- Email: [email protected]
Cystic intracranial mass lesions: Possible role of in vivo MR spectroscopy in its differential diagnosis
Magnetic
Pergamon
Resonance Imaging, Vol. 13, No. 7, pp. 1019-1029, 1995 Copyright 0 1995 Elsevier Science Inc. Printed in the USA. All rights reserved 0730-725X/95 $9.50 + .OO
0730-725x(95)00045-3
0 Original Contribution CYSTIC INTRACRANIAL MASS LESIONS: POSSIBLE ROLE OF IN VIVO MR SPECTROSCOPY IN ITS DIFFERENTIAL DIAGNOSIS RAKESH K. GUPTA,* VIJENDERA K. JAIN,* RAJA ROY ,-f AND RAKESH PANDEY* *MR Section, Departments of Radiology, Neurosurgery, and Pathology, Sanjay Gandhi Post Graduate Institute of Medical Sciences Lucknow 226014, India, and tR.S.1.C. Central Derug Research Institute, Lucknow, India HARISH
POPTANI,*
Thirty-four patients showing cystic intracranial mass lesions on MR imaging were evaluated by in vivo proton MR spectroscopy (MRS) with the aim of detecting lesion-specific spectral patterns that may assist imaging in better tissue characterization. In vlvo spectroscopy was performed using stimulated echo acquisition mode with echo times 20 m and 270 m in all, and spin echo with echo time 135 m in 11 patients. All primary neoplasms (in&a-as well as extra-axial) showed choline (3.22 ppm) resonance along with lipid and/or lactate (1.3 ppm). It was not possible to grade cystic gliomas based on N-ace@ asparate-to-choline ratio. High-grade gliomas (n = 8) showed lipid/lactate and low-grade gliomas (?I = 6) showed only lactate. Seven patients with brain abscess showed resonances only from acetate (1.92 ppm), lactate (1.3 ppm) and alanine (1.5 ppm). Two cases of metastatic adenocarcinoma showed only lipid/lactate. In 7 patients with epidermoid cyst, lactate along with an unassigned resonance at 1.8 ppm was observed and could be easily differentiated from arachnoid cyst (n = 2), which showed only minimal lactate. A case of cystic meningioma could be differentiated from cystic schowannoma by the presence of alanine in the former. It is concluded that MR imaging, when combined with in vivo MRS, may help to better characterize intracranial cystic mass lesions.
Keywords: Brain; Brain tumors; Magnetic resonance spectroscopy; Brain abscess. INTRODUCTION
though differences in metabolites between cystic and solid components of the tumor have been reported,’ there is no report describing in vivo proton MR spectroscopy, primarily in cystic lesions of the brain. In the present study, in vivo proton MRS was performed in 34 cases of intracranial cystic lesions with the aim (1) to study the spectral patterns in different types of cystic lesions and (2) to look for any lesionspecific spectral pattern that may assist MRI in its differential diagnosis in the future.
Magnetic resonance imaging (MRI) is an established technique for the evaluation of intracranial mass lesions. It is highly sensitive, yet relatively nonspecific.’ Cystic, intracranial intra-axial mass lesions usually seen on MRI include gliomas, metastases, abscesses and, rarely, hydatid cyst; cystic extra-axial lesions include meningioma, schwannoma, epidermoid, and arachnoid cyst. The characterisation of these lesions on MRI is not always possible.’ In vivo proton magnetic resonance spectroscopy (MRS) has been performed in different intracranial lesions with the aim of helping to better characterise these lesions.2 Brain tumors usually show high choline containing compounds (Cho), low N-acetyl aspartate (NAA), and the presence of lipid (lip) or lactate (lac).3Y4 An attempt has been made to grade the degree of malignancy of different gliomas based on the difference in quantification of these biochemical signals.3-6 Al-
MATERIALS
AND METHODS
Patients Thirty-four patients (26 males and 8 females) with intracranial cystic mass lesions suggested on MRI, were evaluated with in vivo proton MRS. Their ages ranged between l-65 years. Informed consent was taken from all patients/guardians prior to the study. Of the 34 cases, there were 8 high-grade gliomas (grade III-IV), sociate Professor, MR Section, Dept. of Radiology, Sanjay
RECEIVED 12/S/94; ACCEPTED S/16/95. Address correspondence to Rakesh K. Gupta, MD, As-
Gandhi PGIMS, Lucknow 1019
226014, India.
1020
Magnetic Resonance Imaging l Volume 13, Number 7, 1995
STEAM
20
Table 1. Summary
Tumor type
NAA
Low-grade gliomas (n = 6) High-grade gliomas (n = 8) Metastases (n = 2) Abscesses (n = 7) Epidermoid cyst
Cr I
(n = 7)
CHO,
of results
Spectroscopic
findings*
t lac; 1 NAA, Cho, Cr NAAKho = 0.43 f 0.15 fi lip and/or lac; 1 NAA, Cho, Cr NAA/Cho = 1.14 f 0.75 1 lip and/or
lac
no NAA, Cho, Cr Tlac, ala, actetate, cystosolic amino acids, no NAA, Cho, Cr tlac, unassigned peak (1.8 ppm) no NAA, Cho, Cr Tlac, no NAA, Cho, Cr
Arachnoid cyst (n = 2) Meningioma (n = 1)
r?unassigned (2.04 ppm), t Cho, lac, ala, no NAA, Cr
Schwannoma (n= 1)
tlip and/or lac, Cho, no NAA, Cr
*NAA = N-acetyl aspartate; Cho = choline; Cr = creatine; lac = lactate; lip = lipids; t increase; 1 decrease; I marked increase.
nal intensity ratio of the Cho signal from cystic lesions to normal contralateral brain was also calculated.
r
1 3:s
3.0
2:5
Chemical
Lo
shift
1:5
ii0
0.5
/ ppm
Fig. 1. Normal brain spectrum from a VOI of 2 x 2 x 2 cm3 from the left parietal region using STEAM 20 ms (NAA = N-acetyl aspartate; Cr = total creatine; CHO = total choline; mI = my0 inositol).
