metabolic parameters is important for basic and clinicalresearch (16), but only relative parameters need be imagedin clinicalstudies(14,15,17).

Better methods are needed for selecting appropriatetherapy of brain tumors since treatment of patients withthese tumors is inadequate (18). Tumor blood flow, bloodbrain barrier permeability, and metabolismare importantparameters in selecting appropriate therapy for a patientand in following the effects of therapy on the tumor andnormal brain (16,19,20).

Most clinical PET studies in patients with brain tumorsare performed with FDG. FDG may be obtained using acommercially-available automated synthesis (21); thus,FDG PET is now the most common method used clinically for studying patients with brain tumors.

The combination of FDG and PET has demonstrated clinicalutility in the evaluationof patientswith braintumors.At thetimeof diagnosis,FDGPETprovidesinformationconcerningthe degreeof malignancyand patientprognosis.After therapy,FDGPETis ableto assesspersistenceof tumor,determinedegreeof malignancy,monitorprogression,differentiaterecurrence from necrosis, and assess prognosis. Other studiesusingPETprovideinformationthatmaybeclinicallyuseful.Determination of tumor blood flow and permeability of theblood-brain barrier may help in the selection of appropriatetherapy. Amino acid imaging using 11C-methionineis beingevaluated in patients with brain tumors and provides differentinformationthanFDGimaging.

J NucIMed 1991;32:616—622

PET STUDIES DEMONSTRATINGCLINICAL UTILITY

FDG PET is used clinically in the evaluation of patientswith gliomas (Table 1). The studies are used at the time ofdiagnosis to assess the degree ofmalignancy and to provideinformation related to prognosis and, after therapy, todistinguish between brain damage due to surgery or radiation and tumor persistence, progression, or recurrence.PET centers performing clinical studies have found thatmost patients with suspected or documented primary braintumors are referredfor a FDG PET scan.

PET in the InitialEvaluationof BrainTumors

FDG PET is able to assess the degree of malignancy atthe time of diagnosis since low-grade tumors are lessmetabolicthan high-gradetumors(22). The FDG imagescan be evaluated both visually and quantitatively. Thevisual evaluation determines ifan area ofincreased activityseparate from normal gray matter is present within theconfines of the tumor (Figs. 1 and 2). The quantitativeevaluation determines the absolute metabolic rates foreach tumor. Visual analysis demonstrates increased accumulation of FDG in high-gradetumors but in only 10%of low-grade tumors (22). The absolute metabolic ratesbetween the high-grade and low-grade tumors are alsosignificantly different, but greater overlap exists in thequantitative analysis than in the visual analysis. Di Chiroand Brooks (1 7) have noted the importance of visualinterpretation of PET scans and the limitations of quantitation. Thus, visual analysis of PET scans is used clini

agnetic resonance imaging (MRI) and computedtomography (CT) are excellentanatomic imaging modalitiesfordetectingpatientswith primarybraintumors,butboth ofthese modalities have limitations (1—3).Magneticresonance spectroscopy(MRS) is being evaluated currentlyto determine its role in the evaluation ofbrain tumors (4—10). Positron emission tomography(PET) has several charactenstics that offer significant advantages over other imaging techniques for the evaluation of patients with braintumors. PET provides important research and clinicalinformation in the evaluation of brain tumor metabolism(1 1), blood flow (12), and blood-brain barrier permeability(13). The clinical use of PET has primarily focused on itsuse in studying glucose metabolism using ‘8F-labeledfluorodeoxyglucose (FDG). The clinical indications for PEThave been evaluated by a task force (14) and a workshop(15). Both groups concluded that in patients with braintumors PET is clinically useful for the determination ofthe degree of malignancy and for the differentiation ofrecurrenttumor from necrosis after therapy.

Positron-emitting radionuclides can be readily incorporated into metabolically important substrates, physiologically important compounds, and therapeutic agents(6), allowing many aspects of brain tumors to be characterized. The ability of PET to quantitate physiologic and

Received Nov. 14, 1990; accepted Nov. 14, 1990.For reprints contact: A. Edward Coleman, MD, Box 3949, Duke 1k@iversfty

Medical Center, Durham, NC 27710.

616 The Journal of Nuclear Medicine •Vol. 32 •No. 4 •April 1991

ClinicalApplication of PET for the Evaluationof Brain TumorsR. EdwardColeman,John M. Hoffman,MichaelW. Hanson, H. Dirk Sostman,and S.CliffordSchold

Departments ofRadiology and Internal Medicine (Neurology), Duke University Medical Center, Durham, North Carolina

by on February 5, 2020. For personal use only. jnm.snmjournals.org Downloaded from

.@

\@ )@*@t;'@

TABLE 1ClinicalIndicationsfor PETin the Evaluationof Brain

TumorsDuringInitial Evaluation

Determiningdegreeof malignancyAssessingprognosis

After TherapyAssessingpersistenttumoraftersurgeryGradingdegreeof malignancyAssessingprognosisMonitoringprogressionDifferentiatingrecurrencefromnecrosis

cally for the management of patients with primary braintumors.

