CONTINUING EDUCATION

PET, CT, and MRJ in the Evaluation of Neuropsychiatric Disorders: Current Applications

Paul R. Jolles*, Peter R. Chapmant, and Abass Alavi Division ofNuclear Medicine, Department ofRadiology, Hospital ofthe University of Pennsylvania, Philadelphia, Pennsylvania

Positron emission tomography (PET) is emerging as a very useful clinical tool and is adding a greatdealto ourunderstandingofthepathophysiologyofcentralnervoussystem(CNS) disorders.Althoughcomputedtomography(CT)andmagneticresonanceimaging(MRI)have hada dramaticimpactonpatientmanagement,thereisoftenanimportantassociated functionabnormalitywhichisbestassessedbyPET.Innormalagingandindementia,theCT andMRIbrainchangesofatrophyandwhitematterabnormalitiesarefrequentlynonspecific. PEThasbeenmorediagnostic,showingcharacteristicregionalmetabolicabnormalities. Evaluation of brain tumors such as astrocytomas with PET has demonstrated better correlation with histologic grade compared to CT. Unlike CT or MRI, PET can help to distinguishradiationnecrosisfromrecurrenttumor,andcandifferentiatetheextentof metabolically active tumor from surrounding edema. PET is useful in evaluating stroke patients,providingbetterprognosticinformationanddemonstratingabnormalitiessoonerthan CT.Inepilepsy,PETappearsto besuperiorto MRIinlocalizingseizurefociinpatientswith partialseizures.In headtraumapatients,metabolicpatternsare beingdeSCribedwhichwill likelyhaveaneffecton patientmanagement.Theuseof PETin schizophreniahasyielded very interesting results, with common patterns of metabolic abnormalities being demonstrated. CT and MRI in these patients have not been very useful. PET has also shown promise in movement disorders such as Huntington's disease. It is now clear that PET is alreadydinicallyusefulandcanprovidevaluableinformationunobtainablebyCTandMRI.As newradioligandsaredeveloped,PETiscertainto assumeanevenmoreimportantroleinthe future.

J NuciMed30:1589—1606,1989

he past two decades have been witness to remark the brain have a functional component, sometimes in able progress in our understanding of brain structure the absence of an anatomic lesion. Today, brain func and function, with advances in anatomic and functional tion is best evaluated with positron emission tomogra brain imaging playing a major role. Computed tomog phy (PET). Although lacking the resolution of CT or raphy (CT) and, more recently, magnetic resonance MRI, the advantage of PET over the anatomic imaging imaging (MRI) have provided exceptional anatomic modalities is that biologically important atoms such as detail. However, these modalities are unable to assess nitrogen and carbon can be incorporated into novel the global or regional functional status of the brain. It compounds which allow mapping of regional brain is well known that virtually all abnormalities affecting function and metabolism (Fig. 1). The capabilities of feredby PET enable researchersto measure and image ReceivedMay 19, 1989;revision accepted May 31, 1989. regional blood flow, oxygen and glucose metabolism, Forreprintscontact:AbassAlavi,MD, Div. of NuclearMedi protein synthesis, and neurotransmitter/neuroreceptor cine, Dept. of Radiology, Hospital of the University of Pennsyl systems. With PET now moving from the research vania, Philadelphia,PA 19104. laboratoryinto the clinical nuclear medicine arena, we *pre@ntaddress: Department of Radiology,University of Ar have indeed entered a new era in brain imaging. This kansas for Medical Sciences, 4301 West Markham St., Little Rock, Arkansas72205. paper will deal with the application of PET, CT, and t Present address: Nuclear Medicine & Ultrasound Associates, MRI as currently used in the clinical evaluation of P.O.Box6, Westmead.NSW2145,Australia. common brain abnormalities.

Volume30 •Number10 •October1989 1589 to be a common underlying condition (10,11). How ever, the same changes have been found in subjects without hypertension or other identifiable vascularrisk 0 factors (12), and it is possible that several etiologies may be involved. Our preliminary data suggest that deep white matter lesions do not have a significant effect on regional or global metabolic rates. The findings seen on PET in normal aging are some what controversial. Several investigators have reported that there is probably no significant decrease in whole brain glucose metabolism using fiuonne-l8 (‘8F)fluo rodeoxyglucose (FDG) (13—15),although there are other reports showing a decline (16,1 7). However, a number of investigatorshave describeddiminished re gional glucose metabolism in healthy elderly subjects 1―SSin the temporal, parietal,somatosensory, and especially the frontal regions (1 7-21). Relative preservation of posterior metabolism (particularlyin the calcarine cor tex) has also been reported(20,21). A summary of our regional metabolic data in normal aging is presented in Table 1. Whetheror not there is a definite decreasein cerebral a 0 a * blood flow (CBF) or cerebral oxygen metabolic rate FIGURE 1 (CMRO2)as measuredby PET has not been established. RepresentativetransaxialimagesfromanFDG-PETstudy Some workers have described no change with normal ofa young,normalsubjectareshownfromthelevelofthe high convexity (upper left) to the cerebellum (lower right). aging using the nitrous oxide CBF technique (16,22); Notethatthecorticalandsubcorticalstructuresarewell others have reported diminished CBF and/or CMRO2 delineated,andappearrelativelysymmetricbetweenhem (23-26). Although controversial, these findings suggest ispheres. that there may be regional abnormalities in cerebral metabolism in normal aging using PET. There has been much interest in evaluating the AGING AND DEMENTIA changes which occur in the brain in the dementing disorders. The most common cause of dementia is Evaluation of the brain in aging and dementia has Alzheimer's disease (AD), which accounts for at least been extensive, with functional imaging (PET) appear half of the cases. Other causes include multiple infarc ing to have greater sensitivity in detecting early regional tions (multi-infarctdementia) (MID), AD, and MID in changes than CT or MRI. In the normal elderly popu combination, and other less common etiologies such as lation, pathologic studies have reportedregionalneuron normal pressure hydrocephalus, mass lesions, infec loss in the frontal, parietal and temporal lobes (1—3). tions, AIDS, and metabolic/nutritional disorders (27- Typical changes seen on anatomic imaging in these 30). individuals include generalizedventricularenlargement The CT and MRI findings in AD include ventricular and cortical sulcal atrophy; of these findings, sulcal and cortical sulcal enlargement, which are more pro atrophy and the width of the third ventricle appear to be better correlated with aging using CT (4—6).Both TABLE 1 MRI and CT demonstrate the ventricularsystem quite Significant Regional Hypometabolism: Individual ROll well, with MRI being superiorto CT in delineating the Frontal ______Temporal cortical sulci due to lack of beam hardening artifacts. Mtenor corpuscallosum Middletemporalgyms White matter abnormalities are also common in the Cingulate gyms healthy elderly subject, manifesting as periventricular Frontaleye fields _____Parietal Frontalpole Superiorpanetal and/or focal deep white matter hyperintensities on Middlefrontalgyms lObUle MRI, or hypodensities on CT. MRI is the more sensitive Sensonmotor in detecting these lesions (7). Although there is general Primarysensorycortex agreement that the occurrence of white matter abnor malities increases with age (8,9), the actual underlying . 59 ROls in each hemisphere (1 1 8 comparisons by two-tailed pathologic mechanism is controversial. Most reports t-test[youngvs.oldercontrols]). conclude that these abnormalities represent areas of ROlssignificantatp = 0.001. hypoperfusion or infarction, with hypertension found [Reprintedwith permissionRef.(15)].

1590 Jolles,Chapman,Alavietal TheJournalofNudearMedicine nounced than those seen in normal aging (31-33). However, no findings unique to AD have yet been However, these findings are not specific for AD and described;our own preliminary data showed no signif may be found in many other types of nondementing icant differences between AD patients and controls brain disorders. Focal temporal lobe atrophy with en (37). largement of the surrounding cerebrospinal fluid (CSF) The use of FDG PET in dementia shows promise, as space has been describedin AD patients using CT (34); several types ofdementing disorders have been reported thus far, this appears to be the only anatomic imaging to demonstrate characteristic metabolic patterns. A findingreportedwhich is saidto distinguishAD patients number of studies (21,38,39) have described dimin from normal elderlysubjectswith an 80—90%accuracy. ished regional glucose metabolism in AD patients in White matter disease has also been described in AD the parietal and temporal lobes; the frontal lobes may patients. At present, there is much debate in the litera also demonstrate hypometabolic changes, particularly ture as to whether AD patients have a greater degree of in severe cases (Fig. 2). Relatively preserved glucose white matter abnormality compared to normal elderly metabolism has been noted in the primary visual and controls. Overall,the data suggestthat AD patients and sensonmotor cortices, as well as the cerebellum (Fig. normal controls cannot be reliablydistinguished by the 3). The visual cortex glucose metabolism especially number of focal deep white matter lesions. However, remains quite stable during different stages of AD. the presence of certain patterns of periventricular hy Because of this observation, we usually use the visual perintensity (PVH) on MM might aid in making this cortex metabolic rates for normalization rather than distinction. Fazekas et al. (12) have described a halo of that of the cerebellum as described in the literature. PVH which was seen in six of 12 AD patients, but not These data are presented in Table 2. in controls or MID patients. Normal elderly subjects Hemispheric asymmetry may also be present, with had caps, pencil-thin, or absent PVH. Additional stud more severe hypometabolism on the left compared to ies may help to clarify the diagnostic significance of the right;this may explain the languagedifficulties seen these findings. Patients in this study with MID tended in certain advanced AD patients. Global and regional to have extensive focal and confluent white matter CMRO2 was reported to be decreased in AD patients, abnormalities, associated with “classical―infarcts. especially in the temporal and parietal regions; substan Phosphorus-3l magnetic resonance spectroscopy of tially diminished frontal oxygen metabolism has been the brainhasbeen used to evaluateAD patients(35,36). noted in severe cases. An associated decrease in CBF

@ @‘.