6 low-grade gliomas (grade I-II), 7 epidermoid cysts, 7 pyogenic brain abscesses, 2 metastases, 2 arachnoid cysts, and 1 case each of schwannoma and meningioma. The final diagnosis in all these cases was based on the results of surgery and histopathology. In pediatric cases (1 - 10 yr) sedation was given in appropriate doses. There was no ciinical or laboratory evidence of infection in 4 of 7 patients with brain abscess; they presented only with signs of raised intracranial tension. Comparison of the spectroscopic results from the cystic lesions was either made from the contralateral uninvolved tissue or with 30 age-matched normal controls. To evaluate the tumor grade, the integrals of NAA and Cho signal intensity (NAA/Cho) were calculated in both low-grade and high-grade gliomas. The results were compared with the normal NAA/Cho ratio. To see the difference in the Cho peak from cystic neoplasms from that in the normal contralateral brain, sig-
Magnetic Resonance Imaging and In Vivo Spectroscopy MRI and in vivo proton MRS were performed on a 2T whole-body system (Magnetom, Siemens, Germany) operating at a field strength of 1.5 T using a circularly polarized head coil. MRI. Routine spin echo (SE) T,-(TRITE = 22001 2080) and T,-weighted (TR/TE = 660/15) images were obtained in the axial coronal and sagittal planes. Lesions appearing cystic on MRI were subdivided into intra-axial or extra-axial, depending upon the location of the mass.* The inclusion criteria for in vivo MRS was a lesion size mass of more than 8 ml. A volume of interest (VOI) of 4.09-8 ml was selected, for in vivo spectroscopy, from these images. The VOI was chosen within the confines of the lesion and was confirmed by taking turbo FLASH images (TR/TE/TI/FA = 65.3/ 3/500/P) in the three orthogonal planes. In vivo proton MRS. Volume-selective spectroscopy was performed using STEAM localizing sequence’ with 3 water-suppression chemical shift selective pulses (Band width = 60 Hz). The STEAM sequence parameters used in the present study were TRITE/TM/n = 3000/20,270/27.5/128. In 11 cases, SE sequence using TR/TE/n = 3000/135/128 was also performed. Sequences with different echo times were run: (1) to identify resonances with short Tz values and, (2) to observe and confirm the phase reversal shown by J coupled me-
Cystic intracranial mass lesions, role of MRS 0 H. POPTANI
1021
ET AL.
Fig. 2. High-gradeglioma (Glioblastomamultiforme). (A) T,-weightedaxial imageshowsa well-definedlargehypointenselesionin the left front0 parietal regionextendingto the basalganglia,causingmidlineshift. The box in the centerrepresentsthe VOI. (B) STEAM 20 msspectrumshowsa large resonanceat 1.3ppm and smallresonances at 2.02 ppm and 3.22 ppm assignedto NAA and CHO. At STEAM 270ms the resonanceat 1.3ppm showsa doublet of lactate (LIP = lipid; LAC = lactate; rest as in Fig. 1).
STEAM ‘0
STEAM 270
LIP 1 LAC
LAC
I -1,
3.5
3.0
2.5
2.0
Chemical shift
1.5
1.0
3.5
0:5
/ ppm 03
3.0
2.5
2.0
Chemical shift
1.5
1.0
/ ppm
a.5
‘
Magnetic Resonance Imaging 0 Volume 13,
1022
7, 1995
Ex Vivo Proton Magnetic Resonance Spectroscopy In five cases of brain abscess, high-resolution NMR was performed to confirm the in vivo assignment of the resonance peaks. The sample preparation included immediate freezing of the aspirated fluid at -20°C. The sample was then centrifuged at 3000 rpm for 10 min at 4°C and the supernatant was collected. 450 ~1of the supernatant was taken in a 5 mm NMR tube and 50 ~1 of DzO (containing 7.5% TSP) was added to make 500 pl. High-resolution NMR was performed on a 400 MHz spectrometer (Bruker’s, Switzerland) using a pulse collect sequence with a TR of 4 s, pulse angle 65”, and 256 data collects. Water suppression was achieved by using presaturation pulses. In addition, spin echo fourier transformed (SEFT) spectra were also recorded with TE = 100 ms and TR = 3 s, to see the phase reversal of the J coupled multiplets.
(A) STEAM
Number
20
RESULTS CHO
II
7-m
3.5
310
NAA
215
210
Chemical shift (B)
Controls The in vivo proton MR spectrum of the normal volunteer at STEAM 20 ms (Fig. 1) shows major resonances of NAA (2.02 ppm), Cho (3.22 ppm), Cr (3.02 ppm), and inositols (m1) at 3.56 ppm. Assignments of resonances is based on the previously assigned chemical shifts.” The mean NAA/Cho ratio in 30 cases was 1.85 + .15 (1 SD).
1A
115
110
015
/ ppm
Fig. 3. Low-grade glioma (Astrocytoma Grade I) (A) T,weightedsagittalimageshowswell-definedcysticlesionin the brain stem. (B) STEAM 20 msspectrumshowsresonances from CHO, NAA, and lactate.
tabolites. Voxel shimming of the free induction decay (FID) signal without water suppression yielded a fullwidth half maxium of 4-6 Hz. The total time taken for MRI and MRS ranged between 70-80 minutes. Postprocessing of the FID included zero filling to 4 K data points, followed by line broadening with gaussian multiplication, Fourier transformation, and phase correction. No baseline correction was done. Assignment of resonances was based on the previously reported values and high-resolution in vitro NMR study.lO,”
Cystic Mass Lesions Presence of Cho resonance was a consistent finding in all primary brain neoplasms. However, abnormal to normal choline ratio was less than 1 in all 16 cases (14 gliomas, 1 meningioma, 1 schwannoma). The results of the in vivo spectroscopic data in various cystic intracranial mass lesions is summarised in Table 1. Intra axial High-grade gliomas (n = 8) showed a large resonance assigned to lipid and/or lactate at 1.3 ppm (Fig. 2), in addition to resonances from NAA and Cho. The mean NAA/Cho ratio was 1.14 + 0.75 (1 SD). Low-grade gliomas (n = 6) showed lactate at 1.3 ppm along with resonances of NAA and Cho (Fig. 3). The mean NAA/Cho ratio in these cases was 0.43 f 0.15. Both cases of cystic metastases showed only lipids and/or lactate at 1.3 ppm with no resonances of NAA, Cho, or Cr (Fig. 4). Pyogenic brain abscesses (n = 7) showed large resonances from lactate (1.3 ppm), alanine (1.5 ppm), acetate (1.92 ppm), and cytosolic amino acids (0.9 ppm) with the absence of NAA, Cr, and Cho. At SE 135 ms,
Cystic intracranial mass lesions, role of MRS 0 H. POPTANI ET
STEAM
1023
AL.