Differentresultshavebeenreportedby Tyler et al. (23)who quantified the glucose metabolic rates in the tumorsof 16 untreated patients with suspected high-grade gliomas.Two patients had low-grade tumors (Grade II), and theglucosemetabolic rate in those tumors was not significantly different from the glucose metabolic rates in thetumors of 14 patients with Grade III and IV tumors. Thisstudyis differentfrom the studyof Di Chiro (22), sinceonly patients with suspected high-grade tumors were selected,patientswith previoustreatmentwere excluded,and visual analysis was not performed.

FDG PET can also be used to evaluate patients withmeningiomas to determine the aggressivity and probabilityof recurrence of the tumors. A significant correlation hasbeendemonstratedbetweenthe rate of growth of meningiomas as determined by repeated CT scans and the glucose metabolic rate (24). The glucose utilization rate is asreliableas histologicclassificationand other criteriaforpredicting behavior and recurrence of intracranial meningiomas.

PET in the Post-TherapyEvaluationof BrainTumorsFDG PET is a good indicator of prognosis in patients

with primary brain tumors (22,25). A marked worseningof prognosis is noted as the FDG uptake increases in the

FIGURE1. FDGPETscan(A)demonstratesa hypometaboliclesion in the right frontal area corresponding to the areas ofincreased T2 signal on MRI (B). This abnormality was a lowgradeastrocytoma.

A

617PETin BrainTumors•Colemanet al

A

R

FIGURE2. FDG PET scan(A) demonstratesa focalareahypermetabolismand surrounding hypometabolismin the righttemporal lobe. Contrast-enhancedCT scan reveals contrastenhancementin the right temporallobe within an area ofdecreasedattenuation.Thistumoris a glioblastomamultiforme,a high-gradetumor.

tumor. By comparing the metabolic values of FDG accumulation in the tumor to the opposite normal brain parenchyma, a median metabolic ratio of 1.4 was found for45 patients with high-grade gliomas studied at the NationalInstitutes of Health (NIH). Patients with tumors that hadlow metabolism (ratio < 1.4) had a median survival of 19mo. Patients with tumors that had high metabolism (ratio>1.4) had a median survival of only 5 mo. The PET scanfindings were superior to the histologic grade in predictingprognosis.

Results similar to those from the NIH were obtained ina study at the University of Pennsylvania (25). The patients with hypermetabolic tumors had a median survivalof 7 mo, whereas the patients with normal accumulationor hypometabolic lesions had a median survival of 33 mo.In the patients with high-grade tumors, the FDG PETstudy separated them into groups with a good prognosis(normal or hypometabolic 78% 1-yrsurvival) and a poorprognosis (hypermetabolic 29% 1-yr survival). The FDGPET scan provided an independent assessment of theaggressivenessof the brain tumor.

MRI and CT cannot accurately differentiate persistenttumor from the effects ofsurgery in the earlypostoperativeperiod (1). FDG PET is able to identify persistent tumorafter surgery for brain tumors (26). Brain surgery doesnotresult in increased FDG accumulation at the surgicalsite.Fivepatientswithpartialcomplexseizuresandnohistologic evidence of tumor were studied 6—7days aftertemporal lobectomy and demonstrated no areas of increasedFDG accumulation at the surgicalsite. Seventeenpatientswith primarybraintumors werestudied with FDGPET 1—16days postoperatively to determine if PET couldpredict persistent tumor. Eleven of the 17 patients hadabnormal areas ofFDG accumulation at the surgical margins defined as equal to or greaterthan gray matter. These11 patients had clinical and CT evidence of recurrenttumor 2 mo after surgery.The six patients who had noevidenceof hypermetabolismon the postoperativescan

by on February 5, 2020. For personal use only. jnm.snmjournals.org Downloaded from

A. @;@.B,@,L

;:i::4@F@f-UL

@@ ii

RadiopharmaceuticalParameterStudied150-waterBlood

flow150-oxygenOxygenmetabolism15o@e@

monoxideBloodvolume11C-methionineProteinsynthesis11C-aminoisobutync

acidBlood-brain barrierpermeability‘1C-l-pyruvateLacticacidproduction11C-putrescinePolyamine

metabolism13N-cisplatinChemotherapypharmacoki

netics@Rb-chkrideBlood-brainbarrierpermeability@Ga-EDTABlood-brainbarrier permeability

had no evidenceof recurrent tumor 3—5mo after surgery.Thus, temporal lobectomy and surgical manipulation ofthe brain do not result in increasedFDG accumulation atthe surgical site, and increasedFDG accumulation aftersurgeryaccuratelypredictspersistenttumor.

