View of Role of MR Spectroscopy in Evaluation of Intraaxial Brain Tumors

(1)

1076 http://annalsofrscb.ro

Role of MR Spectroscopy in Evaluation of Intraaxial Brain Tumors

YashBhutada¹, G Balachandran²

1,2 Department of Radiodiagnosis, Sri Lakshmi Narayana Institute of Medical Sciences Affiliated to Bharath Institute of Higher Education and Research, Chennai, Tamil Nadu, India.

ABSTRACT

To define biochemical markers of intraaxialbrain tumors by means of MRspectroscopy.To estimate role of MR spectroscopy in diagnosing and grading of intraaxialbrain tumors with histopathologicalco-relation.To estimate role of MR spectroscopy in determining the infiltrative nature of the intra axial braintumor.

Keywords:

MR spectroscopy,axial braintumor

1. Introduction

Intra axial brain masses are a noteworthy health problem and present numerous imaging challenges. These lesions comprise of primary neoplasm, secondary neoplasm, tumefactive demyelinating lesions, lymphoma, encephalitis and abscesses. [1-3]In intracranial tumor management, imaging plays vital integral role. Magnetic resonance (MR) imaging in specific has developed as the imaging modality most often used to assess intracranial tumors. Importance of MR imaging in the examination of intraaxial tumors can be split into tumor diagnosis with classification, treatment planning and post treatment scrutiny. The advanced MR techniques have developed which offer more than the anatomic information provided by the conventional MR imaging sequences. [4] They produce physiologic data and information on chemical composition.

The current advanced techniques comprise of perfusion imaging, diffusion-weighted imaging, MR spectroscopy, blood oxygen level–dependent (BOLD) imaging and molecularimaging.[5]With only anatomic imaging, distinction between extra axial and intra axial brain tumors is simple; though, the major diagnostic task isnoninvasively and precisely differentiate intraaxial tumors to avoid biopsy and follow-up imaging studies. Incorporation of diagnostic information from advanced magnetic resonance (MR) imaging techniques can further enrich the classification accuracy of conventional anatomic imaging.[6]

MR spectroscopy permits the non-invasive calculation of selected biological compounds in vivo.

Proton spectroscopy has been acknowledged as a secure and noninvasive diagnostic method.

Proton spectroscopy when combined with MRI techniques, [7-9] permits for the association of anatomical and physiological variation in the metabolic and biochemical processes taking place within the previously determined volumes in the brain. MR spectroscopy provides information about the likely extent and nature of changes on a routine MRI scan by examining the presence or ratio of tissue metabolites such as NAA, creatine, choline, and lactateetc.[10]

Extensive usage of quicker MR spectroscopy applications with higher signal-to-noise ratio (SNR) and spatial resolution, allows us to detect functional metabolic changes, which provides more data to recognize the precise nature of the tumor and the morphological and physiological changes occurring in the adjacent brain parenchyma. Longitudinal studies have established that MRS is useful in monitoring disease evolution and treatment effects. MR spectroscopy also has a prognostic implication.[11-13]

(2)

2. Methodology

The study was carried out at the Department of Radiodiagnosis, Sri Lakshmi Narayana Institute of Medical Sciences from December 2018 to July 2020 with aim to evaluate role of MR spectroscopy in intraaxial braintumors.

Source of data:

Patients with clinical features suggestive of intra cranial space occupying lesion referred for MRI study to the Department of the Radiodiagnosis, Sri Lakshmi Narayan Institute of Medical Sciences wereincludedinstudy.TheMRIwasdoneontheadviceofthereferring doctor and no patient was made to undergo MRI for the sole purpose of thisstudy.

STUDY PERIOD: 18 months.

STUDY DESIGN: Observational Inclusion criteria:

The study includes

• All patients with clinical features suggestive of intra cranial space occupyinglesion.

• All patients with incidentally diagnosed intraaxial brain tumor byCT.

Exclusion criteria:

The study will exclude

• Cases with benign lesions after histopathologyconfirmation.

• Patient having history ofclaustrophobia.

• Patient having history of metallic implants insertion, cardiac pacemakers and metallic foreign body insitu.

• Patient clinically unstable.

SAMPLING AND SAMPLE SIZE: The study is time bound study with sample size of 60 cases.

Data acquisition:

Patients with clinically suspected intra cranial lesion referred for MRI study, underwent the examination after contraindications for MRI were excluded and consent was taken. All the MRI scans in this study were performed using Siemens 1.5 T MAGNETOM ESSENZA MRI scanner.