270
LIP/LAG
STEAM
20
Fig. 4. Metastatic adenocarcinoma. (A) T,-weighted axial image showstwo well-defined thickwalled cystic lesions with associated perifocal edema in left periventricular and occipital regions. The box in the center of the lesion in the periventricular region represents the VOI. (B) 20 ms STEAM spectrum showsa large resonance at 1.3 ppm from lipid/lactate. At 270 ms a doublet at 1.3 ppm is clearly seen. 1 3.5
lactate, alanine, and cytosolic amino acids were inverted compared to acetate (Fig. 5b). Ex vivo high-resolution NMR spectrum confirmed the in vivo assignments (Fig. SC). The large resonance seen in vivo at 0.9 ppm is due to the presence of valine and leucine, which are coupled multiplets and show phase reversal on SEFT 100 ms (Fig. 5~). Extra Axial All 7 cases of epidermoid cyst showed a consistent pattern. These showed a resonance of lactate at 1.3 ppm and a small unassigned resonance at 1.8 ppm at 20 ms.
2.5 Chemical
3.0
2.0 shift
1:5
1:o / ppm
0:5
I
At STEAM 270 ms only lactate resonance was observed, suggesting that the resonance at 1.8 ppm has a shorter T2 relaxation time than lactate (Fig. 6). Only minimal lactate was seen in 2 cases of arachnoid cyst (Fig. 7). A case of cystic meningioma showed a strong unassigned broad resonance at 2.04 ppm along with Cho, lactate, and alanine (Fig. 8). At SE 135 ms spectrum the resonance at 2.04 ppm disappeared and the alanine and lactate peaks were inverted. Compared to meningioma, cystic schwannoma (n = 1) showed lipid/lactate and choline resonances.
1024
Magnetic ResonanceImaging 0 Volume 13, Number 7, 1995
STEAM
270
SE 135
LAC
STEAM
20
Fig. 5. Pyogenic brain abscess. (A) T,-weighted axial image shows a cystic lesion in the right parietal region with associated perifocal edema. (B) STEAM 20 ms spectrum shows resonances from lactate (1.3 ppm), alanine (1.5 ppm), acetate (1.92 ppm), and cytosolic amino acids (0.9 ppm). At 135 ms SE, inversion of ala, lac, and cytosolic amino acids is seen in relation to acetate. STEAM 270 ms spectrum shows the same resonances with reduced signal intensity (AL = alanine; AC = acetate; CY = cytosolic amino acids; rest as in Figs. 1 and 2). (C) Ex vivo pulse collect spectrum of the pus confirms the in vivo assignments. Ex vivo SEFT 100 shows the phase reversal of alanine, lactate, and cytosolic amino acids (valine, leucine). (Figure continues)
,
3.5
3.0
2.5
Chemical
2.0
shift
1.5
1.0
/ ppm
0.5
r
Cystic intracranial mass lesions, role of MRS 0 H. POPTANI ET
1025
AL.
SE100
*b
I’v”Im 4.5
8”1”” 4.0
3.5
I “‘I 3.0
1 “I’ 2.5
1 “‘I 2.0
LAC
1 “I’ 1 .s
1 I”’ 1.0
1’ .S
PPtl ((3
Fig. 5 continued.
DISCUSSION
A large Cho resonance with variable NAA, Cr, and lipid/lactate are usual in vivo proton MR spectroscopic findings in cases of predominantly solid brain tumors. There is more choline in tumors compared to normal contralateral brain, with the abnormal-to-normal cho-
line ratio more than 1 in solid gliomas3 However, in the present series with primary cystic neoplasms, a small resonance of choline with abnormal-to-normal choline ratio of less than 1 was observed. The low choline in cystic neoplasms appears to be due to the degraded free pool of choline detected in the cystic fluid as a part of necrosis, and cystic degeneration in primary
1026
Magnetic Resonance Imaging 0 Volume 13, Number 7, 1995
Fig. 6. Epidermoid cyst. (A) T,-weighted axial image shows a mixed intensity mass in the left parasagittal region with mass effect on lateral ventricles. (B) STEAM 20 ms spectrum shows lactate at 1.3 ppm and a resonanceat 1.8 ppm that disappears at STEAM 270 ms. (* = unassigned resonance).
(4 STEAM
20
STEAM
270
LAC
LAC
, 3:s
3:o
2:5
210
Chemical shift
1:s
1:o
/
cl:5
’
ppm
neoplasms with no active cell membrane turnover. The presence of choline in all cystic primary neoplasms appears to be a differentiating marker from other cystic masses, including metastases where no choline resonance is seen.
3.5
3.0
2.5
Chemical
2.0
shift
1.5
1.0
0.5
/ opm
Grading of different gliomas has been predicted on the basis of NAAKr, Cho/Cr, and lactate/NAA ratios.s*6 Recently, the NAA/Cho ratio was found to be correlating well with the grading of gliomas.3 No correlation has been observed in the present study in cys-
Cystic intracranial mass lesions, role of MRS 0 H. POPTANI ET AL.
1027
Fig. 7. Arachnoid cyst. (A) T,-weighted axial image shows a well-defined centrally placed cyst pushing the vermis anteriorly. (B) STEAM 20 ms spectrum shows only lac at 1.3 ppm, which is inverting at SE 135 ms.