FDG PET can identify malignant degeneration of lowgradegliomas(27). In a study of 12 patients,all 12demonstrated a focal area of hypermetabolism at the timeof clinical deterioration. Three patients had FDG PETscans before the malignant degeneration, and the regionof tumor was hypometabolic on the initial scan. Theseresultssupportthe useof FDG PET in determiningthebiologic changein the tumors as they undergomalignantdegeneration.

FDG PET is accurate in the differentiation of recurrenttumor from necrosisafter radiotherapyand/or chemotherapy(28,29). CT (2) and MRI (3) are not accurateinthe differentiation of recurrent tumor from necrosis.Radiationnecrosisisdetectedasan areaof hypometabolism(Fig. 3) and recurrent tumor is detected as an area of focalhypermetabolism (Fig. 4) on the FDG PET study. Patientswith necrosissecondaryto intra-arterial chemotherapyalsodemonstratehypometabolism.

CorrelativeStudiesFDG PET provides useful metabolic information about

the brain tumor and normal brain, but the anatomiclocalization of the metabolic information is frequentlydifficulttodeterminefromthePET scanitself,particularlyin patientswho havehad previoussurgerythat distortsnormalanatomy.The effectsof tumor masson the anatomy and the effects of edema and therapy on the metabolismof the normalbrainmustalsobe consideredin theinterpretationof the FDG braintumor study.The necessity of having good anatomic studies for correlating withthe FDG PET scan in patients with brain tumors cannotbe overstressed.

Incorrect diagnoses can be made on the FDG PET scanifa patient hasa seizurecloseto the time of administration

A

r, 4

I@

________L@;@__FIGURE 4. FDG PET scan (A) demonstratesfocalhypermetabolismat the posterior margin of an area ofhypometabolism in the right frontal lobe. The CT scan (B)demonstratescontrast-enhancementat the posteriorborderofanareaof decreasedattenuation.Thispatienthashadprevioussurgeryand radiationtherapyfor an oligodendroglioma.Thispatternis recurrenttumorinconjunctionwithradiationnecrosis.

of the FDG. Seizurefoci resultin areasof hypermetabolism. Even if the patient does not have tonic-clonic movementsbut hasa subclinical seizure,a focal areaof hypermetabolism could be interpreted incorrectly as a highgrade tumor when the patient may have a low-gradetumor. EEG monitoring during the PET study is helpfulif the patient is suspected of having a seizure disorder.

The method ofhistologic grading oftumors is importantfor comparisonwith the PET scanfindings sincedifferenthistologic grading scales have been used for gliomas. Anotherimportantfactorin comparingthe PET scanresultswith histology is the biopsy site. PET is useful in helpingto identify the appropriate site for biopsy. The metabolically active component of a mass is more likely to yield adiagnosis since the metabolically inactive component isfrequentlycysticand/ornecrotic.Thisdifficultyin obtaining representativetissue for histologic analysis and indetermining the correct degree of malignancy on the histologic sections helps explain the additional prognosticinformation in a PET scanevenafter a histologicdiagnosishas been made.

TABLE 2PETBrainTumorStudiesof PotentialClinicalUtility

FIGURE3. FDGPETscan(A)demonstratesa focalareaofhypometabolismin the left frontoparietalregioncorrespondingtotheareaof increasedT2 signalontheMRI(B).Thisabnormalitywasdemonstratedto beradiationnecrosisonbiopsy.

618 TheJournalof NuclearMedicine•Vol.32 •No.4 •April1991

by on February 5, 2020. For personal use only. jnm.snmjournals.org Downloaded from

A @SB;@@!I

OTHER PET STUDIES OF POTENTIALCLINICALUTILITY

While the clinical role of FDG PET in the evaluation ofpatients with brain tumors is well-accepted (14,15), otherstudiesusing PET have provided important informationabout tumors. These studies have been performed in alimited number ofpatients and provide information abouttumorphysiology.

TumorBloodFlowand MetabolismA wide range of blood flow and metabolism in primary

brain tumorshas beendemonstratedin severalstudies(Fig. 5). In a study ofseven patients with giiomas, includingthree patients who had received previous therapy, a widerangeof tumor blood flow and glucosemetabolism wasfound, but the mean values were similar to the contralateral cortex (30). Cerebral oxygen consumption was depressed in the tumor. In another study often patients withgliomas,markedvariability in tumor blood flow wasnoted(31). No relation was found between tumor blood flowand vascularityof the tumor on arteriography.In a studyof 16 patients with untreated gliomas, a wide range ofblood flow and glucosemetabolic rateswas found in thetumors (23). In the Grade II tumors, the glucose metabolicrateswere lower and the blood flows were higher than inthe Grade IIIand IV tumors. However, the significance ofthis difference is uncertain since only two Grade II tumorswere studied.