Mri protocol:

MRI protocol consisted of thefollowing

• Post-contrast T1W-FS axial, Coronal andsagittal.

TR TE NO. OF SLICES

GAP IN

mm MATRIX FOV

T1WI 500 9.7 20-23 1 288 230

T2WI 4000 101 20-23 1 480 230

(3)

FLAIR 9000 105 20-23 1 240 230

DWI 3800 107 20-23 1 128 230

T2 GRE 68000 20 20-23 1 288 230

Single voxel spectroscopy; multi voxel spectroscopy was performed at TE of 135 ms, TR was at 1500 ms. In single voxel studies the voxel is placed on the lesion so that it covers the maximum area of the solid tumoral area. In multivoxel spectroscopy, the voxel was extendedtocoverperilesionalareainselectivecasesofhighgradetumors, avoiding areas of cysts or necrosis and with minimal contaminationfromthe surrounding non-tumoral tissue.

Ascomparedtoamultivoxelmagneticresonancespectroscopy,the operation is quicker and simpler in single voxel magnetic resonance spectroscopy. As compared to a multi voxel magnetic resonance spectroscopy where it is difficult to space over the total area of interest, a limited volume of interest in a single voxel magnetic resonance spectroscopy allows an admirable space. In case of a single voxel magnetic resonance spectroscopy, there is brilliant spectral quality and peak separation with high signal to noise quantification when equated to amultivoxelmagneticresonancespectroscopywhichshowslowersignal to noise and poses problems with quantification. Due to partial volume and chemical shift displacement effects from adjacent tissues, there is spectral contamination in case of a single voxel MRS. The chemical shift aliasing in a multi voxel MRS is due to the bleeding of spectra from the adjacent voxel. The multi voxel magnetic resonance spectroscopy which takes about 6-8 minutes for 2D imaging and 10-15 minutes for 3D imaging is time consuming as compared to single voxel magnetic resonance spectroscopy which consumes about 3-5 minutes pervoxel.

Study definition:

MRspectroscopyisusedasdiagnostictestfordiagnosingintraaxial brain tumors. An increase choline peak at 3.2ppm, myoinositol peakat3.6ppm, lipid peak at 0.9-1.4ppm, lactate peak at 1.3ppm and reduced NAA peak at 2.0ppm, creatinine peak at 3.0,3.9ppm was considered significant for diagnosing brain tumors. We reported brain tumor as high grade if there was increase in choline/creat ratio of more than 2.3, choline/NAA ratio of more than 1.9, reduced NAA/creatinine of less than1.5.Wereportedbraintumorsaslowgradeifcholine/creatinineratio was less than 2.3, this value was used as a threshold value in order to increase the specificity of detecting brain tumors. Astrocytoma tumors were divided as low grade and high grade by using threshold value for myoinositol/creatinineratioof0.82+/-0.25forlowgradetumorsand0.33

+/- 0.16 for high grade tumors.

Statistical analysis:

In Microsoft excel, data was entered and data sheet and analysis was done. Descriptive statistics, frequencies and proportions were calculated and tabulated. Sensitivity, specificity, negative predictive value,positivepredictivevalueanddiagnosticaccuracytotestthevalidity of MR Spectroscopy with respect to histopathological examination were calculated. Fisher exact test was the test of significance for categorical data. p <0.05 was considered as statisticallysignificant.

(4)

20%

20.00%

18.00% 16.67% 16.67%

16.00%

14.00% 13.33% 13.33%

12.00%

10%

10.00%

8.00% 6.67%

6.00%

4.00% 3.33%

2.00%

0.00%

0-10Yr 11-20Yr 21-30 Yr 31-40Yr 41-50 Yr 51-60Yr 61-70 Yr 71-80Yr

3. Results

In the study, it was observed that majority of the patients with intra-axial brain tumors were between 31 to 40 years of age. They constitute 20%of totalstudysample.Theyoungestpatientwas10monthsoldandtheoldest was 78 years old female.

TABLE – 1: AGE DISTRIBUTION OF SAMPLE

AGE (years) NO. OF CASES PERCENTAGE

0-10 10 16.67%

11-20 8 13.33%

21-30 2 3.33%

31-40 12 20%

41-50 10 16.67%

51-60 4 6.67%

61-70 8 13.33%

71-80 6 10%

TOTAL 60 100%

Graph – 1: Bar Diagram showing Age distribution of the subjects

In the study it was observed that majority i.e. 73.33% of the patients with intraaxial brain tumors

were males. It is evident that there is male preponderance in intraaxial braintumors.