@ . @4, M@893P03 It@893 P03 IPI@

P 20'S 14@ 0 P 22/S 1'G 0

B I('e@: @:1';:, @(r@3 hi3 ),@ @ , :@ • •, FIGURE 2 A: The MRI imageof a moderatelyearlyAlzheimer'sdiseasepatientshowssignificantbilateralcorticalatrophyand ventricularenlargement.Therearealsoa few foci of deepwhite matterhypenntensity.B: The FDG-PETimageshows profound bilateral parieto-occipital hypometabolism with additional abnormality in the right temporal lobe. The remainder ofthecorticalandsubcorticalstructuresarerelativelyintact.

Volume30 •Number10 •October1989 1591 B @,‘,@*t'@p@I•r@ •@WI

“@tP@? ‘°‘

FIGURE 3 A: MRI image of a patient with advanced ftJzheimer's disease shows mild cortical atrophy and moderate ventricular enlargement,with a few foci of deep white matter hypetintensity.Thesefindingsare nonspecificand are frequently seen in elderly individuals. B: Sequential FDG-PET images show marked bilateral parieto-occipital hypometabolism and moderate bilateral fronto-temporal hypometabolism. There is preservation of glucose metabolism in the sensorimotor andcalcanne(primaryvisual)cortices,andinthesubcorticalnuclei,frequentlynotedinAlzheimer'sdisease.

has also been reported (41,42). AD patients with a exhibit more profound glucose hypometabolism in the progressive decrease in glucose metabolism have been frontalregionsthan in the parietaland temporal regions shown to display corresponding cognitive loss. In ad (45). The PET findings seen in demented patients with dition, regions of local hypometabolism were reported Huntington's disease and Parkinson's disease will be to correspond to areas of cortical neuronal loss in AD discussed in the section on movement disorders. (43). It is clear that PET has the potential to characterize With PET, other types of dementia may be able to and further our understanding of the various types of be distinguished from AD. MID patients tend to dem dementia. A number ofcontroversies exist, as discussed onstrate scattered focal regions of diminished glucose above. In addition to these, a Standardizedmethod for metabolism (44), ratherthan the predominantly dimin quantification of metabolic data has yet to be achieved; ished parietal and temporal metabolism seen in AD. also, the need for atrophy correction of metabolic rates This finding, along with the characteristic appearance is being recognized (46). As these problems are ad of multiple lesions on MRI (especially in the white dressed and solved, and selective brain stimulation/ matter), establishes the diagnosis in most patients. In activation studies are utilized, great progress in this area dividuals with Pick's disease have been reported to will be made.

TABLE 2 Normalization of Quantitative Metabolic Data Structuresused to negativeCalcarinenormalizeSensitivitySpecificityTest accuracyPredictive positivevalues cortex89%81%84%72%93%Cerebellum1 00%57%72%56%‘100%Caudate nucleus5%78%52%12%60%Sensoimotor %80%79%80%Wholecortex59%91 brain35%79%63%48%69%[Reprinted

with permiSSiOnRef. (40)].

1592 Jolles, Chapman, Alavi et al The Journal of Nuclear Medicine BRAIN TUMORS in the evaluation of brain tumors is not a new applica tion; Wrenn et at. described a scintillation counter for Since the mid 1970s, rapid advances in high resolu this purpose in 195 1 (55). Since the advent of modern tion anatomic imaging have favored the use of CT and PET scanners,largenumbers ofbrain tumors have been MRI (in recent years) for brain tumor characterization. evaluated, mainly with FDG; ‘‘CL-methionine (reflect In general, both CT and MRI provide excellent deline ing neutral amino acid transport) (56) and “Cputres ation ofbrain tumors. However, MM hasbeen reported cine (57) have also been employed. In most cases, high to be more sensitive than CT due to better resolution, grade tumors appear hypermetabolic, while low grade lack of beam hardening or streak artifacts, and multi tumors are hypometabolic (Figs. 4 and 5). Associated plane imaging capability. In particular,MRI is superior edema, cystic changes, and necrosis also manifest as in evaluatingthe cerebellumand brainstem,wherethere hypometabolic regions. Overall, the data show a striking may be severe beam hardening artifacts on CT. CT, correlation between PET findings and histologic grade, however, is much superior to MM in detecting small with PET being more accurate in predicting degree of foci of calcification (47-50), which may be seen in tumor malignancy than CT. PET was also reported to lesionssuch as meningiomasand oligodendrogliomas. be of better prognostic value than CT in predicting Beforethe introduction ofparamagnetic contrastagents such as gadolinium-diethylenetriaminepentaacetic acid ([Gd]DTPA), CT with iodinated contrast was reported to differentiate tumor from surrounding edema much better than MRI. However, this may no longer be the case except for lesions such as lower grade gliomas, which show less reliable enhancement (51) but gener ally do not cause significant edema. The following discussion will center upon astrocyto mas, which are common brain tumors. Astrocytomas are generallyclassifiedinto four grades:Grades I and II (low grade)and Grades IIIand IV (high grade). By and large,the tumor appearanceon CT or MRJcorresponds with the histologic grade. On CT, low grade tumors typically present as hypodense lesions with little if any mass effect or surrounding edema. Contrast enhance ment is infrequent and, if present, is usually inhomo geneous. The tumor borders may be well defined, at though local infiltration may result in a less distinct margin. Calcification (usually central) may be present. These lesions appearas high signal intensity regions on T2-weighted MR images, and low signal intensity on Tl-weighted images. Again, as on CT, little mass effect or edema is evident, and enhancement with [Gd]DTPA is variable.Punctate calcificationswillprobablynot be detected on MRI, although large calcifications may be 0 delineated (52-53). High gradetumors typically appear .,.1@r@. asirregular, inhomogeneous density lesions on CT with associatedmass effect, surroundingedema, and contrast t ‘%J enhancement. On MRI, these lesions demonstrate high bdP@: P_i (or inhomogeneous) signal intensity on T2-weighted .. images with associated edema, mass effect, and contrast B enhancement with [Gd]DTPA(50). Although CT and MRI usually provide excellent FIGURE 4 delineation of brain tumors when they are originally A: The selectedMR imageof a leftthalamiclymphoma diagnosed,a significantclinical dilemma arises in pa demonstrates tumor enhancement with gadolinium-DTPA. tients who have received radiation therapy. Unfortu Asmalllowsignalintensityregionisseenwfthinthemass, nately, neither imaging modality is able to distinguish representinghemorrhage.Thereis only mildmasseffect. B: The FDG-PETimagesshowa focalhypermetabolic recurrent tumor from radiation necrosis (54). This is focus in the left thalamus, consistent with a very active an areawhere PET can provide a valuablecontribution. tumor.Notethatthereis nosurroundinghypermetabolic The use of positron emitting radiopharmaceuticals zone, which suggests the absence of significant edema.

Volume30 •Number10 •October1989 1593 0 @ TLWQIP03

P2@'S I'GO

0@

B C

FIGURE 5 A: CT scanin a patientwitha low-gradegliomashowsan irregularhypodenseregionin the lighttemporo-panetal region,involvingboththecortexanddeepwhitematterandextendingintotheregionof therightthalamus.Thereis mildmasseffectbutnosignificantcontrastenhancementisdemonstrated.B:TheMR imageshowsa regionof high signalintensitycorrespondingto the abnormalitiesseen on the CT scan. Mild masseffect is also noted. C: The FDG PETimagesshowmarkedhypometabolismintherightparietallobe,extendingintothetemporallobe.Thesefindings areusuallynotedin low-gradetumors.

survival in patients with high grade gliomas (58—62). infarction, there is typically no abnormality seen on Metabolic abnormalities distant from solitarysupraten noncontrast CT. Approximately 12—24hr after the torial lesions have been noted, particularlyin the con ictus, an ill-defined hypodensity is frequently noted, tralateral cerebellar hemisphere (63). This phenomenon and mass effect due to edema may be present (69). has been termed “crossedcerebellar diaschisis―by However, some authors have reported CT findings Baron (64). Data from our laboratory suggest that which may make it possible to detect ischemic infarc ipsilateral hypometabolism may be a remote effect of tion earlier than 12 hr: increased density in a cerebral edema (65). vessel secondary to an embolus or thrombus (70), lower Unlike CT or MRI, PET has been shown to distin peak amplitude (in Hounsfield Units) of the affected guish radiation necrosis from tumor recurrence using hemisphere on dynamic scanning (71), and partial dis FDG (66, 67) and ‘‘CL-methiornne (56), and appears appearance of the lentiform nucleus in patients with to be the only imaging modality currently able to do embolic infarction of the middle cerebral or internal so. As might be expected, an area of radiation necrosis carotid artery distribution (72). Later on in the course appearsas a hypometabolic area, and recurrenttumor ofthe infarction,the affectedareamay become isodense asa hypermetabolicregion(Fig. 6). This featurealone on CT, and then proceed through a stage of contrast would make PET a valuable and unique tool in helping enhancement, termed “luxuryperfusion―.Finally, the plan the management of these patients. Some investi end stage of chronic infarction is reached, manifested gators have advocated the use of three dimensional as an area of well-defined hypodensity associated with correlation of CT, MRI, and PET images in radiation enlargement of the adjacent CSF space. therapy planning to better define the target volume Although therearecases in which CT is able to detect (68). early ischemic infarction, MRI has been reported to be more sensitive overall, showing abnormalities at 1—2 hrs after the ictus in animal studies (73). The typical appearance is an area which is of low signal intensity on Tl-weighted images and high signal intensity on T2- STROKE weighted images, resulting from increased regional wateraccumulation due to edema. In the chronic stage, Advances in anatomic and functional imaging have the infarct appears as a cyst-like CSF-intensity region. significantly furthered our understanding of the patho MRI can distinctly demonstrate hemorrhagic infarcts, physiology of brain ischemia and infarction. Immedi which may be missed on CT. While CT and MRI ately following cerebral arterial thrombosis or embolic provide valuable information on ischemic infarction, event, hypoperfusion of the brain in the involved arte PET is better able to track the actual sequence of rial distribution occurs. At this stage of an ischemic pathophysiologic events, particularlyin the earlystages.