AminoAcid MetabolismThe evaluation of gliomas with amino acids has been

performed by several investigators. In 11 patients withgliomas using@ ‘C-labeledracemic mixtures of tryptophanand valine, accumulation of these tracers was noted inmost tumors (32). The mechanism of accumulation inthese studies could be relatedto either amino acid metabolism or breakdown of the blood-brain barrier. To furtherunderstandamino acid accumulation in gliomas and nor

mal brain, Bergstrom et al. (33) studied five patients withgliomas using ‘‘C-methyl-L-methionine and branchedchain amino acids. The five patients had localization ofthe amino acid in the tumor greaterthan in normal brain.The stereospecificity of methionine accumulation in thebrain was previously demonstrated by this group. Thestudy documented a 35% reduction of the ‘‘C-methyl-Lmethionine in the brain tumor and normal brain with theinfusion of the branched amino acids. A high sensitivityof ‘‘C-methyl-L-methionineimaging has been reported in33 patientswith brain tumors including 17 patients withgliomas (34). Tumor accumulation was 1.2—3.5timesgreater than normal brain accumulation. The accumulation in high-gradetumors tended to be greaterthan in lowgrade tumors. These initial data are encouraging, andamino acid imaging is being used clinically in some institutions.

Blood-BrainBarrierPermeabllftyPreliminary studies have been performed in patients

with brain tumors to evaluate blood-brain barrierpermeability using 82Rband 68Ga-EDTA (13,35). These agentsor the nonmetabolized amino acid aminoisobutyric acid(AIB) (36) can be used to quantitate blood-brain barrierpermeability. This quantification of permeability may beimportant in the selection of appropriate therapy.

Determination of the Effects of TherapyThe effects of various therapies on gliomas have been

studied.PatientswithanaplasticgiiomaswerestudiedwithFDG PET immediately before and after 60 hr of treatmentwith 0.5 mg/kg of intravenousdexamethasoneat 6-hrintervals (37). Visual interpretationand region of interestanalysis were not significantly changed by the therapy.High-dose steroid therapy did not influence the interpretation of FDG PET scans.

FDG PET scans have been used to evaluate patientsafter radiation therapy and/or chemotherapy (18,38—40).In eight patients with gliomas, the pretherapy FDG PETstudy was compared to the study obtained within 1 mo ofcombined radiotherapy and chemotherapy (38). Six patients had decreased glucose metabolism in the tumorsafter therapy, with regression of the tumors on CT scansand clinical remissions of 1—13mo. One patient demonstrated increased glucose metabolism in the tumor aftertherapy, showing an increase in tumor size by CT and noclinical improvement. The patient with no change inglucose metabolism after therapy had a temporary response with initial regression of the tumor size on CTscan, but tumor progression was noted at 2 mo. Similarresults were published in a later study by the same group(39). The authors concluded that the glucose metabolicrate was a good indicator of therapeutic effectiveness.

Rozental et al. (19) studied the effects of an 8-drugs-in-l-day chemotherapeutic regimen. The tumor-to-contralatera! normal brain ratio increased 20%-100% after therapy,and the ratio decreased to 22% above and 35% below

FIGURE5. Oxygen-i5 bOIUSwatercerebralbloodflowstudiesdemonstrate(A)increasedtumorbloodflowcomparedto normalbrainin deepsubcorticalstructuresincludingthe caudateandstnatumand(B)decreasedtumorbloodflow in a deep,righttemporopanetaltumor. Both tumorsare biopsy-provenhighgradeastrocytomas.

619PETin BrainTumors•Colemanet al

by on February 5, 2020. For personal use only. jnm.snmjournals.org Downloaded from

-

——r@pa—..

-@@;

F—-

.

I.@@-

L1_.@

L

baseline at 28 days. The reason for these changes is uncertam,andtherelationbetweenthesechangesandoutcomewas not studied.

AddftionalPotentialUsesTherapeutic drugs such as cisplatin have been labeled

with nitrogen-13 to assess the pharmacokinetics in braintumors (41). The increased production of lactic acid bycerebraltumors hasbeendemonstratedusing “C-l-pyruvate (42). Malignant brain tumors have an increased metabolic demand for polyamines, which is met by an increased production of putrescine; ‘‘C-putrescinehas beendemonstrated to localize in human giiomas (43). Preiminary data in seven patients with gliomas suggest that theaccumulation of ‘‘C-putrescine relates to the degree ofmalignancy.

COMPARISONOF PET WITh MRI AND MRS

Compared with the extensive experience with PET ingradinggliomas,therehasbeenlittle work with MRI. Onerecentstudy (44) in 36 patientsshowedsignificant differences between low-grade astrocytoma, anaplastic astrocytoma, and glioblastoma groups in mean MM scores. Whensubsequent biopsies were considered, the accuracy of neuropathologic diagnosis was 94% compared with 83% forone observerand 8 1%for a second observer.A systematicstudy including Gd-DTPA enhanced MRI has not beendone. The differentiation of recurrentor persistent tumorfrom therapy-induced brain injury and tumor necrosis isgenerally considered inaccurate by MRI (3). The use ofcontrast-enhancedMRI shows areasofblood-brain barrierbreakdown well, but quantification has not been accomplished to date.