(A) STEAM
SE 135
20 LAC
3.5
3.0 2.5 Chemical
210
shift
1.5
/
1.0 PPm
3.5
0.5
3.0 2.5 Chemical
2.0
shift
1:5
1.0
0.5
/ PPm
(B)
tic gliomas. In fact, the NAA/Cho ratio was lower in low-grade gliomas than in high-grade gliomas. Brain abscess can be confidently diagnosed on contrast-enhanced CT and MRI when associated with fever, leucocytosis, and the presence of a focus of infection in other parts of the body. CT and MRI features are nonspecific in 50% of the brain abscessesthat present clinically with only intracranial mass lesions.‘,‘* In the present series, 4 of the 7 patients with brain abscess presented with space-occupying lesions without signs of infections. Proton spectra were con-
sistent in all 4 cases and showed resonances from acetate, lactate, alanine, and cytosolic amino acids with no resonance from NAA, Cho, and C. Lactate and acetate in the abscess appear to be due to enhanced glycolysis along with fermentative pathways resulting in turnover of pyruvate to acetate and lactate. The cytosolic amino acids seen in vivo are actually the accumulated high concentration of valine and leucine, which are endproducts of the degradation of proteins by proteolytic enzymes released by the polymorphonucleocytes in the pus.13 It is surprising to note the similarity
Magnetic Resonance Imaging 0 Volume 13, Number 7, 1995
1028
Fig. 8. Meningioma. (A) Tr-weightedaxial imageshowsa well-definedcysticmassin theright frontal region.(B) STEAM 20 msspectrumshowsCho, ala, lac, and a broad resonance at 2.04ppm that remainsunassigned.At SE 135ms,the resonanceat 2.04ppm hasdisappearedandala and lac haveinverted. The resonancesfrom lactate and alanineare better definedon SE 135,dueto disappearance of metaboliteswith short T, values(* = unassigned).
(4 STEAM
20
SE 135
*
CHO CHO
3.5
310
2.5
Chemxal
2.0
shift
1.5
110
015
I
/ ppm
of spectra reported in a caseof brain abscesswith metastatic adenocarcinoma.4 This is in contradiction to what has been observed in our series, where we have confirmed in vivo results with ex vivo high-resolution study. In vivo proton MRS in a caseof hydatid cyst”
3:s
310
2:5
Chemical
2.0
shift
i-5
1.0
0:5
/ PPm
shows resonancesfrom pyruvate, acetate, lactate, and alanine. The presence of amino acids in the abscesses may help in their differentiation from hydatid cyst. AI1 epidermoids appearing hypointense on T, -weighted imagesshowed consistent spectrum (i.e., lactate and a
Cystic intracranial mass lesions, role of MRS 0 H.
small resonance at 1.8 ppm that remains unassigned). The presence of only minimal lactate in the arachnoid cyst, which may show similar intensity as that of epidermoid on imaging, can be differentiated on MR spectroscopy. Alanine has been shown by most workers in solid meningiomas and described as characteristic of this tumor.2*3,6 We demonstrated alanine even in a cystic meningioma, suggesting that it is a differentiating feature of meningiomas whether cystic or solid, from other extra axial lesions like cystic schwannoma and epidermoids. Presence of a broad unassigned resonance at 2.04 ppm at STEAM 20 ms in a meningioma appears to be a metabolite with short T2 value because it disappeared on SE 135 ms. Based on the above observations, it appears that pathologically dissimilar cystic mass lesions show different spectral patterns. We feel that combined MRI and in vivo proton MRS may help better characterize these lesions in the future. REFERENCES 1. Sartor, K. MR imaging of the skull and brain: A correlative text atlas. Heidelberg: Springer-Verlag; 1992. 2. Howe, F.A.; Maxwell, R.J.; Saunders, D.E.; Brown, M.M.; Griffiths, J.R. Proton Spectroscopy In vivo. Magn. Reson. Q. 9:31-59; 1993. 3. Kugel, H.; Heindel, W.; Ernestus, RI.; Bunke, J.; Mesnil, R.; Friedmann, G. Human brain tumors: spectral patterns detected with localised H-l MR spectroscopy. Radiology 183:701-709; 1993. 4. Ott, D.; Hennig. J.; Ernst, T. Human Brain Tumors: Assessment with in vivo proton MR spectroscopy. Radiology 186:745-752; 1993.
POPTANI ET AL.
1029
5. Arnold, D.L.; Shoubridge, E.A.; Villemeure, J.G.; Feindel, W. Proton and phosphorus magnetic resonance spectroscopy of human astrocytomas in vivo. Preliminary observations on tumor grading. NMR Biomed. 3:184189; 1990. 6. Gill, S.S.; Thomas, D.G.T.; Bruggen, N.V.; Gadian, D.G.; Peden, C.J.; Bell, J.D.; Cox, I.J.; Menon, D.K.; Iles, R.A.; Bryant, D.J.; Coutts, G.A. Proton MR spectroscopy of intracranial tumors: In vivo and in vitro studies. J. Comput. Assist. Tomogr. 14(4):497-504; 1990. 7. Bruhn, H.; Michaelis,T.; Merboldt, K.D.; Hanicke, W.; Gyngell, M.L.; Hamburger, C.; Frahm, J. On the interpretation of proton NMR spectra from brain tumors in vivo and in vitro. NMR Biomed. 5:253-258; 1992. 8. Curnes, J.T. MR imaging of peripheral intracranial neoplasms. Extraaxial vs intracraxial masses. J. Comput. Assit. Tomogr. 11(6):932-937; 1987. 9. Frahm, J.; Merboldt,K.D.; Hanicke, W. Localised proton spectroscopy using stimulated echoes. J. Magn. Reson. 72:502-508; 1987. 10. Michaelis, T.; Merboldt, K.D.; Hanicke, W.; Gyngell, M.L.; Bruhn, H.; Frahm, J. On the identification of cerebral metabolites in localized ‘HNMR spectra of human brain in vivo. NMR Biomed. 4:90-98; 1991. 11. Kohli, A.; Gupta, R.K.; Poptani, H.; Roy, R. In vivo proton MR spectroscopy (MRS) in a case of intracranial hydatid cyst. Neurology 45:562-564; 1995. 12. Klug, N.; Ellams, I.D. Difficulties in the differential diagnosis of brain abscesses. In: W. Schiefer, M. Klinger, M. Brock (Eds). Advances in Neurosurgery 9. Berlin: Springer-Verlag; 1981: pp. 61-67. 13. Mendz, G.L.; McCall, M.N.; Kuchel, P.W. Identification of methyl resonances in the 1H NMR Spectrum of incubated blood cell lysates. J. Biol. Chem. 264 (4):21002107; 1989.