1594 Jolles,Chapman,Alavietal TheJournalof NuclearMedicine @ ;1II@@@@ r@

PET has been reportedto demonstrate abnormalities are more extensive than would be predicted by CT or earlier than CT, and often the metabolic abnormalities MRI. In addition, the abnormal metabolic pattern seen on PET has been related to the type of clinical syn drome, and also to the degree ofeventual recovery (74, 75) (Fig. 7). Using 1NH@and @O-laheledagents. PET @sable to monitor cerebral blood flow and oxygen metabolism in

@(4@ P

B

FIGURE 6 A:Thisbraintumorpatientwhowasstudiedafterradiation therapy demonstrates a large high signal intensity abnor B mality on MRI in the left paneto-occupitalregion, with associated mass effect and edema. This MR appearance FIGURE 7 may be seenwith eithertumorrecurrenceor radiation A:Noncontrast-enhancedCTimageofa patientwitha left necrosis. B: FDG-PET images show an extensive hypo basal ganglia infarct shows a small hypodense region metabolicabnormalityinthesameregionasthe MRIscan. extendinganteriotlyfrom the superioraspectof the head These findings are consistent with radiation necrosis rather of the leftcaudatenucleus.Corticalatrophyis present, thanrecurrenttumor.(Withpermission:JollesPR,Chap particularly on the left. B: The FDG-PET image reveals manPR,AlaviA. Determinationof regionalcerebralfunc marked hypometabolismin the region of the left basal tin andstructureusingcomputedtomography,magnetic ganglia,as well as bifrontalhypometabolism.Note the resonance imaging and positron emission tomography. more extensive abnormalities seen on PET compared to AdvFuncNeuroimag1989;2:7—15). CT.

Volume30 •Number10 •October1989 1595 stroke patients. As may be expected in early infarction, mat injury and frequently does not correlate with the local cerebral blood flow (LCBF) decreases; there is a degree of neurologic deficit or with the level of con concomitant rise in the local oxygen extraction fraction sciousness as assessed by the Glasgow Coma Score (LOEF) in an attempt to deliver the maximal amount (GCS) (86). The latter is probably true for MM as well. of oxygen to the compromised region of brain. As the MRI is vastly superior to CT after 48—72hr post stroke evolves to the “luxuryperfusion―stage, LCBF trauma (in the subacute and chronic phases) for the typically increases compared to LOEF ( 76). However, detection of all intracranial lesions, most especially of groups of transient ischemic attack (TIA) and stroke nonhemorrhagic primary intra-axial lesions (83, 84, patients, some with occlusion of a carotid artery, have 86-88). A recent MRI study categorized these lesions been described with diminished LCBF at this stage; this into (a) diffuse axonal injury (DAI), (b) cortical contu has been termed “miseryperfusion―(77). Currently, it sion, (c) subcortical gray-matterinjury, and (d) brain is not certain whether these patients are at increased stem injury (88). The most common of these is DAI risk for infarction (78, 79); one patient presenting with which results from sheering stresses in the white matter a TIA and “miseryperfusion―syndrome on PET was tracts. DAI is multifocal and tends to be associated with describedrecently,who progressedto an infarction(80). more severe initial impairment of consciousness and a Areas of chronic infarction tend to exhibit diminished poorer prognosis (83, 87). Although MM is superior to LCBF and oxygen/glucose metabolism. CT in detecting DAI, both appearto be quite insensitive Additional hypometabolic foci distant from a cere in diagnosing these lesions. bral infarction are often noted on PET studies, typically CT does not reliably predict the outcome or extent in the absence ofcorresponding CT or MRI abnormal of recovery, in part due to its insensitivity in detecting ities (75, 81). These may occur in the ipsilateral or white matter lesions (86). Conversely, the higher sen contralateral cerebral hemisphere, or may be seen in sitivity of MM may lead to improved predictors of the contralateral cerebellar hemisphere; the latter was prognosis(83, 85, 87). termed “crossedcerebellar diaschisis―(64). These ab The literature on PET in head trauma is limited. normalities imply the interruptionofcrossing pathways FDG-PET does not differentiate structural lesions from in the brain (63), and can be seen with other supraten parenchymal dysfunction, all of which are manifested torial abnormalities such as brain tumors and head as reduced glucose metabolism (89). Functional trauma. changesmay extend beyond anatomic abnormalitieson Phosphorus-3 1 magnetic resonance spectroscopy CT or MRI (89, 90), and are often present in areas (MRS) has been used in the investigation of chronic adjacentto and remote from the focal damage. Cortical ischemia infarctions in humans (82). No findings spe contusion, intracranial hematoma and resultant en cific for infarction were described. Further research is cephalomalaciatend to have associatedmetabolic ef requiredto determine the diagnostic utility of MRS in fects that are confined to the site and extent of the stroke. anatomic lesion as seen on MM or CT. However, subdural and epidural hematomas cause widespread HEAD TRAUMA hypometabolism, not infrequently involving the contra lateral cerebral hemisphere, even when CT and MM The development of CT has had a dramatic impact are unremarkable in these regions (91) (Fig. 8). DAI upon the diagnosis and management of patients with can result in widespread cortical hypometabolism, but head trauma. More recently, MRI has been shown to a strikingfinding is ofprofound hypometabolism in the be superiorto CT in all clinical settings except in acute visual cortex (92), normally one ofthe most metaboli head trauma (83) where CT remains the modality of cally active regions. Preliminary work in our laboratory choice in the diagnosis of surgically correctable lesions suggeststhat supratentoriallesionswithhead injury can (84). CT is superior in detecting subarachnoid hemor cause cerebellar hypometabolism in both the ipsi- and rhage (83) and bony fracture, and in differentiating contralateralsides (Fig. 9). acute parenchymal hemorrhage from edema. Evalua There is correlationwith the severityofhead trauma, tion is slower and logistically more complex with MM, the GCS, and the degree of whole brain hypometabo particularlywith patients on life-supportsystems.MM lism (90). Global and regional metabolic rates tend to is increasingly becoming an effective alternative (85), improve along with clinical recovery (90, 91). DAI with multiplanar reconstruction enabling improved lo induced cortical hypometabolism has shown improve calization and assessment of a variety of lesions. Also, ment as earlyas 3 wk on serialPET scans (91)@Severely with effective means of monitoring acutely ill patients injured areasshow persistenthypometabolism. in the magnetic field, with shorter imaging time, and One study suggeststhat PET, but not CT, correlates with the better detection of parenchymal lesions, MM closelywith neuropsychologic(NP) and languagetest may play a more active role in the investigation of head ing in the assessment of site and extent of cerebral trauma.CT underestimatesnonhemorrhagicparenchy dysfunction with possible ramifications for rehabilita

1596 Jolles,Chapman,Alaviet al TheJournalofNuclearMedicine FIGURE 8 A 45-year-oldpatientwhowasinvolvedina motorvehicleaccident.A:TransaxialCT demonstratesa largeepidural hematomainthelefttemporalparietalregionat thetimeof presentation.B:A repeatCT scan4 dayslaterfollowing surgical decompression demonstrates marked reduction in both the size of the epidural hematoma and mass effect. C: FDG-PETat thetimeof therepeatCT scanshowswidespreadlefthemispherichypometabolismwhichisdispropor tionatelylargerthanthe originalsizeof the anatomiclesionseenon CT.

@4 FIGURE 9 A: TransaxialCT scanof a patientwhosustainedhead traumaina motorvehicleaccidentshowsa largeill-defined area of hypodensity in the right frontopanetal region with somelateralventriculareffacement.B: Protondensityand T2-weighted images show a large region of hyperintensity a intherighthemispherewithSparingof theOCcipitallobe, r'.r@‘@n. andrecenthemorrhageintherightfrontallobe.C: FDG PETdemonstratesmarkedhypometabolismof theentire right cerebral hemisphere induding the occipital lobe, and mild reduction in the left frontal lobe, extending beyond theMRIabnormality.Crossedcerebellardiaschisisisalso •1 present.

