio of PNET and in other previously reported cases of ADEM [16]
tumors [12] was also significantly higher compared to studied with MRS, the choline signal was normal, and the
gliomas. Based on the relative high Cr level, we favored the region of interest was hyperintense on T2-weighted
diagnosis of a high-grade astrocytoma. This was confirmed sequences. Over time, the hyperintense signal gradually
by open biopsy. resolved (patient 3). Abnormal hyperintense T2-weighted
The cases of patient 3 and patient 4 suggest that MRS signal with normal levels of choline may indicate that the
may have a role in the differential diagnosis of acute underlying pathologic process is related more to the local
disseminated encephalomyelitis (ADEM). During the acute edema than to demyelination. In this setting, a normal
312 Childs Nerv Syst (2007) 23:305–314

Fig. 5 A 15-year-old girl was admitted with a history of left- which are mildly hypointense. e Spectrum of brain stem mass lesion.
sided hemiparesis and disequilibrium. a, b Axial T2 and T1 after Values of NAA (2.0 ppm), Cr (3.0 ppm), and Cho (3.2 ppm) reflect
contrast images show a round hyperintense signal in the right upper signal peak integrals (arbitrary units). Metabolite levels relative to the
pons with a ring-like enhancement after contrast. c Axial diffusion- normal cerebellum: NAA 32%; Cho 86%. A lipid-dominant peak is
weighted image show the hyperintense pontine lesion. d After observed in the spectrum, as well as moderate amounts of lactate
6 months, axial T2 image shows some residual changes in the pons,

choline level might have prognostic implications for NAA levels are decreased in lesions known to involve
recovery, a hypothesis that is potentially of great value neuronal and axonal loss (e.g., infarcts, brain tumors,
but will require testing in a larger cohort of patients [3]. seizure foci, MS plaques). Therefore, it has been suggested
Another important disorder in the differential diagnosis of that the NAA signal may be used as a marker of permanent
ADEM is mitochondrial encephalomyopathy. It has been neuronal or axonal loss. However, it is now becoming
reported that mitochondrial diseases [3] have elevated apparent that, as for other amino acids, NAA levels depend
cerebral and CSF lactate levels on MRS. In the case of not only on neuronal density but also on fluxes of synthetic
patient 3, the absence of lactate in the region of interest and and degradation pathways [3]. A case report of a 2-year-old
CSF was helpful in downgrading the diagnostic likelihood child with mental retardation and normal brain MR imaging
of mitochondrial encephalopathy. Elevated lactate levels findings describes a complete absence of cerebral NAA
have been reported in mitochondrial encephalopathies such [17], presumably resulting from a defect in the synthetic
as mitochondrial encephalopathy, lactacidosis, and stroke pathway of NAA. Therefore, decreased levels of NAA
(MELAS); myoclonic epilepsy and ragged fibers (MERFF); should not automatically be interpreted as resulting from
Kearns-Sayre syndrome; and Leigh disease. However, the neuronal/axonal loss or viability. Reversible NAA
sensitivity and specificity of the spectroscopic detection of decreases have been demonstrated in patients with single
lactate as a diagnostic test for these and other mitochondrial transient acute demyelinating MS lesions, in patients with
diseases remain to be determined. Small elevations of lactate mitochondrial encephalopathy and in one patient with
have also been reported in fulminant demyelinating lesions [3] ADEM [3]. In patients with a single large demyelinating
but were not observed in our cases (patient 3 and patient 4). lesion, De Stefano et al. [6] found a 34% to 45% NAA
Childs Nerv Syst (2007) 23:305–314 313