Newer MRI techniques such as diffusion-weighted images have the potential to distinguish between tumor,edema and necrosis, or cyst formation (45,46). A systematic study of thesetechniquesfor grading glioma or evaluating post-therapy effects has not been reported. Thisapproach has the potential to quantify tissue perfusionwithout using tracers or contrast agents. The technicalproblems in quantifying tissue perfusion are formidable,however, and it is unclear whether this method can quantify perfusion using standard clinical MRI systems.

MRSof3'P (Fig.6)and ‘Hisbeingusedin the researchevaluation of human brain tumors (4-10). In a study of13 patients with primary brain tumors using image-guided3'P spectroscopy of a 4 X 4 x 4 cm@volume of interest,metaboliteconcentrationswere reduced 20%-70% inbrain tumors compared with normal brain (4). The pH ofbrain tumors was more alkaline than that ofnormal brain.A study of 43 large brain tumors used image-guided 3'Pspectroscopy of a 5 X 5 x 5 cm@volume of interest (8).Meningiomas demonstrated marked differences, but malignant giiomas showed less distinct changes from normalbrain tissue. Malignant gliomas had a mild reduction in

n1 t:iciA

.k@

p

CNB

FIGURE 6. FDG PET scan (A) demonstratesfocalhypermetabolismintheregionoftheleftthatamusinananaplasticastrocytoma. The T2-weighted MRI (B) and 31Pspectra areshown for the tumor (T), perilesionatnormal brain (PLNB),andcontralateral normal brain (CNB) on MRI. The spectra showdifferencesin tumor pH and in ratios with inorganic phosphateand phosphod@sters.

phosphocreatine and a major reduction in phosphodiesterswitha suggestionofa splitpeak.

Phosphorus-31 MRS and FDG PET were performed in23 patients, including 20 patients with gliomatous tumorsand 3 with meningiomas (6). A large degree of variabilitywas noted in the metabolic rate of glucose in each histologic group. A better separation ofthe more benign gliomasfrom the malignant gliomas was obtained by using theratio of metabolism in the tumor core to the contralateralhemisphere. Low-grade gliomas usually had normal 31Pspectroscopy, and high-grade gliomas had reduced andoften split phosphodiester peaks and alkaline pH. Meningiomas had variable glucose metabolic rates by PET, andthe MRS study showed low phosphocreatine levels, reduced phosphodiesters, and alkaline pH. Early metabolicchanges have been demonstrated after chemotherapy with3'P MRS (47), but correlation with clinical change wasnot seen.

Three patients with primary brain tumors have beenstudiedwith‘Hspectroscopicimages;oneofthesepatientsalso had FDG PET (10). Metabolicmaps of N-acetylaspartate, choline, lactate, and creatine concentrationswere obtained. Lactate was observed in all patients. Choline usually was elevated in tumors, and N-acetyl aspartateusually was reduced. Regional variations in tumors werenotable. In the patient with a PET scan, the abnormallyincreased FDG accumulation correlated with increased

620 The Journal of Nuclear Medicine •Vol. 32 •No. 4 •April 1991

- 7 PUll

by on February 5, 2020. For personal use only. jnm.snmjournals.org Downloaded from

lactate concentration. Since lactate is the end product ofglycolysis, this correlation is expected (9).

Other studies using ‘HMRS have demonstrated specificabnormalities between different tumor types (48) anddifferent grades of tumor (49), as well as heterogeneousmetabolism within subregions ofbrain tumors (50). Whilepreliminary, these observations suggest a potential role intumor typing that might be similar to that of FDG PET.

Finally, early results have appeared in which MR imaging of endogenous 23Na in human tumors (51) andadministered ‘9F-deoxyglucosein rat tumors (52) wasperformed. These are to date simply feasibility studies.

Standard ‘HMRI produces excellent anatomic images,but is unlikely to replacePET for tumor grading,analysisof post-therapy changes, or physiologic (e.g., blood-brainbarrier permeability) quantification. Diffusion-weightedimages have interesting potential for differential diagnosis,but experienceto datewith brain tumors is too limited tosuggest the ultimate clinical utility of this method. MRS,although it yields metabolic data, is quite different fromPET in that it studiessteady-statemetabolismratherthantracer kinetics. Proton MRS potentially can produce spatial resolution similar to that of PET and the evaluationof lactatemayyieldinformationsimilarto FDGuptake.Phosphorus MRS is always likely to have inferior spatialresolution compared with PET. The information providedwith respect to pH, energy status, and phospholipid metabolism is distinct from that obtainable with FDG PET.Experience with both ‘Hand 3'P MRS is still too limitedto indicate clearly the clinical role of MRS by itself or inconjunction with PET.