Volume30 •Number10 •October1989 1597 tion (93). Our preliminary data suggest substantial dif ity) (102—105),and (b) relative basal ganglia hyperme ferences between the sites of dysfunction identified by tabolism (105—107),with increased subcortical to FDG-PET and NP results. Further prognostic data from cortical ratios reflecting preservation of activity in sub PET, MRI and NP testing are required. cortical structures relative to decreased cortical metab olism (106) (Fig. 10). Blood flow studies using H215O have also demonstrated increasedflow to the left globus SCHIZOPHRENIA pallidus (108). Other findingsinclude left hemisphere Schizophrenia appears to be a heterogeneous disor hypometabolism, temporal lobe hypometabolism der, and this may in part account for the conflicting (105), and a normal anterior to posterior gradient. In and various results of many studies of brain morphol such studies patient populations, neuropsychological ogy and function. Hypotheses of the pathogenesis of assessmentand absenceor chronicityofmedication are schizophrenia are abundant. However, symptoms can not necessarily uniform. Furthermore, assessment of be broadlycategorizedas negativeand positive. Positive local cerebral glucose metabolic rates has tended not to symptoms tend to be much more responsive to neuro consider volume averagingeffects because of the prox leptic therapy than negative symptoms. imity of basal ganglia to lateral ventricles or of cortex Since the first report of ventriculomegaly on CT in to widened sulci. Short term neuroleptic treatment schizophrenia (94), it has been the most consistent tends not to significantly change either the relative finding in over 90 CT studies of structural abnormali hypofrontality (105) or the subcortical to cortical ratio ties. Other less frequent abnormalities include global (106, 109), but does increase global metabolic rates and frontal cortical atrophy and sulcal widening, cere (105). Hypofrontality may relate to chronicity of illness bellar atrophy, cerebral asymmetry, third ventricular or medications (103, 104). dilatation, decreasedcranial size and alteredbrain den The concept of hypofrontality is consistent with the sity when compared with normal controls. These have hypothesis of frontal lobe dysfunction (97, 105). The been variably correlated with prognosis and preponder frontalcortex is believed to be important for the highest ance of positive or negative symptoms (95, 96). The integrative,adaptive, and executivefunctions. Frontal poor specificity and low frequency (3—40%)(96, 97) of lobe dysfunction could account for the prominence of ventriculomegaly has not warranted routine CT exam negativesymptoms(97, 105).Furthermore, FDG-PET ination in suspected schizophrenic patients. studies with specific frontal activation tests of sustained Ventriculomegaly does occur in young, first episode attention by continuous performance tasks, have dem patients (97) and tends not to progress as would be onstrated reduced activation of the frontal lobe (97, expected in a degenerative process such as dementia 104, 110). In one study, the degree of accentuation of (95). These raise questions as to the onset of the path hypofrontality correlated with the preponderance of ological process and temporal relationship to clinical negative, rather than positive, symptoms (104). How manifestations, and to the possibility of the disorder ever, some recent studies have not documented hypo resulting from a “staticlesion―early in brain develop frontality (111). ment (98). In most patients positive and negative symptoms Published studies ofbrain morphologic abnormalities occur simultaneously. It has been suggested that the on MRI are difficult to compare because of methodo logic differences. As with CT, lateral ventricular dila tation has been noted with MM (95, 99). The better resolution of MRI has enabled detection of corpus callosal abnormalities, with both increased and de creased size being reported (99). This is relevant as failure of interhemispheric communication occurs in some schizophrenics (100). As with CT, no consistent focal structural abnormality has been detected to ac count for the ventriculomegaly (95). This may suggest a generalizedprocess or that the resolution ofMRI may still be inadequate to detect small regions of atrophy, such as in the medial limbic structures, as has recently been described in neuropathologic studies (101). FDG-PET is employed on the assumption that aber rant neuronal activity will be manifested as abnormal FIGURE 10 FDG-PETina youngpatientwithschizophreniashowsa glucose metabolism. FDG-PET studies have demon relative striatal hypermetabolism, and increased striatal to stratedconflicting results. The two most frequent find cortical ratio. The entire cortex appears to be uniformly ings are (a) frontal lobe hypometabolism (hypofrontal affectedbytheunderlyingprocess.

1598 Jolles,Chapman,Alavietal TheJournalofNuclearMedicine temporal and frontal lobes, together with their dopa in nonpsychotic affective illnesses (119). Iffurther stud minergic interconnections through the basal ganglia, be ies confirm a definite increase in D2 receptor density, considered a neurobehavioral unit, and that dysfunc this will significantly affect future research. tion of this unit may be involved in the positive symp Although experience is limited to date, neurorecep toms of schizophrenia (105). Frontal lobe dysfunction tor/neurotransmitter imaging and the evaluation of the would not only lead to therapy-resistantnegative symp effects of drugs on regional brain function and chem toms, but to overaction of subcortical and limbic do istry should increase in importance, and facilitate in paminergicsystems that are normally modulated by the vivo testing of neurophysiologic theories of schizophre frontal cortex, resulting in positive symptoms (112), nia and other psychiatricdisorders. sensitive to neuroleptics. This may be the basis for the various findings of hypofrontality and relative basal ganglia hypermetabolism seen in FDG-PET studies. MOVEMENT DISORDERS Anatomic and metabolic studies have not unravelled the pathophysiology of schizophrenia. Validation of Huntington's Disease theories, such as the above, will require in vivo studies Huntington's disease (HD) is a lethal neurodegener of neurotransmitter systems. This is an exciting area of ative disease ofthe caudate and putamen, of autosomal greatpotential for PET research. dominant inheritance with onset usually in the third Pharmacologicand neuropathologicstudies have im through fourth decades, and characterized by progres plicated the dopamine system in the pathogenesis of sive chorea and dementia. Recently, an abnormality of schizophrenia, the so-called “dopaminehypothesis―. the short arm of chromosome 4 has been described Neuroleptics cause D2 receptor blockage proportional (120). to antipsychotic efficacy, and amphetamines, which CT and MM identify the pathologically described elevate synapticdopamine levels, induce a schizophren changes of atrophy of the caudate nuclei, appearing as iform psychosis and exacerbate symptoms in schizo loss of convexity of the lateral aspect of the adjacent phrenia. Post-mortem studies have shown a bimodal lateralventricle(121, 122). With progressionof demen patternof D2 receptordensity in brains of schizophren tia and with disease duration, generalizedcerebralatro ics, with half of the subjects having normal and half phy becomes worse(121, 122). FDG-PET demonstrates having increaseddensities (113) in the caudate nucleus, hypometabolism of the caudate and putamen nuclei in putamen and nucleus accumbens. D, receptornumbers all symptomatic subjects(123, 126). In contrast to AD, have remained normal. Controversy exists as to whether global and local cortical metabolic rates are normal, the increase in D2 receptor binding and density results and are independent of disease severity, duration, and from an adaptation to chronic neuroleptic blockade or medications(124). from an intrinsic biochemical defect. Post-mortem There is debate as to whether PET can detect asymp studies of drug-free or naive patients have shown in tomatic HD carriers. In two series of at-risk subjects, creased D2 receptor binding (114). striatal hypometabolism preceded the onset of symp Severalagents have been reportedfor the imaging of toms and CT changes(123-125). Clinical follow-up has CNS D2 receptors, including butyrophenone deriva shown that four of 18 (123) and one of three (125) tives, 3-N-(' ‘C)methyl spiperone ([‘‘C]NMSP),and such subjects with stnatal hypometabolism, but none [@F]NMSP, and the benzamide, [“C]raclopnde.Stud with normal metabolism, have with time developed ies on patients undergoing therapy show suppression of symptomatic HD. More protracted follow-up is re striatal radioligand uptake and affinity, which is in quired.It does appearthat an abnormal FDG-PET scan versely related to antipsychotic therapy dose and plasma could indicate that the subject may be presymptomatic level (115). A debate currently exists about the striatal for some undefined period. A normal FDG-PET scan D2 receptor density in drug-naive schizophrenics. In would suggest either that the subject is not an HD creased density was found using [“C]NMSP[This oc heterozygote or that clinical manifestation may be at curred not only in drug-naive but in previously treated least some years away (125). However, another study but drug-free patients (116)]. However, using [“C]ra could not demonstrate caudate hypometabolism in any clopride the receptor density was normal (117). at-riskindividual (126). (“C)NMSPis less specific, having some serotomn In a study of FDG-PET and a polymorphic human receptoractivity, and binds irreversiblyto D2receptors. linked DNA marker (D4S1O), in seven of 13 at-risk It is possiblethat [‘‘Cjraclopride,which binds reversibly subjects there was concordance ofthese two tests in the and dissociates rapidly, may be more subject to com indication of either a high or low risk of developing petition from endogenousdopamine, which is signifi HD. Five subjects had abnormal DNA results and cantly elevated in actively psychotic schizophrenics normal PET, and one subject had the reserve. This (118). Elevated striatal D2 receptor density has been study suggeststhat FDG-PET may be useful to confirm reportedin other psychotic affective disorders,but not abnormal DNA studies, to check for the possibility of

Volume30 •Number10 •October1989 1599 an erroneously abnormal DNA test because of DNA nesia and development of mild to moderate dementia recombination, or to suggest that there may be an (133). The mechanism of the mild diffuse cortical undefined period, perhaps years, before symptoms be hypometabolism,which is unrelated to disease dura gin if PET is normal and DNA results are abnormal tion, is unclear(133). Severedementia ofPD may have (125). an FDG-PET appearance indistinguishable from AD Stnatal glucose hypometabolism is not entirely spe with selective parietal hypometabolism (133). cific for HD, occurring in hereditary chorea and Lesch Reduced striatal dopamine concentrations result Nyhan syndrome. Therefore, DNA and PET studies from decreased synthesis from dopa and decreased stor are complementary in evaluating at-riskindividuals for age. As would be expected, PET with ‘8Fdopa in PD HD (125). demonstrates reduced basal ganglia activity especially D2 receptor binding studies with (‘‘C)NMSPhave in patients with the “on/off―phenomenon (136). In demonstrated a reduced D2receptor density in the basal hemiparkinsonism, there is predominant reduction in ganglia of patients with HD (127—129),consistent with the contralateral striatum, but also some irregular re postmortem autoradiographicfindings (127). duction in the ipsiateral side (137). Nomifensine is a potent inhibitor of presynaptic do Parkinson's Disease pamine and norepinephrine reuptake sites. In normal Parkinson's disease (PD) was the first illness in which ity, the highest regional brain uptake occurs in the an abnormality of a neurotransmitter was established. striatal dopaminergic nerve terminals of the nigrostri PD is characterizedby the triad ofbradykinesia, tremor atal pathway and in the norepinephrine uptake sites in and rigidity with progression to dementia in 20—30%of the thalamus (136). There is an age-dependent decline patients. There is loss of pigmented neurons in the (138). In PD, striatal uptake of “Cnomifensine is substantia nigra ofmidbrain and locus coerulus of pons markedly reduced in patients less than 65 yr of age. with reduced production of dopamine from decarbox Older PD patients have striatalbinding similar to older ylation of dopa, reduced storage of dopamine, and controls. In hemiparkinsomsm thereis reducedbilateral nigrostriatal system dysfunction. It has been hypothe striatal uptake, especially on the side contralateral to sized that early in PD this denervation is associated symptoms (138). Development of an “Cnomifensine with a hypersensitivitydue to “upregulation―of striatal enantiomer would refine this technique. dopamine receptors (130) and that down regulation Consistent with the post-synaptic supersensitivity may occur as the disease progresses. theory, PET with [‘‘C]NMSPsuggesteda mild increase CT has little role in PD, detecting no specific striatal in binding in early disease, and decreased binding in abnormality and occasionally only mild, nonspecific advanced disease, consistent with subsequent down ventricular, and sulcal enlargement. The major feature regulation (127). Patients with hemiparkinsonism have of PD on MRI appearsto be a decrease in the width of demonstrated bilateral variability in striatal uptake the pars compacta of the substantia nigra, probably (127, 139). reflecting loss of pigmented cells, but possibly due to Treatment with dopaminergic medication would be iron deposition (131). Moderate or marked cortical expected to decrease [“CJNMSPbinding in the basal atrophy tends to occur more frequently than in controls. gangiia because of competition with therapy-induced In parkinsonian-like syndromes MM shows abnormal dopamine, down-regulation of the dopamine receptor intensities in the putamen and to a lesser degree in the with therapy, and because of decreased receptor num caudate nuclei and substantia nigra, suggestiveof iron bers with age. Surprisingly, one group found a bilateral deposition (132). Iron is present in many enzyme sys increase in receptor binding following therapy, suggest tems, and deposition may reflect changes in these sys ing eitherpersistentpost-synapticD2receptorhypersen tems with pathology or age. sitivity, or an alteration in local pharmacokinetics As the synapse is a site of significant glucose utiiza (139). tion, nigrostriatal atrophy may be expected to cause Exposure to the “designerheroin―,l-methyl-4- reductionin localstriatalglucosemetabolicrates.Con phenyl 1,2,3,6-tetrahydropyndine (MPTP), can cause versely, in the early untreated state mild basal ganglia an acute parkinsonism. MPTP exposed monkeys and hypermetabolism has been reported (133), and in humans demonstrate reduced striatal ‘8Fdopa uptake hemiparkinsonism, hypermetabolism of the basal gan (140, 141), and marked up-regulation of D2 receptors glia contralateral to the symptomatic side has been seen using [‘‘C]raclopride(142) and [‘8F]NMSP(143). The (134). Another group detected no significant striatal changes seen in MPTP subjects are more marked than change (124, 135). Levodopa therapy does not alter in PD. However, MPTP animals may serve as a useful local or global metabolic rates (133, 135), despite sig model for study of PD. nificantclinicalimprovement.PD patientstend to have Although much is known about the biochemical mild, uniform cerebral hypometabolism (average 18% abnormalities of PD, the functional problems in the decrease), which worsens with the severity of bradyki basal ganglia remain unclear. In vivo neurotransmitter