Pergamon
Resonance Imaging, Vol. 13, No. 7, pp. 1019-1029, 1995 Copyright 0 1995 Elsevier Science Inc. Printed in the USA. All rights reserved 0730-725X/95 $9.50 + .OO
0730-725x(95)00045-3
0 Original Contribution CYSTIC INTRACRANIAL MASS LESIONS: POSSIBLE ROLE OF IN VIVO MR SPECTROSCOPY IN ITS DIFFERENTIAL DIAGNOSIS RAKESH K. GUPTA,* VIJENDERA K. JAIN,* RAJA ROY ,-f AND RAKESH PANDEY* *MR Section, Departments of Radiology, Neurosurgery, and Pathology, Sanjay Gandhi Post Graduate Institute of Medical Sciences Lucknow 226014, India, and tR.S.1.C. Central Derug Research Institute, Lucknow, India HARISH
POPTANI,*
Thirty-four patients showing cystic intracranial mass lesions on MR imaging were evaluated by in vivo proton MR spectroscopy (MRS) with the aim of detecting lesion-specific spectral patterns that may assist imaging in better tissue characterization. In vlvo spectroscopy was performed using stimulated echo acquisition mode with echo times 20 m and 270 m in all, and spin echo with echo time 135 m in 11 patients. All primary neoplasms (in&a-as well as extra-axial) showed choline (3.22 ppm) resonance along with lipid and/or lactate (1.3 ppm). It was not possible to grade cystic gliomas based on N-ace@ asparate-to-choline ratio. High-grade gliomas (n = 8) showed lipid/lactate and low-grade gliomas (?I = 6) showed only lactate. Seven patients with brain abscess showed resonances only from acetate (1.92 ppm), lactate (1.3 ppm) and alanine (1.5 ppm). Two cases of metastatic adenocarcinoma showed only lipid/lactate. In 7 patients with epidermoid cyst, lactate along with an unassigned resonance at 1.8 ppm was observed and could be easily differentiated from arachnoid cyst (n = 2), which showed only minimal lactate. A case of cystic meningioma could be differentiated from cystic schowannoma by the presence of alanine in the former. It is concluded that MR imaging, when combined with in vivo MRS, may help to better characterize intracranial cystic mass lesions.
Keywords: Brain; Brain tumors; Magnetic resonance spectroscopy; Brain abscess. INTRODUCTION
though differences in metabolites between cystic and solid components of the tumor have been reported,’ there is no report describing in vivo proton MR spectroscopy, primarily in cystic lesions of the brain. In the present study, in vivo proton MRS was performed in 34 cases of intracranial cystic lesions with the aim (1) to study the spectral patterns in different types of cystic lesions and (2) to look for any lesionspecific spectral pattern that may assist MRI in its differential diagnosis in the future.
Magnetic resonance imaging (MRI) is an established technique for the evaluation of intracranial mass lesions. It is highly sensitive, yet relatively nonspecific.’ Cystic, intracranial intra-axial mass lesions usually seen on MRI include gliomas, metastases, abscesses and, rarely, hydatid cyst; cystic extra-axial lesions include meningioma, schwannoma, epidermoid, and arachnoid cyst. The characterisation of these lesions on MRI is not always possible.’ In vivo proton magnetic resonance spectroscopy (MRS) has been performed in different intracranial lesions with the aim of helping to better characterise these lesions.2 Brain tumors usually show high choline containing compounds (Cho), low N-acetyl aspartate (NAA), and the presence of lipid (lip) or lactate (lac).3Y4 An attempt has been made to grade the degree of malignancy of different gliomas based on the difference in quantification of these biochemical signals.3-6 Al-
MATERIALS
AND METHODS
Patients Thirty-four patients (26 males and 8 females) with intracranial cystic mass lesions suggested on MRI, were evaluated with in vivo proton MRS. Their ages ranged between l-65 years. Informed consent was taken from all patients/guardians prior to the study. Of the 34 cases, there were 8 high-grade gliomas (grade III-IV), sociate Professor, MR Section, Dept. of Radiology, Sanjay
RECEIVED 12/S/94; ACCEPTED S/16/95. Address correspondence to Rakesh K. Gupta, MD, As-
Gandhi PGIMS, Lucknow 1019
226014, India.
1020
Magnetic Resonance Imaging l Volume 13, Number 7, 1995
STEAM
20
Table 1. Summary
Tumor type
NAA
Low-grade gliomas (n = 6) High-grade gliomas (n = 8) Metastases (n = 2) Abscesses (n = 7) Epidermoid cyst
Cr I
(n = 7)
CHO,
of results
Spectroscopic
findings*
t lac; 1 NAA, Cho, Cr NAAKho = 0.43 f 0.15 fi lip and/or lac; 1 NAA, Cho, Cr NAA/Cho = 1.14 f 0.75 1 lip and/or
lac
no NAA, Cho, Cr Tlac, ala, actetate, cystosolic amino acids, no NAA, Cho, Cr tlac, unassigned peak (1.8 ppm) no NAA, Cho, Cr Tlac, no NAA, Cho, Cr
Arachnoid cyst (n = 2) Meningioma (n = 1)
r?unassigned (2.04 ppm), t Cho, lac, ala, no NAA, Cr
Schwannoma (n= 1)
tlip and/or lac, Cho, no NAA, Cr
*NAA = N-acetyl aspartate; Cho = choline; Cr = creatine; lac = lactate; lip = lipids; t increase; 1 decrease; I marked increase.
nal intensity ratio of the Cho signal from cystic lesions to normal contralateral brain was also calculated.
r
1 3:s
3.0
2:5
Chemical
Lo
shift
1:5
ii0
0.5
/ ppm
Fig. 1. Normal brain spectrum from a VOI of 2 x 2 x 2 cm3 from the left parietal region using STEAM 20 ms (NAA = N-acetyl aspartate; Cr = total creatine; CHO = total choline; mI = my0 inositol).