(5)

Figure 1 - Anaplastic astrocytoma.

T1axial,sagittalandcoronalshowlargeheterogeneouslyhypointenselesioninrighttemporo- parietalregion.

T2/FLAIR axial show heterogeneously hyperintense lesion with small cystic area in right temporo-parietal region with moderate perilesional edema.

DWI and ADC show no restricted diffusion in the lesion.

(6)

T1 post-contrast axial, sagittal and coronal show few areas of enhancement within the lesion.

In this study it was observed that majority i.e. 73.33% of the patients were having intraaxial brain tumors in supratentorial location. It was observed that most common location for intraaxial brain tumor is supratentorial

.

TABLE – 3 : DISTRIBUTION OF SAMPLE BASED ON LOCATION

NO. OF CASES PERCENTAGE

SUPRATENTORIAL 44 73.33%

INFRATENTORIAL 12 20%

BOTH 4 6.67%

TOTAL 60 100%

Graph-3 : Pie Diagram showing distribution of sample based on location

In this study majority of the patients i.e., 43.33% had hypointense signal on T1.

TABLE – 4 : DISTRIBUTION OF SAMPLE BASED ON SIGNAL CHARACTERISTICS ON T1W

ISOINTENSE 12 20%

(7)

HYPOINTENSE 26 43.33%

HYPERINTENSE - 0%

HETEROINTENSE 22 36.67%

TOTAL 60 100%

Graph-4 : Bar diagram showing distribution of sample based on signal characteristics on T1

In this study majority of the patients i.e., 60% had heterogenous signal on T2.

TABLE – 5 : DISTRIBUTION OF SAMPLE BASED ON SIGNAL CHARACTERISTICS ON T2W

ISOINTENSE 4 6.67%

HYPOINTENSE - -

HYPERINTENSE 20 33.33%

HETEROINTENSE 36 60%

TOTAL 60 100%

(8)

Graph-5 : Bar diagram showing distribution of sample based on signal characteristics on T2

In this study majority of brain tumors i.e., 53.33% had no blooming on gradient and 46.67%

showed blooming, out of which most common cause was bleed i.e., 92.8% and the rest 7.14%

was due to calcification within the tumor.

In this study it was observed that majority of tumors i.e., 80% showed perilesional edema. It is evident that most of the brain tumors present with perilesional edema..

Figure 2 – Hemangioblastoma

(9)

T1 axial and sagittal show hypointense cystic lesion with isointense nodule in posterior fossa.

T2/FLAIR axial show hypointense cystic lesion with hypertense nodule in posterior fossa with no perilesional edema.

DWI and ADC show no restricted diffusion in the lesion.

T1 post-contrast axial and coronal show homogeneous enhancement of mural nodule. MRS shows increased Choline and reduced NAA

TABLE – 7 DISTRIBUTION OF SAMPLE BASED ON PERILESIONAL EDEMA

PRESENT 48 80%

ABSENT 12 20%

TOTAL 60 100%

Graph-8 : Pie diagram showing distribution of sample based on perilesional edema.

In this study it was observed that majority of brain tumors i.e. 63.33% had intense post contrast enhancement. It is evident that most of brain tumors show intense enhancement on post contrast study.

(10)

Graph-12 : Bar diagram showing distribution of sample based on MR SPECTROSCOPY findings of intraaxial brain tumors.

TABLE – 12 : DISTRIBUTION OF CASES ACCORDING TO PATHOLOGY

SERIAL

NO. INTRAAXIAL BRAIN TUMOR NO. OF

CASES PERCENTAGE

1. HIGH GRADE GLIOMA (GRADE3

AND GRADE4 GLIOMA) 28 51.85%

2. LOW GRADE GLIOMA (GRADE1

AND GRADE2 GLIOMA) 4 7.41%

3. OLIGODENDROGLIOMA 4 7.41%

4. GLIOMATOSIS CEREBRI 4 7.41%

5. EPENDYMOMA 4 7.41%

6. MEDULLOBLASTOMA 4 7.41%

7. METASTASIS 2 3.70%

8. CHOROID PLEXUS PAPILLOMA 2 3.70%

9. LYMPHOMA 2 3.70%

TOTAL 54*

*OUT OF 60 CASES, HISTOPATHOLOGY WAS NOT DONE IN 6 CASES.

(11)

Graph – 13 : Distribution of cases according to pathology

TABLE – 13 : DISTRIBUTION OF CASES BASED ON MRI DIAGNOSIS IN CORELATION WITH HISTOPATHOLOGICAL DIAGNOSIS

SL.