1600 Jolles,Chapman,Alaviet al TheJournalof NudearMedicine studies should aid in the understanding of PD and assessment of therapy, such as the use of caudate im plants.

EPILEPSY The epilepsies are a heterogeneousgroup of disorders constituting, as a group, one of the more common serious neurologic illnesses. Not all patients with gen eralized (GS) or partial seizures (PS) can be controlled with medical therapy. Intractable complex PS can be cured by surgery in appropriatelyselected candidates. The preoperative localization of the ictal focus entails multiple investigations including FDG-PET, electro encephalography (EEG), CT, and MM. PS most frequentlyoriginate from a structurallesion in the temporal lobe, resulting in a complex constella tion of psychomotor phenomena. The most common pathology is mesial temporal sclerosis (MTS) in which there are one or more foci of neuronal loss and gliosis, A involving the hippocampus, dentate, subiculum and amygdala(144).Occasionally,tumors, vascularmalfor mations or tuberous sclerosis are the cause. CT has a limited role, being particularly hampered in the temporal lobe by bone artifact, but will detect R larger tumors (145-148). MM affords better resolution and parenchymalresolution without bone artifact.MM is very sensitive in the detection of nonsclerotic lesions, but has been reportedas being poor in detecting MTS, being normal in a seriesof 18 patients with MTS (148). B In another series, only one of 14 patients with patho FIGURE 11 logically confirmed MTS had focal MM signal abnor A: TransaxialMR imagein a patientwith complexpartial malities (149) (Fig. 11). Atrophy, as determined by seizures shows focal hypenntensities consistent with me gross asymmetry, sulcal and temporal horn enlarge sialtemporalsclerosis.B: InterictalFDG-PETscanshows ment, occurs more frequently (150). These changes left temporalhypometabolismin an area largerthan that may not always correlate with pathologic evidence of seen on MRI. MTS (149). EEG is vital in the diagnosis and classification of The extent to which the interictal hypometabolic epilepsy, and in the preoperative localization of the focus representsthe anatomic abnormalityand/or func epileptogenic focus. Scalp EEG detects primarily coth tional derangement of metabolism is not entirely clear, cal activity and, having low spatial resolution, cannot but it probablymore representsthe latter. Such a focus differentiate primary cortical activity from that propa usually involves the mesial temporal lobe and may gated from another site. Stereotactically implanted extend to the entire lobe, to the frontalor parietallobes, depth electrodes improve localization, but are clearly or to the ipsilateral hemisphere. There is individual invasive and thereforehave sampling limitations. variation of size and shape of the zone. Since the calculation of metabolic rates requires The degree of hypometabolism does not correlate steady state conditions for 30—40mm following injec with the frequency of EEG spike activity (151, 152), tion before equilibrium is attained, FDG-PET has poor nor with seizure frequency or severity (152). PET and temporal resolution of changes in glucose metabolism. EEG appear to measure different aspects of cerebral Thus, any transient metabolic changes, as occur with a dysfunction. seizure, will be averaged with the pre and postictal In localizing the seizure focus, there was poor corre metabolicrates.Furthermore,the onset ofa PS may be lation when PET was compared with scalp and depth unpredictable. Therefore, most FDG-PET studies are EEG separately.When combined EEG data were corn performed in the interictal phase. The most consistent pared with PET, the correlation was good, with both finding,occurringin -@-7O%of PS patients, is of one or localizing the focus in @@-7O%of patients (153). When more foci of hypometabolism in the region of the the EEG focus was unilateral and well defined, there primary epileptogenic focus or foci (151). wasgood correlationbetween PET and scalpand sphen

Volume 30 •Number 10 •October1989 1601 oidal EEG. Diffuse or shifting EEG abnormalities were onstrate a diffuse hypermetabolism with postictal hy usually associated with normal PET studies (154). Both pometabolism (157). Petit mat seizures were associated PET and EEG can cause false positive localization in with a diffuse increase in cerebral metabolism relative terms of histological findings (144). The very presence to post-therapy seizure-controlled state (160). of depth electrodes in situ during PET may cause foci With furtherimprovement in spatial resolution with of hypometabolism (144). PET has better spatial reso PET and higher Tesla MR instruments the relative lution than EEG (155), but the zone of hypometabo contributions of these technologies will become more lism may at times be larger than the region suggested refined. Much is still to be learnt about the electrophys by EEG (152). iological and pathophysiologic significance ofthe inter MRI is clearly superior to PET in detecting non ictal zone of hypometabolism in PS patients. sclerotic structural lesions that cause seizures (148). PET appearsto be superior to MM in the detection of ACKNOWLEDGMENTS MTS. In the abovementioned series,of 18 patients with normal MRI, 56% had temporal lobe hypornetabolism This workwassupportedby NIH Grants NS14867-l0and (148). However, hypometabolism cannot specify his NIH AG03934-07. topathology. The extent of hypometabolisrn is generally larger REFERENCES than the structural abnormality on histopathology of 1. Brody H. Organization of the cerebral cortex. III. A the resected specimen (144). This suggests a functional study ofaging in the human cerebral cortex. J Comp inhibition of adjacent neuronal elements (144). Fur Neurol1955;102:511—556. thermore, these functional changes appear to be, in 2. Tomlinson BE, Blessed 0, Roth M. Observations on part, reversible. During a seizure the hypometabolic the brains of nondemented old people. J Neurol Sci zone may become hypermetabolic (155). Hypometa 1968; 7:33 1—356. 3. Anderson JM, Hubbard BM, Coghill GR, et at. The bolic regions in a patient with Lennox-Gestaut syn effectof advanced old age on the neurone content of drome normalized following comrnissurotomy (156). the cerebral cortex: observations with an automatic Postictal scans may show an increased number of hy image analyser point counting method. J Neurol Sci pornetabolic foci when compared to interictal scans 1983; 58:235—246. (154). Therefore, regions of hypometabolism may vary 4. Haug G. Age and sex dependence of the size of normalventricleson computedtomography.Neuro and should not be solely relied upon for localization. radiology1977;14:201—204. MRI, FDG-PET, and EEG are complementary in 5. Zatz LM, JerniganTL, Ahumada AJ, Jr.Changeson foci localization. MRI is the preferable initial imaging computed cranial tomography with aging: intracra investigation, excluding most tumors. Consistent local nial fluid volume. AJNR 1982; 3:1—11. ization by scalp EEG and PET may eliminate the need 6. de leon MJ, George AE, Ferris SH, et at. Positron emission tomography and computed tomography as for depth EEG (154). sessment of the aging human brain. JCAT 1984; Ictal FDG-PET is of limited value in PS patients 8:88—94. because of poor temporal resolution, averaging of met 7. George AE, de Leon Mi, Kalnin A, et at. Leukoen abolic rates, significant individual variation and the cephalopathy in normal and pathologic aging: 2. MRI variability of ictus onset. In general, the interictal focus ofbrain lucencies.AJNR 1986;7:567—570. 8. Brant-Zawadski M, Fein G, Van Dyke D, et at. MR of hypometabolism becomes hypermetabolic during a imaging of the aging brain: Patchy white-matter le seizure (154, 157) and may spread to become diffuse. sions and dementia. AJNR 1985;6:675—682. The period of postictal depression is associated with 9. George AE, de Leon MJ, Gentes CI, et at. Leukoen marked hypometabolism. cephalopathy in normal and pathologic aging: 1. CT In patients with epilepsia partialis continua or with ofbrain lucencies. AJNR 1986; 56 1—566. 10. Braffman BH, Zimmerman RA, Trojanoqski JQ, et multiple frequent seizures, there tends to be widespread at. Brain MR: Pathologic correlation with gross and hypermetabolism rather than a discrete hypermetabolic histopathology.2. Hyperintensewhite-matterfoci in focus. The areasofictal hypermetabolismdo not always theelderly.AJNR 1988;9:629—636. correspond to sites ofinterictal hypometabolism (157). 11. Drayer BP, Imaging of the aging brain. Part 1. Ra The use of the shorter lived positron emitter ‘@Ohas diology1988;166:785—796. 12. Fazekas F, Chawluk JB, Alavi A, et at. MR signal not aided localization, as reports of CBF, CMRO2, and abnormalities at 1.5 T in Alzheimer's dementia and OER in interictal and ictal phases have given variable normal aging.AJNR 1987;8:421—426. results (158, 159). 13. Duara R, Margolin RA, Robertson-Tchabo EA, et Intenctal and ictal FDG-PET studies on patients with at. Cerebral glucose utilization as measured with pos generalized tonic/clonic and petit mal seizures, except itron emission tomography in 21 resting healthy men between the ages of 21 and 83 years. Brain 1983; with Lennox-Gastaut syndrome, have failed to lateral 106:761—775. ize or localize an abnormal focus (152). Scans per 14. deLeon M, George A, Tomanelli J, et at. PET scan formed during electroconvulsive therapy tend to dem fling demonstrates that aging in man is accompanied