6 low-grade gliomas (grade I-II), 7 epidermoid cysts, 7 pyogenic brain abscesses, 2 metastases, 2 arachnoid cysts, and 1 case each of schwannoma and meningioma. The final diagnosis in all these cases was based on the results of surgery and histopathology. In pediatric cases (1 - 10 yr) sedation was given in appropriate doses. There was no ciinical or laboratory evidence of infection in 4 of 7 patients with brain abscess; they presented only with signs of raised intracranial tension. Comparison of the spectroscopic results from the cystic lesions was either made from the contralateral uninvolved tissue or with 30 age-matched normal controls. To evaluate the tumor grade, the integrals of NAA and Cho signal intensity (NAA/Cho) were calculated in both low-grade and high-grade gliomas. The results were compared with the normal NAA/Cho ratio. To see the difference in the Cho peak from cystic neoplasms from that in the normal contralateral brain, sig-
Magnetic Resonance Imaging and In Vivo Spectroscopy MRI and in vivo proton MRS were performed on a 2T whole-body system (Magnetom, Siemens, Germany) operating at a field strength of 1.5 T using a circularly polarized head coil. MRI. Routine spin echo (SE) T,-(TRITE = 22001 2080) and T,-weighted (TR/TE = 660/15) images were obtained in the axial coronal and sagittal planes. Lesions appearing cystic on MRI were subdivided into intra-axial or extra-axial, depending upon the location of the mass.* The inclusion criteria for in vivo MRS was a lesion size mass of more than 8 ml. A volume of interest (VOI) of 4.09-8 ml was selected, for in vivo spectroscopy, from these images. The VOI was chosen within the confines of the lesion and was confirmed by taking turbo FLASH images (TR/TE/TI/FA = 65.3/ 3/500/P) in the three orthogonal planes. In vivo proton MRS. Volume-selective spectroscopy was performed using STEAM localizing sequence’ with 3 water-suppression chemical shift selective pulses (Band width = 60 Hz). The STEAM sequence parameters used in the present study were TRITE/TM/n = 3000/20,270/27.5/128. In 11 cases, SE sequence using TR/TE/n = 3000/135/128 was also performed. Sequences with different echo times were run: (1) to identify resonances with short Tz values and, (2) to observe and confirm the phase reversal shown by J coupled me-
Cystic intracranial mass lesions, role of MRS 0 H. POPTANI
1021
ET AL.
Fig. 2. High-gradeglioma (Glioblastomamultiforme). (A) T,-weightedaxial imageshowsa well-definedlargehypointenselesionin the left front0 parietal regionextendingto the basalganglia,causingmidlineshift. The box in the centerrepresentsthe VOI. (B) STEAM 20 msspectrumshowsa large resonanceat 1.3ppm and smallresonances at 2.02 ppm and 3.22 ppm assignedto NAA and CHO. At STEAM 270ms the resonanceat 1.3ppm showsa doublet of lactate (LIP = lipid; LAC = lactate; rest as in Fig. 1).
STEAM ‘0
STEAM 270
LIP 1 LAC
LAC
I -1,
3.5
3.0
2.5
2.0
Chemical shift
1.5
1.0
3.5
0:5
/ ppm 03
3.0
2.5
2.0
Chemical shift
1.5
1.0
/ ppm
a.5
‘
Magnetic Resonance Imaging 0 Volume 13,
1022
7, 1995
Ex Vivo Proton Magnetic Resonance Spectroscopy In five cases of brain abscess, high-resolution NMR was performed to confirm the in vivo assignment of the resonance peaks. The sample preparation included immediate freezing of the aspirated fluid at -20°C. The sample was then centrifuged at 3000 rpm for 10 min at 4°C and the supernatant was collected. 450 ~1of the supernatant was taken in a 5 mm NMR tube and 50 ~1 of DzO (containing 7.5% TSP) was added to make 500 pl. High-resolution NMR was performed on a 400 MHz spectrometer (Bruker’s, Switzerland) using a pulse collect sequence with a TR of 4 s, pulse angle 65”, and 256 data collects. Water suppression was achieved by using presaturation pulses. In addition, spin echo fourier transformed (SEFT) spectra were also recorded with TE = 100 ms and TR = 3 s, to see the phase reversal of the J coupled multiplets.
(A) STEAM
Number
20
RESULTS CHO
II
7-m
3.5
310
NAA
215
210
Chemical shift (B)
Controls The in vivo proton MR spectrum of the normal volunteer at STEAM 20 ms (Fig. 1) shows major resonances of NAA (2.02 ppm), Cho (3.22 ppm), Cr (3.02 ppm), and inositols (m1) at 3.56 ppm. Assignments of resonances is based on the previously assigned chemical shifts.” The mean NAA/Cho ratio in 30 cases was 1.85 + .15 (1 SD).
1A
115
110
015
/ ppm
Fig. 3. Low-grade glioma (Astrocytoma Grade I) (A) T,weightedsagittalimageshowswell-definedcysticlesionin the brain stem. (B) STEAM 20 msspectrumshowsresonances from CHO, NAA, and lactate.
tabolites. Voxel shimming of the free induction decay (FID) signal without water suppression yielded a fullwidth half maxium of 4-6 Hz. The total time taken for MRI and MRS ranged between 70-80 minutes. Postprocessing of the FID included zero filling to 4 K data points, followed by line broadening with gaussian multiplication, Fourier transformation, and phase correction. No baseline correction was done. Assignment of resonances was based on the previously reported values and high-resolution in vitro NMR study.lO,”
Cystic Mass Lesions Presence of Cho resonance was a consistent finding in all primary brain neoplasms. However, abnormal to normal choline ratio was less than 1 in all 16 cases (14 gliomas, 1 meningioma, 1 schwannoma). The results of the in vivo spectroscopic data in various cystic intracranial mass lesions is summarised in Table 1. Intra axial High-grade gliomas (n = 8) showed a large resonance assigned to lipid and/or lactate at 1.3 ppm (Fig. 2), in addition to resonances from NAA and Cho. The mean NAA/Cho ratio was 1.14 + 0.75 (1 SD). Low-grade gliomas (n = 6) showed lactate at 1.3 ppm along with resonances of NAA and Cho (Fig. 3). The mean NAA/Cho ratio in these cases was 0.43 f 0.15. Both cases of cystic metastases showed only lipids and/or lactate at 1.3 ppm with no resonances of NAA, Cho, or Cr (Fig. 4). Pyogenic brain abscesses (n = 7) showed large resonances from lactate (1.3 ppm), alanine (1.5 ppm), acetate (1.92 ppm), and cytosolic amino acids (0.9 ppm) with the absence of NAA, Cr, and Cho. At SE 135 ms,
Cystic intracranial mass lesions, role of MRS 0 H. POPTANI ET
STEAM
1023
AL.