NO.

INTRAAXIAL BRAIN TUMOR

HISTOPATHOLOGICAL

DIAGNOSIS MRI

DIAGNOSIS

1. GBM (Grade 4) 22 24

2. High Grade Glioma (Grade 3 ) 4 6

3. Low Grade Glioma (Grade 2 ) 6 4

4. Oligodendroglioma 4 2

5. GliomatosisCerebri 4 4

6. Ependymoma 4 2

7. Medulloblastoma 4 4

8. Metastasis 2 2

9. Choroid Plexus Papilloma 2 2

10. Lymphoma 2 2

11. Neurocytoma - 2

TOTAL 54 54

(12)

Graph-14 :Intraaxial brain tumors

TABLE – 14 : VALIDITY OF MR SPECTROSCOPY WITH HISTOPATHOLOGY AS A DIAGNOSTIC TEST

GlioblastomaMultiforme

Histopathology

GBM Others Total Fisher Exact Test MR

Spectroscopy

GBM 20 4 24

p =

0.00000001

Others 2 28 30

Total 22 32 54

There is significant association between MR Spectroscopy findings and Histopathological findings for Glioblastoma Multiforme

95% Confidence Limit Sensitivity 90.91% 70.84% to 98.88%

Specificity 87.5% 71.01% to 96.49%

Positive Predictive Value 83.33% 66.45% to 92.66%

Negative Predictive Value 93.33% 78.77% to 98.14%

Diagnostic Accuracy 88.89% 77.37% to 95.81%

(13)

Graph-19 : Bar Diagram Showing Diagnostic Accuracy of MR Spectroscopy in diagnosing various Brain Tumors

4. Discussion

In our study, all age group patients were included. Brain neoplasms were most commonly found in 31-40 (n=12) years age group followed by 41-50 (n=10) years age group and 0-10 (n=10) years age group. (Refer Table-1, Graph-1). P A McKinney studied the incidence of brain neoplasms in all age group and found that primary brain neoplasms occur most commonly in 7thdecade. In our study, difference is due to small samplesize.[14-16]Out of 60 patients in our present study, incidence of brain neoplasms was more in males 73.33% (n=44). (Refer Table-2, Graph- 2)In our study of 60 cases, 73.33% (n=44) neoplasms were supratentorial, 20%(n=12) were infratentorial and 6.67%(n=4) were both supra and infratentorial in location. Infratentorial tumors were less common than supratentorial tumors in our study.

In our study, glioma cases were reported as low grade (diffuse infiltrative astrocytoma) or high grade astrocytoma (anaplastic astrocytoma and glioblastoma multiformae), oligodendroglioma,ependymomaand gliomatosiscerebriaccording to the MR characterization of tumors. Both conventional sequences and different parameters of MR spectroscopy was used to optimize for better results.[17,18]Glioma constituted 70% (n=42) out of the total 60 cases in our study. It was the most frequent brain neoplasm found in our study.Out of 42 gliomas cases diagnosed on Magnetic Resonance Imaging, 24 were GBM, 6 were anaplastic astrocytoma, 4 were diffuse infiltrative astrocytoma, 2 case of oligodendroglioma, 4 cases of gliomatosiscerebri and 2 case ofependymoma.[19]

In our study 40 out of 42 (95%) cases of glioma had perilesional edema. The only two case which did not show perilesional edema were ependymoma(n=2). Intense enhancement was showed by all GBM. Moderate enhancement was showed by anaplastic astrocytoma. Minimal enhancement was showed by diffuse infiltrative astrocytoma cases. [20] Intense enhancement was showed by oligodendroglioma, ependymomaand two case of gliomatosiscerebri. Mild enhancement wasshowedbytheothertwocasesofgliomatosiscerebri. Ourfindings are in agreement with study conductedby R Felix, W Schörneretal.[21]

(14)