1602 Jolles,Chapman,Alavietal TheJournalof NuclearMedicine by loss of integration of regional brain activity with size in patients with presenile dementia of the Alz out concurrent reduction in absolute regional cere heimer type. Arch Neurol 1984;4 1:1258—1263. brat metabolic rates for glucose.J Cereb Blood Flow 33. Damasio H, Eslinger P, Damasio AR, et at. Quanti Metab 1985; 5:S119. tative computed tomography analysisin the diagnosis 15. Alavi A. The aging brain. J Neuropsychiatry 1989; ofdementia. ArchNeurol 1983;40:715—719. 1(suppll):S5 1—555. 34. George AE, Stylopoulos LA, de Leon MJ, et at. 16. DasturDK. Cerebralblood flow and metabolismin Temporal lobe CT diagnostic features of Alzheimer's normal human aging, pathological aging, and senile disease [Abstract].Presented at the 25th Annual dementia. J Cereb Blood Flow Metab 1985;5:1—9. Meeting ofthe ASNR, New York, May 1987.AJNR 17. Kuhl DE, Metter El, Rieger WH, Phelps ME. Effects 1987;8:931. of human aging on patterns oflocat cerebral glucose 35. Pettegrew JW, Moossy J, Withers G, et at. 31-P utilization determined by the [‘8F]fluorodeoxyglu nuclear magnetic resonance study of the brain in cose method. J Cereb Blood Flow Metab 1982; Alzheimer's disease.JNeuropatholExp Neurol 1988; 2: 163—171. 47:235—248. 18. Chawluk JB, Alavi A, JamiesonDO, et at. Changes 36. Pettegrew JW, Koop SJ, Minshew NJ, et at. 31-P in local cerebral glucose metabolism with normal nuclear magnetic resonance studies of phosphoglyc aging, the effects of cardiovascular and systemic eride metabolism in developing and degenerating health factors. J Cereb Blood Flow Metab 1987; brain: Preliminary observations. J Neuropathol Exp 7(suppl l):S4l1. Neurol 1987; 6:419—430. 19. de Leon M, GeorgeA, Tomanelli J, et at. Positron 37. Alavi A, Lenkinski R, Chawluk J, et at. Correlation emission tomography studies of normal aging, a rep of positron emission tomography (PET) and mag lication of PET III and 18-FDG using PET VI and netic resonance spectroscopy (MRS) in CNS disor II-CDG. NeurobiolAging 1987; 8:319—323. ders[Abstract].JNuclMed 1988;29:910. 20. Chawluk J, Maw A, Hurtig H, et at. Altered patterns 38. Duara R, Grady C, Haxby J, et at. Positron emission of regionatcerebralglucosemetabolismin agingand tomography in Alzheimer's disease. Neurology 1986; dementia. J Cereb Blood Flow Metab 1985; 5(suppl 36:879—887. 1):Sl21—S122. 39. Alavi A, Chawluk J, Hurtig H, Ctat. Determination 21. AlaviA, Dann R, ChawlukJ,et at.Positronemission of patternsof regionalcerebralglucose metabolism tomographyimagingofregionatcerebralglucoseme in normalaginganddementia[Abstract].JNucl Med tabolism. Semin NuclMed 1986; 16:2—34. 1985;26:P69. 22. Sokoloff L. Cerebral circulatory and metabolic 40. Alavi A, JollesPR, JamiesonDO, et at. Anatomical changesassociatedwith aging. Res Publ Assoc Res and functionalchangesof the brainin normalaging Nerv Ment Dis 1966; 41:237—254. and dementia as demonstratedby MM, CT, and 23. Frackowiak RSJ, WiseRJS, Gibbs JM, et at. Positron PET.Nud MedAnn 1989:49—79. emission tomographicstudiesin agingand cerebro 41. Frackowiak RSJ, Pozzilli C, LeggNJ, Ctat. Regional vasculardiseaseatHammersmithHospital.AnnNeu cerebral oxygen supply and utilization in dementia: rol1984;l5(suppl):Sl12—118. aclinicalandphysiologicalstudywithoxygen-15and 24. Pantano P, Baron JC, Lebrum-Grandie P, et at. positron tomography. Brain 1981; 104:753—778. Regional cerebral blood flow and oxygen consump 42. Frackowiak RSJ, Pozzilli C, Legg NJ, et at. A pro tion in human aging. Stroke 1984; 15:635—641. spective study of regional cerebral blood flow and 25. Frackowiak RSJ, Lenzi GL, Jones T, et at. Quanti oxygen utilizationin dementia using positronemis tative measurement of regional cerebral blood flow sion tomography and oxygen-l5. J Cereb Blood Flow and oxygen metabolismin man using 150 and posi Metab 1981; 1:5453—5454. tron emission tomography;theory, procedureand 43. JamiesonDG, ChawlukJB,AlaviA, et at.The effect normal values.JCAT 1980;4:727—736. of disease severity on local cerebral glucose metabo 26. Yamaguchi T, Kanno I, Uemura K, et at. Reduction lism in Alzheimer's disease. J Cereb Blood Flow in regionat cerebral metabolic rate of oxygen during Metab 1987;7(supp l):S4l0. human aging.Stroke1986;17:1220—1228. 44. Benson F, Kuhi DE, HawkinsME, et at. The fluo 27. Beck JC, Benson DF, Scheibel AB, et at. Dementia rodeoxyglucose ‘28Fscan in Alzheimer's disease and in the elderly: the silent epidemic. Ann Intern Med multi-infarct dementia. Arch Neurol 1983;40:711. 1982;97:231—241. 45. Kamo H, McGeer R, Harrop R, et at. Positron 28. Jellinger K. Neuropathological aspects of dementia emission tomographyand histopathologyin Pick's resulting from abnormal blood and cerebrospinat disease.Neurology1987;37:439. fluid dynamics. Acta NeurolBelg 1976;76:83—102. 46. Chawluk JB, Alavi A, Dann R, et at. Positron emis 29. Ho DD, Rota TR, Schooley RT, et at. Isolation of sion tomography in aging and dementia: effect of HTLV-IIIfromcerebrospinatfluidandneuraltissues cerebral atrophy. JNuclMed 1987;28:431—437. of patientswith neurologicsyndromesrelatedto the 47. Bradley WG, Jr, Watuch V, Yadley RA, et at. Com acquired immunodeficiency syndrome. N Engl J parison ofCT and MR in 400 patients with suspected Med 1985;313:1493—1497. diseaseof the brainand cervicalspinalcord.Radio! 30. Tomlinson BE, BlessedG, Roth M. Observations on ogy 1984; 152:695—702. the brains ofdemented old people.JNeurolSci 1970; 48. Kucharczyk W, Brant-Zawadski M, Sobel D, Ct at. 11:205—242. Central nervous system tumors in children: detection 31. Gado MH, Hughes CP, Danziger W, et at. Volumet by magnetic resonance imaging. Radiology 1985; tic measurements ofthe cerebrospinal fluid spaces in 155:131—136. subjects with dementia and in controls. Radiology 49. FelixR, SchornerW, LaniadoM, et at.Braintumors: 1982; 144:535—538. MR imaging with gadolinium-DTPA. Radiology 32. Albert M, Naeser MA, Levine HL, et at. Ventricular 1985;156:681—688.