270
LIP/LAG
STEAM
20
Fig. 4. Metastatic adenocarcinoma. (A) T,-weighted axial image showstwo well-defined thickwalled cystic lesions with associated perifocal edema in left periventricular and occipital regions. The box in the center of the lesion in the periventricular region represents the VOI. (B) 20 ms STEAM spectrum showsa large resonance at 1.3 ppm from lipid/lactate. At 270 ms a doublet at 1.3 ppm is clearly seen. 1 3.5
lactate, alanine, and cytosolic amino acids were inverted compared to acetate (Fig. 5b). Ex vivo high-resolution NMR spectrum confirmed the in vivo assignments (Fig. SC). The large resonance seen in vivo at 0.9 ppm is due to the presence of valine and leucine, which are coupled multiplets and show phase reversal on SEFT 100 ms (Fig. 5~). Extra Axial All 7 cases of epidermoid cyst showed a consistent pattern. These showed a resonance of lactate at 1.3 ppm and a small unassigned resonance at 1.8 ppm at 20 ms.
2.5 Chemical
3.0
2.0 shift
1:5
1:o / ppm
0:5
I
At STEAM 270 ms only lactate resonance was observed, suggesting that the resonance at 1.8 ppm has a shorter T2 relaxation time than lactate (Fig. 6). Only minimal lactate was seen in 2 cases of arachnoid cyst (Fig. 7). A case of cystic meningioma showed a strong unassigned broad resonance at 2.04 ppm along with Cho, lactate, and alanine (Fig. 8). At SE 135 ms spectrum the resonance at 2.04 ppm disappeared and the alanine and lactate peaks were inverted. Compared to meningioma, cystic schwannoma (n = 1) showed lipid/lactate and choline resonances.
1024
Magnetic ResonanceImaging 0 Volume 13, Number 7, 1995
STEAM
270
SE 135
LAC
STEAM
20
Fig. 5. Pyogenic brain abscess. (A) T,-weighted axial image shows a cystic lesion in the right parietal region with associated perifocal edema. (B) STEAM 20 ms spectrum shows resonances from lactate (1.3 ppm), alanine (1.5 ppm), acetate (1.92 ppm), and cytosolic amino acids (0.9 ppm). At 135 ms SE, inversion of ala, lac, and cytosolic amino acids is seen in relation to acetate. STEAM 270 ms spectrum shows the same resonances with reduced signal intensity (AL = alanine; AC = acetate; CY = cytosolic amino acids; rest as in Figs. 1 and 2). (C) Ex vivo pulse collect spectrum of the pus confirms the in vivo assignments. Ex vivo SEFT 100 shows the phase reversal of alanine, lactate, and cytosolic amino acids (valine, leucine). (Figure continues)
,
3.5
3.0
2.5
Chemical
2.0
shift
1.5
1.0
/ ppm
0.5
r
Cystic intracranial mass lesions, role of MRS 0 H. POPTANI ET
1025
AL.
SE100
*b
I’v”Im 4.5
8”1”” 4.0
3.5
I “‘I 3.0
1 “I’ 2.5
1 “‘I 2.0
LAC
1 “I’ 1 .s
1 I”’ 1.0
1’ .S
PPtl ((3
Fig. 5 continued.
DISCUSSION
A large Cho resonance with variable NAA, Cr, and lipid/lactate are usual in vivo proton MR spectroscopic findings in cases of predominantly solid brain tumors. There is more choline in tumors compared to normal contralateral brain, with the abnormal-to-normal cho-
line ratio more than 1 in solid gliomas3 However, in the present series with primary cystic neoplasms, a small resonance of choline with abnormal-to-normal choline ratio of less than 1 was observed. The low choline in cystic neoplasms appears to be due to the degraded free pool of choline detected in the cystic fluid as a part of necrosis, and cystic degeneration in primary
1026
Magnetic Resonance Imaging 0 Volume 13, Number 7, 1995
Fig. 6. Epidermoid cyst. (A) T,-weighted axial image shows a mixed intensity mass in the left parasagittal region with mass effect on lateral ventricles. (B) STEAM 20 ms spectrum shows lactate at 1.3 ppm and a resonanceat 1.8 ppm that disappears at STEAM 270 ms. (* = unassigned resonance).
(4 STEAM
20
STEAM
270
LAC
LAC
, 3:s
3:o
2:5
210
Chemical shift
1:s
1:o
/
cl:5
’
ppm
neoplasms with no active cell membrane turnover. The presence of choline in all cystic primary neoplasms appears to be a differentiating marker from other cystic masses, including metastases where no choline resonance is seen.
3.5
3.0
2.5
Chemical
2.0
shift
1.5
1.0
0.5
/ opm
Grading of different gliomas has been predicted on the basis of NAAKr, Cho/Cr, and lactate/NAA ratios.s*6 Recently, the NAA/Cho ratio was found to be correlating well with the grading of gliomas.3 No correlation has been observed in the present study in cys-
Cystic intracranial mass lesions, role of MRS 0 H. POPTANI ET AL.
1027
Fig. 7. Arachnoid cyst. (A) T,-weighted axial image shows a well-defined centrally placed cyst pushing the vermis anteriorly. (B) STEAM 20 ms spectrum shows only lac at 1.3 ppm, which is inverting at SE 135 ms.