In our study, all anaplastic astrocytoma and glioblastoma multiformecases and 2 case of oligodendrogliomawere heterogenouslesion withnecrotic and solid component together. The 4 casesof

diffuse infiltrative astrocytoma and 2 cases of ependymoma were solid lesions without any necrotic center. In cases of gliomatosiscerebri, two case were solid and the other two were heterogenous with solid and necroticcomponent.[22] hypointense on T1W and hyperintense on T2W. Lesions were solid to solid and cystic. They showed minimal enhancement. No blooming were observed on T2 GRE sequence.Two cases did not correlate histopathologically, it was diagnosed as anaplastic astrocytoma. Howeverwe got diagnostic accuracy of 96.3% and a significant association between MRS and histopathology findings with p=0.00004743 (p<0.05 being significant). We got 100% sensitivity and 96% specificity. Our findings were similartostudydoneMauricioCastilloaetal.[23]Therewas increased cho/creatratio of 2.03(±0.42), increased cho/NAA ratio of 1.9(±0.34) and reduced NAA/creatpeak at 0.9(±0.33). mI/creatratio was lower at 0.80(±0.25). Both cases showed no choline MRSI all tumors showed increased choline peak, reduced NAA, reduced mIpeak at 3.6 ppm and reduced creat. There was increased cho/creatratio of 6.5(±0.55), increased cho/NAA ratio of 3.5(±0.22) and reduced NAA/creatpeak at 0.8(±0.33). mI/creatratio was lower at 0.15(±0.15). All the cases showed increased choline peak with raised cho/creatratio in perilesional edema probably due to tumoralinfiltration.[24-27]

In our study we evaluated four patients witholigodendroglioma, two of which were misdiagnosed as GBM on MRI. All cases were histopathologicallyproven as anaplastic oligodendroglioma. All the tumors were found in adults in 2ndand 4thdecade. [28]

On conventional MR sequences, lesion appeared heterogeneous to hypointense mass on T1W and heterogeneous to hyperintense on T2W. Two out of four cases showed ill-defined margins, having necrotic and solid component together. Cortical bone thinning was noted in all the cases.

Foci of blooming were observed on T2 GRE sequence due to calcification.

Our study has certain limitations. First being, Perfusion MRI was not done, it may have been useful in preoperative assessment of tumor grade. [28] Second being misclassification of oligodendrogliomawith anaplastic astrocytoma, ependymomawith neurocytomaand misgradingof anaplasticastrocytoma with diffuse infiltrative astrocytoma. This can be because of faulty allocation of volume of interest due to tumor heterogeneity and small sample study in the limitedtime.

5. Conclusions

On the basis of MRS alone, accurate grading of gliomas may be difficult. Combining MRS withconventional andother advancedMR imaging techniques, grading becomes moreprecise.Some features of tumors on conventional MRI (e.g. contrast enhancement, surrounding edema, signal heterogeneity, necrosis, hemorrhage and midline crossing) suggest a high grade. MRS is complementary and helpful for glioma grading. High grade gliomas demonstrate marked elevation o f Cho, decreased NAA and presence of Lactate and Lipid.

Myoinositol is raised in low grade gliomas and reduced with increasing grades of tumors.

Our study also demonstrates that spectroscopic MR measurements in the region surrounding the tumor can be used to demonstrate differences in solitary metastases and high-grade gliomas and also peritumoralinfiltrative nature of certain intraaxialbrain tumor.

Funding: No funding sources

Ethical approval: The study was approved by the Institutional Ethics Committee

(15)

Conflict of interest

The authors declare no conflict of interest.

Acknowledgments

The encouragement and support from Bharath University, Chennai is gratefully acknowledged.

For provided the laboratory facilities to carry out the research work References

[1] Sanghvi DA. Recent advances in imaging of the brain tumors. Indian Journal of Cancer. 2009; 46(issue 2):82-7.

[2] Riyadh N. Al-Okaili, JaroslawKrejza, Sumei Wang et al. Advanced MR Imaging Techniques in the Diagnosis of Intraaxial Brain Tumors in Adults. RadioGraphics 2006;26:S173–S189.

[3] Bulakbasi N. Clinical applications of proton MR spectroscopy in the diagnosis of brain tumors. Spectroscopy 2004; 18(2):143-145.

[4] Bottomley PA. Spatial localization in NMR spectroscopy in vivo. Ann N Y AcadSci1987;508:333–348.

[5] FrahmJ.LocalizedProtonSpectroscopyusingstimulatedechoes.JMagnReson1987;72(3):5 02–508.

[6] BrownTR,KincaidBM,UgurbilK.NMRchemicalshiftimaginginthree dimensions. Proc Natl AcadSci U S A1982;79(11):3523–3526.

[7] Bruhn H, Frahm J, Gyngell ML, et al. Noninvasive differentiation of tumors with use of localized H-1 MR spectroscopy in vivo: initial experienceinpatientswithcerebraltumors.Radiology1989;172(2):541– 548.

[8] Langkowski JH, Wieland J, Bomsdorf H, et al. Pre-operative localized in vivo proton spectroscopy in cerebral tumors at 4.0 Tesla--first results. MagnReson Imaging 1989;7(5):547–555.