Volume 30 •Number 10 •October1989 1603 50. Kelly WM, Brant-Zawadski M. Magnetic resonance JCAT 1987;11:948—954. imaging and computed tomography of supratentonal 69. Wang A-M, Lin JC-T, Rumbaugh CL. What is ex tumors. In: Taveras JM, Ferrucci JT, eds. Philadel pected of CT in the evaluation of a stroke? Neurora phia: Radiology: diagnosis-imaging-intervention. diol1988;30:54—58. Chap 53. J.P. Lippincott. 1988:1—21. 70. Pressman BD, Tourje El, Thompson JR. An early 5 1. Hesselink JR, Press GA. MR contrast enhancement sign of ischemic infarction: increased density in a of intracranial lesions with Gd-DTPA. Radio! Clin cerebral artery. AJNR 1987; 8:645—648. NorthAm 1988;26:873—887. 71. Shih TTF, Huang K-M. Acute stroke: detection of 52. Brant-Zawadski M, Davis PL, Crooks LE, et at. changesin cerebral perfusion with dynamic CT scan NMR demonstration ofcerebral abnormalities: corn ning. Radiology 1988; 169:469—474. parison with CT. Am JRoentgenol 1983; 140:847. 72. Tomura N, Uemura K, Inugami A, et at. Early CT 53. Brant-Zawadski M, Badami JP, Mills CM, et at. finding in cerebral infarction: obscuration ofthe len Primary intracranial tumor imaging: a comparison tiformnucleus.Radiology1988;168:463—467. of magnetic resonance and CT. Radiology 1984; 73. Spetzler RF, Zabramski JM, Kaufman B, et at. Acute 150:435—440. NMR changes during MCA occlusion: a preliminary 54. Dooms OC, Hecht 5, Brant-Zawadski M, et at. Brain study in primates. Stroke 1983; 14:185—190. radiation lesions: MR imaging. Radiology 1986; 74. Kushner M, Reivich M, Fieschi C, et at. Metabolic 158:149—155. and clinical correlates of acute ischemic infarction. 55. WrennFR,GoodML, HandlerP.Theuseof posi Neurology1987;37:1103—1110. tron emitting radioisotopes for the localization of 75. Kuhl DE, Phelps ME, Kowell AP, et at. Effect of braintumors.Science1951;113:525—527. stroke on local cerebral metabolism and perfusion: 56. O'Tuama LA. Methionine transport in brain tumors. mapping by emission computed tomography of JNeuropsychiatry 1989; 1(suppl 1):S37—S44. ‘8p@j@and ‘3NH3.Ann Neurol 1980; 8:47—60. 57. Hiesiger E, Logan J, Wolf AP, et at. Serial PET 76. Lassen NA. The luxury perfusion syndrome and its studies of human cerebral malignancy with [1-'‘C] possiblerelation to acute metabolic acidosislocalized putrescine (1lC-PUT) and [1-'‘C]2-deoxy-D-glucose within the brain. Lancet 1966;2:1113—1115. (“C-2DG)[Abstract].JNuclMed 1986;27:889. 77. Baron JC, Bousser MG, Rey A, et at. Reversal of 58. Di Chiro G. Positron emission tomography using focal “misery-perfusionsyndrome―by extracranial [‘8f@fluorodeoxyglucose in brain tumors. A powerful bypass in hemodynamic cerebral ischemia. Stroke diagnostic and prognostic tool. Invest Radio! 1986; 1981;12:454—459. 22:360—371. 78. Ackerman RH, Correia JA, Alpert NM, et at. Posi 59. Di Chiro G, DeLaPaz RL, Brooks RA, et at. Glucose tron imaging in ischemic stroke disease using com utilization of cerebral gliomas measured by ‘8Ffluo pounds labeled with oxygen 15. Arch Neurol 1981; rodeoxyglucose and positron emission tomography. 38:537—543. Neurology 1982; 32:1323—1329. 79. FrackowiakRSJ,WiseMS. Positrontomographyin 60. Patronas NJ, Brooks RA, DeLaPaz RL, et at. Gly ischemic cerebrovascular disease. Neurologic Clinics colytic rate (PET) and contrast enhancement (CT) in 1983; 1:183—200. human cerebral gliomas. AJNR 1983;533—535. 80. Itoh M, Hatazawa J, Pozzilli C, et at. Positron CT 61. Alavi J, Alavi A, Dann R, et at. Metabolic brain imaging of an impending stroke. Neuroradiol 1988; imaging correlated with clinical features and brain 30:276—279. tumors [Abstract].J Nuc!Med 1985; 18:P64. 81. Pappata S, Tran Dinh S, Baron JC, et at. Remote 62. Alavi JB, Alavi A, Chawluk J, et at. Positron emission metabolic effectsofcerebrovascularlesions: magnetic tomography in patients with glioma. A predictor of resonance and positron tomography imaging. Neu prognosis.Cancer1988;62:1074—1078. roradiol1987;29:1—6. 63. Kushner M, Tobin M, Alavi A, et at. Cerebellar 82. Welch KMA, Ewing JR. Chopp M, et at. In vivo3l- glucoseconsumption in normal and pathologic states P NMR of adult human brain.J CerebBlood Flow using fluonne-FDG and PET. J Nuc! Med 1987; Metab 1985; 5(suppl l):S59l—S592. 28:1667—1670. 83. KellyAB,ZimmermanRD, SnowRB, et at. Head 64. Baron JC, BousserMG, Comar D. Crossedcerebellar trauma: comparison of MR and CT-experience in diaschisis;a remote functional depression secondary 100patients. AJNR 1988;9:699—708. to supratentoriat infarction in man. J Cereb Blood 84. Levin HS, Amparo E, EisenbergHM, et at. Magnetic Flow Metab 198 1; 1(suppl l):S500—S501. resonance imaging and computerized tomography in 65. Alavi JB, Alavi A, Powe J, Hackney D, Reivich M. relation to the neurobehavioural sequelae of mild Metabolic and structural findings related to white and moderate head injuries. J Neurosurg 1987; matter disorders as shown by PET, MRI and CT 66:706—713. [Abstract]. JNuc!Med 1986; 27:899. 85. Hadley DM, Teasdale GM, Jenkins A, et at. Magnetic 66. Doyle WK, Budinger TF, Valk PE, et at. Differentia resonance imaging in acute head injury. Clin Radio! tion of cerebral radiation necrosis from tumor recur 1988;39:131—139. rence by [‘8flf@J@and 82Rb positron emission to 86. Gentry LR, Godersky JC, Thompson B, et at. Pro mography.JCAT 1987;11:563—570. spectivecomparative study of intermediate-field MR 67. Patronas NJ, Di Chiro G, Brooks RA, et at. Work in and CT in the evatuation ofclosed head trauma. Am progress: (l8F) fluorodeoxyglucose and positron JRoentgeno!1988;150:673—682. emission tomography in the evaluation of radiation 87. Zimmerman RA, Bilaniuk LT, Hackney DB, et at. necrosis ofthe brain. Radiology 1982; 144:885—889. Head injury: early results ofcomparing CT and high 68. Schad LR, Boesecke R, Schlegel W, et at. Three fieldMR. Am J Roentgenol1986;147:1215—1222. dimensional correlation ofCT, MR. and PET studies 88. Gentry LR, Godersky JC, Thompson B. MR imaging in radiotherapy treatment planning of brain tumors. of head trauma: reviewof the distribution and radio

1604 Jolles,Chapman,Alavietal TheJournalof NuclearMedicine pathologic features oftraumatic lesions. Am JRoent 109. Gur RE, Resnick SM, Gur RC, et at. Regional brain geno! 1988; 150:663—672. function in schizophrenia: II. Repeated evaluation 89. Langfitt TW, Obnst WD, Alavi A, et at. Computer with positron emission tomography. Arch Gen Psy ized tomography, magnetic resonance imaging, and chiatry 1987; 44:126—129. positron emission tomography in the study of brain 110. Cohen RM, Semple WE, Gross M, et at. From syn trauma. Preliminary observations. JNeurosurg 1986; drome to illness: delineating the pathophysiology of 64:760—767. schizophrenia with PET. Schizophr Bull 1988; 90. Alavi A, Fazekas T, Alves W, et at. Positron emission 14:169—176. tomography in the evaluation ofhead injury. JCereb 111. Kling AS, Metier El, Riege WH, et at. Comparison Blood Flow Metab 1987; 7(suppl l):S646. of PET measurement of local brain glucose metabo 91. George JK, Alavi A, Zimmerman RA, et at. Mets lism and CAT measurement of local brain atrophy bolic (PET) correlates ofanatomic lesions (CT/MM) in chronic schizophrenia and depression. Am J Psy produced by head trauma [Abstract]. J Nuci Med chiatry 1986; 143:175—180. 1989; 30:802. 112. Bannon MJ, Roth RH. Pharmacology of mesocorti 92. Alavi A. Functional and anatomic studies of head catneurons.Pharmaco!Rev1983;35:53—68. injury. J Neuropsychiatry 1989; l(suppl l):S45—50. 113. Seeman P. Dopamine receptors and the dopamine 93. Rao N, Turski PA, Polcyn RE, et at. ‘8Fpositron hypothesis of schizophrenia. Synapse 1987; 1:133— emission computed tomography in closed head in 152. jury.ArchPhysMedRehab 1984;65:780—785. 114. Owen F, Crow TJ, Poulter M, et at. Increased dopa 94. JohnstoneEC,CroweTJ,FrithCD,et at.Cerebral mine receptor sensitivity in schizophrenia. Lancet ventricular size and cognitive impairment in chronic 1978; ii:223—226. schizophrenia. Lancet 1976;2:924—926. 115. Smith M, Wolf AP, Brodie JD, et at. Serial (‘8F)N- 95. Kelsoe JR, Cadet JL, Pickar D, et at. Quantitative methylspiroperidol PET studies to measure changes neuroanatomy in schizophrenia. Arch Gen Psychiatry in antipsyotic drug D-2 receptor occupancy in schiz 1988;45:533—541. ophrenic patients. BiolPsychiatry 1988;23:653—663. 96. Pfefferbaum A, Zipursky RB, Lim KO, et at. Corn 116. Wong DF, Wagner HN Jr, Tube LE, et at. Positron puted tomographic evidence for generalised sulcat emission tomography reveals elevated D2 dopamine and ventricularenlargementin schizophrenia.Arch receptors in drug-naive schizophrenics. Science 1986; Gen Psychiatry 1988; 45:633—640. 234:1558—1563. 97. Anreasen NC. Brain imaging: applications in psy 117. Farde L, Wiesel FA, Hall H, et at. No D2 receptor chiatry. Science 1988;239:1381—1388. increase in PET study of schizophrenia. Arch Gen 98. Weinberger DR. Implications ofnormat brain devel Psychiatry1987;44:671—672. opment for the pathogenesis of schizophrenia. Arch 118. Seeman P, Guan HC, Niznik HB. Endogenous do Gen Psychiatry 1987;44:660—669. pamine lowers the dopamine D2 receptor density as 99. Rossi A, Stratta P, Gallucci M, et at. Brain morphol measured by (3H)raclopride: implications for posi ogy in schizophrenia by magnetic resonance imaging tronemissiontomographyof the human brain.Syn (MRI).ActaPsychiatrScand1988;77:741—745. apse 1989; 3:96—97. 100. Beaumont J, Dimond S. Brain disconnection and 119. Pearlson GD, Wong DF, Ross CA, et at. D2 dopa schizophrenia. Br J Psychiatry 1973; 123:661—662. mine receptors increased in psychotic, but not non 101. Bogerts B, Meertz E, Schonfeldt-Bausch R. Basal psychotic affective disorder. The Johns Hopkins Fifth ganglia and limbic system pathology in schizophre Annual Symposium 1989: PET imaging of brain nia. Arch Gen Psychiatry 1985; 42:784—791. chemistry. 102. Buchsbaum MS, Ingvar DH, Kesslar R, et at. Cere 120. Gusella JF, Wexler NS, Conneally PM, et at. A brat glucography with positron tomography. Arch polymorphic DNA markergeneticallylinked to Hunt Gen Psychiatry 1982; 39:25 1—259. ington'sdisease.Nature1983;306:234—238. 103. Farkas T, Wolf A, Jaeger J, et at. Regional brain 121. Kuhi DE, Phelps ME, Markham CH, et at. Cerebral glucose metabolism in chronic schizophrenia. Arch metabolism and atrophy in Huntington's disease de Gen Psychiatry 1984; 41:293—300. termined by I8@f@and computed tomographic scan. 104. VolkowND, WolfAP, VanGelderP,etat.Phenom Ann Neurol 1982; 12:425—434. enologicat correlates of metabolic activity in 18 pa 122. Lukes SA, Aminoff Mi, Crooks J, et at. Nuclear tients with chronic schizophrenia. Am J Psychiatry magnetic resonance imaging in movement disorders. 1987;144:151—158. Ann Neuro!1983;13:690—691. 105. Wolkin A, JaegarJ, BrodieJD, et at. Persistenceof 123. Mazziotta JC, Phelps ME, PaM JJ, et at. Reduced cerebral metabolic abnormalities in chronic schizo cerebral glucose metabolism in asymptomatic sub phrenia as determined by positron emission tomog jects at risk for Huntington's disease.N Eng! J Med raphy. Am JPsychiatry 1985; 142:564—571. 1987; 3 16:357—362. 106. Resnick SM, Gur RE, Alavi A, et at. Positron emis 124. Kuhl DE, Metter U, Riege WH, et at. Patterns of sion tomography and subcortical glucosemetabolism cerebral glucose utilization in Parkinson's disease and in schizophrenia. Psychiatry Res 1988; 24:1—11. Huntington's disease. Ann Neurol 1984; l5(suppl): 107. Gur RE, Resnick SM, Alavi A, et at. Regional brain @119—S125. function in schizophrenia: I. A positron emission 125. Hayden MR, Hewitt J, Stoessl AJ, et at. The com tomography study. Arch Gen Psychiatry 1987; bined useofpositron emission tomography and DNA 44:119—125. polymorphisms for preclinical detection of Hunting 108. Early TS, Reiman EM, Raichle ME, et at. Left globus ton'sdisease.Neurology1987;37:1441—1447. pallidus abnormality in never-medicated patients 126. Young AB, Penney JB, Starosta-Rubinstein S, et at. with schizophrenia. Proc Nat! Acad Sci USA 1987; Normal caudate glucose metabolism in persons at 84:561—563. risk for Huntington's disease. Arch Neurol 1987;