CONCLUSION

PET studiesprovideunique informationconcerningbrain tumors. The ability to study tumor blood flow,blood-brainbarrier permeability,oxygen,glucose,andamino acid metabolism provides a better understandingof the tumors and the effects of therapy on tumors. Thesephysiologic and metabolic studies are being used to improve the design and selection of therapeutic measuressince currenttherapy is effective in a minority of patients.Parameters that can be monitored by PET may demonstratean important factor in the unusual therapeutic sensitivity of tumors in some patients.

On the basisof currentknowledge,FDG PET studiesare clinically helpful in the treatment of patients withprimarybrain tumors. The ability to determine the degreeof malignancy(22,25,27)hasimportanttherapeuticimplications. Patients with low-grade tumors are followedclinically without therapy, whereas patients with highgrade tumors receive radiotherapy and/or chemotherapy.The appropriate time to begin therapy in patients withinitially low-grade tumors is not easily determined, andthe data from Francavillaet al. (27) demonstrate that PETscanning can be used to determine malignant degenera

tion. PET scanning is more accurate than CT or MRI inthe detection of persistent tumor in the postoperativeperiod (37) and in the differentiation of recurrent tumorfrom necrosis after therapy (28,29). The clinical role ofPET scanning in primary brain tumors will increase asmore studies are performed validating the initial results,providing new data on selecting appropriatetherapy, andfollowing the results of the therapy.

REFERENCES

1.LaohaprasitV, SilbergeldDL, OjemannGA, EskridgeJM, WinnHR.Postoperative CT contrast enhancement following lobectomy for epilepsy.JNeurosurg 1990;73:392—395.

2. MartinsAN,JohnstonJH,HenryJM,HenryJM,StoffelTi, Di ChiroG.Delayednecrosisofthe brain.I Neurosurg1977;47:336—345.

3. DoomsGC,HechtS, Brant-ZawadskiM, BerthiaumeY, NormanD,Newton TH. Brain radiation lesions MR imaging. Radiology 1986;158:149—155.

4. HubeschB, Sappey-MarinierD, RothK, MeyerhoffDi, MatsonGB,Weiner MW. P-31 MR spectroscopy of normal human brain and braintumors. Radiology 1990;174:401—409.

5. ThomsenC,JensenKE,AchtenE,Henriksen0. In vivomagneticresonance imaging and “Pspectroscopy oflarge human brain tumors at 1.5tesla. Ada RadioL 198829:77—82.

6. Heiss WD, Heindel W, Herholz K, Rudolfi, Bunke i, ieske J, Friedmann0. Positron emission tomography offluo@ne-18..deoxyglucoseand imageguided phosphorus-31 magnetic resonance spectroscopy in brain tumors@JNuc!Med 1990;31:302—310.

7. Segebarth CM, Baleriaux DF, Arnold DL, Luyten PR, den Hollander iA.MRimage-guidedP-31MRspectroscopyin theevaluationofbraintumortreatment.Radiology1987;165:215—219.

8. HeindelW,BunkeJ,GlatheS.SteinbrichW,MollevangerL.Combined‘H-MRimaging and localized “Pspectroscopy of intracranial tumors in43patients.J CompulAssistTomogr1988;12:907—916.

9. LearJL. Glycolysis:link betweenPET andprotonMR spectroscopicstudies ofthe brain. Radiology 1990;174:328—330.

10.LuytenPR,MarienAJH,HeindelW,ci al.Metabolicimagingofpatientswith intracranial tumors: H-i MR spectroscopic imaging and PET. Radiology i990;176:791—799.

11. BeaneyRP.Positronemissiontomographyinthestudyofhumantumors.SeminNuciMed 1984;14:324—341.

12.MineuraK, YashudaT, KowadaM, etal.PETevaluationof histologicmalignancy in gliomas using oxygen 15- and fluorine-18-fluorodeoxyglu.cose. Neurologica/Res 1986;8:164—168.

13. Doyle WK, Budinger iT, VaIk PE, Levin VA, Gutin PH. Differentiationof cerebral radiation necrosis from tumor recurrence by [‘1F]FDGand82Rb positron emission tomography. J Comput Assist Tomogr1987;i1:563-570.

14. KuhI DE, Wagner HN, A1aVIA, et al. Positron emission tomography:clinical status in the United States in 1987. 1 NucI Med i98829:l 136—1143.

15. Al-Aish M, Coleman RE, Larson SM, et al. Advances in clinical imagingusing positron emission tomography. National Cancer Institute WorkshopStatement. Arch Intern Med 1990;150:735—739.