[9] Arnold DL, Shoubridge EA, Villemure JG, et al. Proton and phosphorus magnetic resonance spectroscopy of human astrocytomas in vivo. Preliminary observations on tumor grading. NMR Biomed 1990;3(4):184–189.

[10] Hourani R, Horska A, Albayram S, et al. Proton magnetic resonance spectroscopic imaging to differentiate between nonneoplastic lesions and brain tumors in children. J MagnReson Imaging2006;23(2):99–107.

[11] Gill SS, Thomas DG, Van Bruggen N, et al. Proton MR spectroscopy of intracranial tumors: in vivo and in vitro studies. J Comput Assist Tomogr 1990;14(4):497–504.

[12] HoweFA,BartonSJ,CudlipSA,etal.Metabolicprofilesofhumanbrain tumors using quantitative in vivo 1H magnetic resonance spectroscopy. MagnReson Med2003;49(2):223–232.

[13] Preul MC, Leblanc R, Caramanos Z, et al. Magnetic resonance spectroscopy guided brain tumor resection: differentiation between recurrent glioma and radiation change in two diagnostically difficult cases. Can J NeurolSci1998;25(1):13–22.

(16)

[14] Ricci PE, Pitt A, Keller PJ, et al. Effect of voxel position on single-voxel MR spectroscopy findings. AJNR Am J Neuroradiol 2000;21(2): 367– 374.

[15] SenftC,HattingenE,PilatusU,etal.Diagnosticvalueofprotonmagneticresonance

spectroscopy in the noninvasive grading of solid gliomas: comparison of maximum and mean choline values. Neurosurgery 2009;65(5):908–913. discussion 13.

[16] MauricioCastillo,J.KeithSmith,andLesterKwock.CorrelationofMyo- inositol Levels and Grading of Cerebral Astrocytomas AJNR Am J Neuroradio.l October200021:1645–1649.

[17] Preul MC, Caramanos Z, Collins DL, et al. Accurate, noninvasive diagnosis of human brain tumors by using proton magnetic resonance spectroscopy. Nat Med1996;2(3):323–325.

[18] Tate AR, Griffiths JR, Martinez-Perez I, et al. Towards a method for automated classification of 1H MRS spectra from brain tumors. NMR Biomed1998;11(4–5):177–

191.

[19] De Edelenyi FS, Rubin C, Esteve F, et al. A new approach for analyzing proton magnetic resonance spectroscopic images of brain tumors: nosologic images. Nat Med2000;6(11):1287–1289.

[20] Tate AR, Majos C, Moreno A, et al. Automated classification of short echo time in in- vivo 1H brain tumor spectra: a multicenter study. MagnReson Med2003;49(1):29–36.

[21] Fan G, Sun B, Wu Z, et al. In vivo single-voxel proton MR spectroscopy in the differentiation of high-grade gliomas and solitary metastases. ClinRadiol 2004;59(1):77–85.

[22] ChiangIC,KuoYT,LuCY,etal.Distinctionbetweenhigh-gradegliomas and solitary metastases using peritumoral 3-T magnetic resonance spectroscopy, diffusion, and perfusion imagings. Neuroradiology 2004;46(8):619–627.

[23] IshimaruH,MorikawaM,IwanagaS,etal.Differentiationbetweenhigh- grade glioma and metastatic brain tumor using single-voxel proton MR spectroscopy.

EurRadiol2001;11(9):1784–1791.

[24] Law M, Cha S, Knopp EA, et al. High-grade gliomas and solitary metastases:

differentiation by using perfusion and proton spectroscopic MR imaging.

Radiology2002;222(3):715–721.

[25] Spampinato MV, Smith JK, Kwock L, Ewend M, Grimme JD, Camacho DL, Castillo M. Cerebral blood volume measurements and proton MR spectroscopy in grading of oligodendroglial tumors. AJR Am J Roentgenol. 2007Jan;188(1):204-12.

[26] OmuroAM,LeiteCC,MokhtariK,etal.Pitfallsinthediagnosisofbrain tumors. Lancet Neurol2006;5(11):937–948.

[27] Saindane AM, Cha S, Law M, et al. Proton MR spectroscopy of tumefactive demyelinating lesions. AJNR Am J Neuroradiol 2002;23(8):1378–1386.

[28] Remy C, Grand S, Lai ES, et al. 1H MRS of human brain abscesses in vivo and in vitro. MagnReson Med1995;34(4):508–514.