(A) STEAM
SE 135
20 LAC
3.5
3.0 2.5 Chemical
210
shift
1.5
/
1.0 PPm
3.5
0.5
3.0 2.5 Chemical
2.0
shift
1:5
1.0
0.5
/ PPm
(B)
tic gliomas. In fact, the NAA/Cho ratio was lower in low-grade gliomas than in high-grade gliomas. Brain abscess can be confidently diagnosed on contrast-enhanced CT and MRI when associated with fever, leucocytosis, and the presence of a focus of infection in other parts of the body. CT and MRI features are nonspecific in 50% of the brain abscessesthat present clinically with only intracranial mass lesions.‘,‘* In the present series, 4 of the 7 patients with brain abscess presented with space-occupying lesions without signs of infections. Proton spectra were con-
sistent in all 4 cases and showed resonances from acetate, lactate, alanine, and cytosolic amino acids with no resonance from NAA, Cho, and C. Lactate and acetate in the abscess appear to be due to enhanced glycolysis along with fermentative pathways resulting in turnover of pyruvate to acetate and lactate. The cytosolic amino acids seen in vivo are actually the accumulated high concentration of valine and leucine, which are endproducts of the degradation of proteins by proteolytic enzymes released by the polymorphonucleocytes in the pus.13 It is surprising to note the similarity
Magnetic Resonance Imaging 0 Volume 13, Number 7, 1995
1028
Fig. 8. Meningioma. (A) Tr-weightedaxial imageshowsa well-definedcysticmassin theright frontal region.(B) STEAM 20 msspectrumshowsCho, ala, lac, and a broad resonance at 2.04ppm that remainsunassigned.At SE 135ms,the resonanceat 2.04ppm hasdisappearedandala and lac haveinverted. The resonancesfrom lactate and alanineare better definedon SE 135,dueto disappearance of metaboliteswith short T, values(* = unassigned).
(4 STEAM
20
SE 135
*
CHO CHO
3.5
310
2.5
Chemxal
2.0
shift
1.5
110
015
I
/ ppm
of spectra reported in a caseof brain abscesswith metastatic adenocarcinoma.4 This is in contradiction to what has been observed in our series, where we have confirmed in vivo results with ex vivo high-resolution study. In vivo proton MRS in a caseof hydatid cyst”
3:s
310
2:5
Chemical
2.0
shift
i-5
1.0
0:5
/ PPm
shows resonancesfrom pyruvate, acetate, lactate, and alanine. The presence of amino acids in the abscesses may help in their differentiation from hydatid cyst. AI1 epidermoids appearing hypointense on T, -weighted imagesshowed consistent spectrum (i.e., lactate and a
Cystic intracranial mass lesions, role of MRS 0 H.
small resonance at 1.8 ppm that remains unassigned). The presence of only minimal lactate in the arachnoid cyst, which may show similar intensity as that of epidermoid on imaging, can be differentiated on MR spectroscopy. Alanine has been shown by most workers in solid meningiomas and described as characteristic of this tumor.2*3,6 We demonstrated alanine even in a cystic meningioma, suggesting that it is a differentiating feature of meningiomas whether cystic or solid, from other extra axial lesions like cystic schwannoma and epidermoids. Presence of a broad unassigned resonance at 2.04 ppm at STEAM 20 ms in a meningioma appears to be a metabolite with short T2 value because it disappeared on SE 135 ms. Based on the above observations, it appears that pathologically dissimilar cystic mass lesions show different spectral patterns. We feel that combined MRI and in vivo proton MRS may help better characterize these lesions in the future. REFERENCES 1. Sartor, K. MR imaging of the skull and brain: A correlative text atlas. Heidelberg: Springer-Verlag; 1992. 2. Howe, F.A.; Maxwell, R.J.; Saunders, D.E.; Brown, M.M.; Griffiths, J.R. Proton Spectroscopy In vivo. Magn. Reson. Q. 9:31-59; 1993. 3. Kugel, H.; Heindel, W.; Ernestus, RI.; Bunke, J.; Mesnil, R.; Friedmann, G. Human brain tumors: spectral patterns detected with localised H-l MR spectroscopy. Radiology 183:701-709; 1993. 4. Ott, D.; Hennig. J.; Ernst, T. Human Brain Tumors: Assessment with in vivo proton MR spectroscopy. Radiology 186:745-752; 1993.
POPTANI ET AL.
1029
5. Arnold, D.L.; Shoubridge, E.A.; Villemeure, J.G.; Feindel, W. Proton and phosphorus magnetic resonance spectroscopy of human astrocytomas in vivo. Preliminary observations on tumor grading. NMR Biomed. 3:184189; 1990. 6. Gill, S.S.; Thomas, D.G.T.; Bruggen, N.V.; Gadian, D.G.; Peden, C.J.; Bell, J.D.; Cox, I.J.; Menon, D.K.; Iles, R.A.; Bryant, D.J.; Coutts, G.A. Proton MR spectroscopy of intracranial tumors: In vivo and in vitro studies. J. Comput. Assist. Tomogr. 14(4):497-504; 1990. 7. Bruhn, H.; Michaelis,T.; Merboldt, K.D.; Hanicke, W.; Gyngell, M.L.; Hamburger, C.; Frahm, J. On the interpretation of proton NMR spectra from brain tumors in vivo and in vitro. NMR Biomed. 5:253-258; 1992. 8. Curnes, J.T. MR imaging of peripheral intracranial neoplasms. Extraaxial vs intracraxial masses. J. Comput. Assit. Tomogr. 11(6):932-937; 1987. 9. Frahm, J.; Merboldt,K.D.; Hanicke, W. Localised proton spectroscopy using stimulated echoes. J. Magn. Reson. 72:502-508; 1987. 10. Michaelis, T.; Merboldt, K.D.; Hanicke, W.; Gyngell, M.L.; Bruhn, H.; Frahm, J. On the identification of cerebral metabolites in localized ‘HNMR spectra of human brain in vivo. NMR Biomed. 4:90-98; 1991. 11. Kohli, A.; Gupta, R.K.; Poptani, H.; Roy, R. In vivo proton MR spectroscopy (MRS) in a case of intracranial hydatid cyst. Neurology 45:562-564; 1995. 12. Klug, N.; Ellams, I.D. Difficulties in the differential diagnosis of brain abscesses. In: W. Schiefer, M. Klinger, M. Brock (Eds). Advances in Neurosurgery 9. Berlin: Springer-Verlag; 1981: pp. 61-67. 13. Mendz, G.L.; McCall, M.N.; Kuchel, P.W. Identification of methyl resonances in the 1H NMR Spectrum of incubated blood cell lysates. J. Biol. Chem. 264 (4):21002107; 1989.