deficit during the acute phase of the disease, followed by sive treatment and its side effects in patients destined to have
almost complete NAA recovery (70% to 80% of the normal better outcomes than patients with diffuse pontine gliomas.
contralateral NAA levels).
It is often difficult to make the correct diagnosis of ring-
like enhanced lesions on gadolinium-enhanced MR brain
References
images, as in the case of patient 5. The appearance of a new
lesion 8 years after radiation treatment for primary brain
1. Barkovich AJ (2005) Pediatric neuroimaging, 4th edn. Lippincott-
tumor may represent recurrence of the tumor, infection, Raven, Philadelphia, pp 321–347
radiation-induced demyelination [2, 15], radiation necrosis 2. Biousse V, Newman NJ, Hunter SB et al (2003) Diffusion
of the brain, ischemic stroke secondary to radiation-induced weighted imaging in radiation necrosis. J Neurol Neurosurg
Psychiatry 74:382–384
vasculopathy, secondary tumor, or occult vascular malfor-
3. Bizzi A, Ulug AM, Crawford TO et al (2001) Quantitative proton
mation. We believe that tumor recurrence was unlikely in MR spectroscopic imaging in acute disseminated encephalomy-
our patient. The pontine lesion was remote from the tumor’s elitis. AJNR Am J Neuroradiol 22:1125–1130
primary site, and PET did not show increased uptake and 4. Castillo M, Green C, Kwock L et al (1995) Proton MR
spectroscopy in patients with neurofibromatosis type 1: evaluation
spectroscopy did not show any choline peak. Pontine
of hamartomas and clinical correlation. AJNR 16:141–147
infarction was suggested by the initial MRI. However, the 5. Curless RG, Bowen BC, Pattany PM et al (2002) Magnetic
follow-up MR images argued against it in that a typical resonance spectroscopy in childhood brainstem tumors. Pediatr
ischemic infarct should develop into a lacunar defect. Neurol 26(5):374–378
6. De Stefano N, Matthews PM, Arnold DL (1995) Reversible
Spectroscopy showed a high lipid-dominant peak, as
decreases in N-acetylaspartate after acute brain injury. Magn
evidence of necrosis at the level of the pons. On the other Reson Med 34:721–727
hand, the presence of hypointense areas on the T2-weighted 7. Fulham MJ, Bizzi A, Dietz MJ et al (1992) Mapping the brain
images, compatible with small areas of hemorrhage, favors tumor metabolites with proton MR spectroscopic imaging: clinical
the diagnosis of radiation-induced necrosis. relevance. Radiology 185:675–686
8. Jones AP, Gunawardena WJ, Coutinho CM (2001) 1H MR
Previous MRS studies [7, 9] of radiation necrosis have spectroscopy evidence for the varied nature of asymptomatic focal
shown the presence of a high lipid-dominant peak, low brain lesions in neufibromatosis type 1. Neuroradiology 43(1):62–67
NAA peak, and a lower Cho peak compared with 9. Kimura T, Sako K, Tanaka K et al (2001) In vivo single-voxel
metastasis or glioblastoma. The lipid-dominant peak may proton MR spectroscopy in brain lesions with ring-like enhance-
ment. NMR Biomed 14:339–349
reflect the presence of the necrotic part of the radiation 10. Kriger MD, Blüml S, McComb JG (2003) Magnetic resonance
necrosis. Fulham et al. [7] reported that a high lactate- spectroscopy of atypical diffuse pontine masses. Neurosurg Focus
dominant peak was observed in subacute radiation necrosis. 15(1):E5
In the subacute phase of cerebral infarction, a ring-like 11. Laprie A, Pirzkall A, Haas-Kogan DA et al (2005) Longitudinal
multivoxel MR spectroscopy study of pediatric diffuse brainstem
enhanced lesion with mild swelling is found in MR images, gliomas treated with radiotherapy. Int J Oncol Biol Phys 62(1):
and it is difficult to differentiate them from tumors. Typical 20–31
of cerebral infarction is the lactate peak. As in the case of 12. Möller-Hartmann W, Herminghaus S, Krings T et al (2002) Clinical
radiation necrosis, lipid-dominant peaks have been ob- application of proton magnetic resonance spectroscopy in the
diagnosis of intracranial mass lesions. Neuroradiology 44:371–381
served from the central nonenhanced region in the cerebral 13. Nafe R, Herminghaus S, Raab P et al (2003) Preoperative proton-
infarction cases [9]. MR spectroscopy of gliomas-correlation with quantitative nuclear
MRS helped in the case of patient 5 to rule out morphology in surgical specimen. J Neurooncol 63:233–245
recurrence, that is, proliferative process. Unfortunately, the 14. Naressi A, Couturier C, Devos JM et al (2001) Java-based
graphical user interface for the MRUI quantitation package.
final diagnosis remains unclear in this case. MAGMA 12:141–152
15. Park SH, Chang KH, Song IC et al (2000) Diffusion-weighted MRI
in cystic or necrotic intracranial lesions. Neuroradiology 42:716–721
16. Tan HM, Chan LL, Chuah KL et al (2004) Monophasic, solitary
Conclusion
tumefactive demyelinating lesion: neuroimaging features and
neuropathological diagnosis. Br J Radiol 77(914):153–156
MRS plays a major role in the evaluation of brainstem 17. Thiel T, Capone A, Schneider JF et al (2000) N-acetyl-
lesions. We believe that MRS is important in the differential aspartate: a market for viable neurons? [abstract]. In: Book of
International Society for Magnetic Resonance in Medicine’s
diagnosis between proliferative and nonproliferative lesions
(ISMRM’s) 8th Scientific Meeting and Exhibition. Denver, CO,
in patients without neurofibromatosis. ISMRN, p 1909
Unfortunately, in cases of NF 1, MRS can have a rather 18. van den Boogaart A, van Ormondt D, Pijnappel WWF et al (eds)
misdiagnosis role. In situations such as NF 1, Cho peak (1994) Mathematics in signal processing III. Clarendon Press,
Oxford, pp 175–195
should not necessarily be considered as a sign of a high
19. van den Boogaart A, Van Hecke A, Van Huffel P et al (1996)
malignant lesion. With the knowledge of child’s age, MRS MRUI: a graphical user interface for accurate routine MRS data
controls should be performed, thereby minimizing aggres- analysis. p 318
314 Childs Nerv Syst (2007) 23:305–314

20. van den Boogaart A (1997) MRUI MANUAL V. 96.3. A user’s magnetic resonance spectroscopic imaging. Ann Neurol 47
guide to the magnetic resonance user interface software package. (4):477–484
Delft Technical University Press, Delft 23. Warren KE, Frank JA, Black JL et al (2000) Proton magnetic
21. Vanhamme L, van den Boogaart A, Van Huffel S (1997) Improved resonance spectroscopy imaging in children with recurrent
method for accurate and efficient quantification of MRS data with primary brain tumors. J Clin Oncol 18:1020–1026
use of prior knowledge. J Magn Reson 129:35–43 24. Zoula S, Herigaut G, Ziegler A et al (2003) Correlation between the
22. Wang PY, Kaufmann WE, Koth CW et al (2000) Thalamic occurrence of 1H-MRS lipid signal, necrosis and lipid droplets
involvement in neurofibromatosis type 1: evaluation with proton during C6 rat glioma development. NMR Biomed 16(4):199–212