16. Fowier iS, Hoffman EJ, Larson SM, et al. Positron emission tomographyin oncology. JAMA 1988;259:2126—2131.

17. Di Chiro G, Brooks RA. PET quantitation: blessing and curse. JNucl Med1988;29:1603—1604.

18.GreenSB, Dyar DP, WaiderMD, et al Compar,sonof carmustine,procarbazine, and high dose methylprednisolone as additions to surgeryand radiotherapy for the treatment of malignant glioma. Cancer Treat Rep1983;67:121—132.

19.RozentalJM, LevinRL, NicklesRi, DobkiniA. Glucoseuptakebygliomas after treatment. A positron emission tomographic study. ArchNeurol 1989;46:1302—1307.

20. Phillips PC, Dhawan V, Strother SC, et al. Reduced cerebral glucosemetabolism and increased brain capillary permeability following high-dosemethotrexate chemotherapy: a positron emission tomographic study. AnnNeurol1987;21:59—63.

21.FowlerJS,HoffmanEl,LarsonSMetal.Cyclotronsandradiopharmaceuticais in positron emission tomography. JAMA 1988;259: 1854—1860.

PETinBrainTumors•Colemanetal 621

by on February 5, 2020. For personal use only. jnm.snmjournals.org Downloaded from

22. Di ChiroG. Positronemissiontomographyusing[‘8F]fluorodeoxyglucosein brain tumors: a powerful diagnostic and prognostic tool. Invest Radio!1986;22:360—371.

23. Tyler IL, Diksic M, Villemure 1G. EvansAC, Yamamoto YL, Feindel W.Metabolic and hemodynamic evaluation of gliomas using positron emission tomography. JNuc!Med 1987;28:l 123—1133.

24. Di Chiro G, Hatazawa I, Katz DA, Rizzoli HV, De Michele Di. Glucoseutilization by intracranial memngiomas as an index of tumor aggressivityand probability ofrecurrence: a PET study. Radio!ogy 1987;164:521—526.

25. Alavi lB. Alavi A, Chawluk I, et al. Positron emission tomography inpatients with glioma: a predictor ofprognosis. Cancer 1988;62:1074—1078.

26. Hanson MW, Hoffman JM, Glantz Mi, et al. FDG PET in the earlypostoperative evaluation of patients with brain tumor [Abstract]. J Nuc!Med 1990;31:799.

27. FrancavillaTL, MiletichRS,Di ChiroG, etal.Positronemissiontomography in the detection of malignant degeneration of low-grade gliomas. JNeurosurg1989;24:l—5.

28. Di Chiro G, Oldfield E, Wright DC, et al. Cerebral necrosis after radiotherapy and/or intra-aflerial chemotherapy for brain tumors: PET andneuropathologic studies. AJNR l987;8:1083—l091.

29. Valk PE, Budinger TF, Levin VA, Silver P. Cutin PH, Doyle WK. PET ofmalignant cerebral tumors after interstitial brachytherapy: demonstrationof metabolic activity and correlation with clinical outcome. J Neurosurg1988;69:830—838.

30. Rhodes CG, Wise RiS, Gibb JM, et al. In vivo disturbance ofthe oxidativemetabolism of glucose in human cerebral gliomas. Ann Neuro!1983:14:614—626.

31. Lammertsma AA, Wise RJS, Cox T@S,Thomas DGT, Jones T. Measurement of blood flow, oxygen utilization, oxygen extraction ratio, and fractional blood volume in human brain tumours and surrounding oedematoustissues. BrJRadio! 1985;58:725—734.

32. HubnerKF, PurvisiT, MahaleySM, et al. Braintumorimagingbypositron emission computed tomography using “C-labeledamino acids. JCompul Assist Tomogr 1982;6:544—550.

33.BergstromM, EricsonK, HagenfeldtL, etal.PETstudyof methiomneaccumulation in gliomas and normal brain tissue; competition withbranched chain amino acids. I CompuzAssist Tomogr 1987;l 1:208—213.

34.O'TuamaLA,LaFranceND,DannalsRF,etal.Quantitativeimagingofneutral amino acid transport by human brain tumors [Abstract]. I CerebB!oodF!owMetab 1987;7(suppl l):5517.

35. Kessler RM, Globe IC, Bird JH, Ctal. Measurement ofblood-brain barrierpermeability with positron emission tomographyand [@GaJEDTA. JCerebB!ood F!ow Metab 1984;4:323—328.

36. Washburn LC, Blair LD, Byrd BL, Sunn iT. Comparison of@Ga-EDTA,(1.1 1 C) aipha-aminoisobutyric acid, and [@“Tc]sodiumpertechnetate inan experimental blood-brain barrier lesion. In: J Nuc! Med Bio!1984:12:267—269.

37. Glantz Mi, Hoffman JM, Coleman RE, et al. The role ofF-18-FDG PETimaging in predictingearly recurrence ofprimary brain tumors. Ann Neuro!l990:in press.

38. Mineura K, Yashuda T, Kowada M, et al. Positron emission tomographicevaluation of radiochemotherapeutic effect of regional cerebral hemocirculation and metabolism in patients with gliomas. J Neuro Onco!1987;5:277—285.

39. OgawaT, UemuraK, ShishidoF, et a!.Changesofcerebralbloodflow,and oxygen and glucose metabolism following radiochemotherapy ofgliomas: a PET study. J Compus Assist Tomogr 1988;12:290—297.

40. Phillips PC, Dhawan V. Strother SC, et al. Reduced cerebral glucosemetabolism and increased brain capillary permeability following high-dosemethotrexate chemotherapy: a positron emission tomographic study. AnnNeuro! 1987;21:59—63.

41.GinosJZ,CooperAJL,DhawanV. et al. [‘3N)cisplatinPETto assesspharmacokinetics of intra-aflerial versus intravenous chemotherapy formalignant brain tumors. JNuc!Med 198728:l844-1852.

42.TsukiyamaT, HaraT, ho M, KidoG,TsubokawaT. Preferentialaccumulation ‘‘Cin human brain tumors after intravenous injection of “C-l-pyruvate. Eur JNuc!Med 1986;12:244—248.

43.HiesigerE,FowleriS,Wo1fAP,etal.SerialPETstudiesofhumancerebralmalignancy with [l-―CJputrescine and Il-C] 2-deoxy-D-glucose. J Nuc!Med 1987;28:125l—126l.

44. Dean B!, Drayer BP, Bird CR, Flom RA, Hodak IA, Coons SW, CareyRE. Gliomas: classification with MR imaging. Radio!ogy 1990;174:411-415.

45. Le Bihan D, Breton E, Lattemand D, Rubin ML, Vignand I, Laval-JeantetM. Separationofdiffusionandperfusionin intravoxelincoherentmotionimaging. Radio!ogy 1988;l68:497—505.

46. Merholdt K-D, Bruhn K, Frahm J, Gyngell ML, Hanicke W, Deimling M.MRI of “diffusion―in humanbrain:newresultsusinga modifiedCEFAST sequence. Magn Reson Med 1989;9:423—429.

47. Arnold DL, Shoubridge EA, Emrich I, Feindel W, Villemure J-G. Earlymetabolic changes following chemotherapy of human gliomas in vivodemonstrated by phosphorus magnetic resonance spectroscopy. Invest Radio! 1989,24:958—961.

48.BruhnH, FrahmH, GyngellML, et al.Noninvasivedifferentiationoftumors with use oflocalized H-l MR spectroscopy in vivo: initial expenence in patients with cerebral tumors. Radio!ogy l989;1972:54l—548.

49.PeelingJ,SutherlandGR,GirvinI. 1-HNMRanalysisofmetabolitelevelsin human epileptic brain tissue and brain tumors. Proceedings ofthe 8thAnnual Meeting, Society ofMagnetic Resonance in Medicine, Amsterdam,1989:296.

50. Segebarth CM, Baleriux DF, Luyten PR, Den Hollander JA. Detection ofmetabolic heterogeneity of human intracranial tumors in vivo by 1-HNMR spectroscopic imaging. Magn Reson Med 1990;13:62-76.

51. Turski PA, Houston LW, Perman WA, et al. Experimental and humanbrain neoplasms: detection with in vivo sodium MR imaging. Radio!ogy1987;163:245—249.

52. NadadaT, KweeIC. Heterogeneity ofregional cerebral glucose metabolismdemonstrated in situ in rat brain by “FNMR rotating frame zeugmatography:one-dimensional chemical shift imaging ofnormal and gliosarcomaimplanted brain. Magn Reson Imaging l987;5:259—266.

622 The Journal of Nuclear Medicine •Vol. 32 •No. 4 •April 1991

by on February 5, 2020. For personal use only. jnm.snmjournals.org Downloaded from

1991;32:616-622.J Nucl Med.   R. Edward Coleman, John M. Hoffman, Michael W. Hanson, H. Dirk Sostman and S. Clifford Schold  Clinical Application of PET for the Evaluation of Brain Tumors

http://jnm.snmjournals.org/content/32/4/616This article and updated information are available at:

  http://jnm.snmjournals.org/site/subscriptions/online.xhtml

Information about subscriptions to JNM can be found at:  

http://jnm.snmjournals.org/site/misc/permission.xhtmlInformation about reproducing figures, tables, or other portions of this article can be found online at:

(Print ISSN: 0161-5505, Online ISSN: 2159-662X)1850 Samuel Morse Drive, Reston, VA 20190.SNMMI | Society of Nuclear Medicine and Molecular Imaging

is published monthly.The Journal of Nuclear Medicine

© Copyright 1991 SNMMI; all rights reserved.

by on February 5, 2020. For personal use only. jnm.snmjournals.org Downloaded from