Volume 30 •Number 10 •October1989 1605 44:254—257. 445:61—67. 127. Hagglund J, Aquilonius S-M, Eckernas S-A, et at. 143. Perlmutter JS, Kilbourn MR, Raichle ME, et at. Dopamine receptor properties in Parkinson's disease MPTP-induced up-regulation of in vivo dopamin andHuntington'schoreaevaluatedby positronemis ergic radioligand-receptor binding in humans. Neu sion tomography using ‘‘C-N-methyl-spiperone.Acta rology 1987; 37:1575—1579. NeurolScand 1987; 75:87—94. 144. Engel J Jr, Brown WJ, KuhI DE, et at. Pathological 128. Wong DF, Links JM, Wagner HN, et a!. Dopamine findings underlying focal temporal lobe hypometab and serotonin receptors measured in vivo in Hun olism in partial epilepsy. Ann Neurol 1982; 12:518— tington's disease with c-l l-N-methylspiperone PET 528. imaging[Abstract].J NuclMed 1985;26:P107. 145. Blom Ri, Vinuela F, Fox AJ, et at. Computed to 129. Leenders KL, Frackowiak RSJ, Quinn N, et at. Brain mography in temporal lobe epilepsy. JCAT 1984; energy metabolism and dopaminergic function in 8:401—405. Huntington's disease measured in vivo using positron 146. LatackJT,Abou-KhahilBW,SiegelGJ,etat.Patients emission tomography. Movement Disorders 1986; with partial seizure:evaluation by MR, CT, and PET 1:69—77. imaging.Radiology1986;159:159—163. 130. Marsden CD. Basal ganglia disease. Lancet 1982; 147. Theodore WH, Dorwart R, Holmes M, et at. Neu ii:l141—1147. roimaging in refractory partial seizures. Comparison 13 1. Braffman BH, Grossman RI, Goldberg HI, et at. MR ofPET, CT and MM. Neurology1986;36:750—759. imaging of Parkinson's disease with spin-echo and 148. Sperling MR, Wilson G, Engel J Jr, et at. Magnetic gradient sequences. Am JRoentgeno! 1989; 152:159— resonance imaging in intractable partial epilepsy: cor 165. relative studies. Ann Neurol 1986;20:57—62. 132.DrayerBP,OlanowW, BurgerP, et at.Parkinson 149. Fobben ES, Zimmerman RA, Sperling MR, et at. plus syndrome diagnosis using high field MR imaging MR imaging in temporal lobe epilepsy [Abstract]. ofbrain iron. Radiology 1986; 159:493—498. Radiology 1988; 169(suppl):l42. 133. Rougemont D, Baron JC, Collard P, et at. Local 150. Alavi A, Sperling M, Zimmerman RA, et at. Corn cerebral glucose utilisation in treated and untreated parison of positron emission tomography and patients with Parkinson's disease.JNeurolNeurosurg magnetic resonance imaging in partial epilepsy [Ab Psychiatry 1984; 47:824—830. stract]. JNuclMed 1988; 29:910. 134. Martin WRW, Beckman JH, Catne DB, et at. Cere 151. Engel J Jr, Kuhl DE, Phelps ME, et at. Interictal brat glucose metabolism in Parkinson's disease. Can cerebral glucose metabolism in partial epilepsy and JNeurolSci1984;11:169—173. its relation to EEG changes. Ann Neurol 1982; 135. Kuhi DE, Metter El, Riege WH. Patterns of local 12:510—517. cerebral glucose utilization determined in Parkin 152. Theodore WH, Newmark ME, Sato 5, et at. son's diseaseby the @‘8F@fluorodeoxyglucosemethod. (‘8F)Fluorodeoxyglucosepositron emission tomog Ann Neurol 1984; 15:419—424. raphy in refractory complex partial seizures. Ann 136. LeendersKL, Palmer AJ, Quinn N, et at. Brain Neurol1983;14:429—437. dopamine metabolism in patients with Parkinson's 153. Engel J Jr, Kuhl DE, Phelps ME, et at. Comparative disease measured with positron emission tomogra localization of epileptic foci in partial epilepsy by phy. J Neurol Neurosurg Psychiatry 1986; 49:853— PCT and EEG. Ann Neuro!1982;12:529—537. 860. 154. Abou-Khalil BW, Siegel GJ, Sackellares JC, et at. 137. Garnett ES, Nahmias C, Firnau G. Central dopamin Positron emission tomography studies of cerebra! ergic pathways in hemiparkinsonism examined by glucose metabolism in chronic partial epilepsy. Ann positron emission tomography. Can J Neuro! Sci Neurol1987;22:480—486. 1984;11:174—179. 155. EngelJ Jr. The useofpositron emission tomographic 138. TedrroffJ, Aquilonius S-M, Hartvig P, et at. Mono scanning in epilepsy. Ann Neurol 1984; l5(suppl): amine re-uptakesites in the human brainevaluated @180—@191. by means of ‘‘C-nomifensineand positron emission 156. Gur RC, Sussman NM, Alavi A, et a!. Positron tomography: the effects of age and Parkinson's dis emission tomography in two cases of childhood epi ease.ActaNeurolScand1988;77:192—201. leptic encephalopathy (Lennox-Gestaut syndrome). 139. Wijnand A, Rutgers F, Lakke JPW, et at. Tracing of Neurology1982;32:1191—1194. dopaminereceptorsin hemiparkinsonismwith posi 157. EngelJ Jr, KuhI DE, PhelpsME. Patternsof human tron emission tomography (PET). J Neuro!Sci 1987; local glucose metabolism during epileptic seizures. 80:237—248. Science 1982; 218:64—66. 140. Chiueh CC, Firnau G, Burns RS, et at. Determina 158. Bernardi 5, Trimble MR, Frackowick RS, et at. An tion and visualization of damage to striatal dopa interictal study of partial epilepsy using positron minergic terminals in l-methyl-4-phenyl-1,2,3,6-tet emission tomography and the 0-15 inhalation tech rahydropyridine-inducedparkinsomsm by [‘8F]-la nique. J Neurol Neurosurg Psychiatry 1983; 46:473— beled 6-fluoro-L-dopaand positron emission tomog 477. raphy.Adv Neuro!1986;45:167—169. 159. Franck G, Sadzot B, Salmon E, et at. Regional cere 141. Catne DB, Langston JW, Wayne WR, et at. Positron brat blood flow and metabolic rates in human focal emission tomography after MPTP: observations re epilepsy and status epilepticus. In: Delgrado-Escueta lating to the cause of Parkinson's disease. Nature AV,WardAAJr,WoodburyDM et at.,eds.Advances 1985;317:246—248. in neurology. Vol 44. New York: Raven Press, 142. Leenders KL, Aquilonius S-M, Bergstrom K, et a!. 1986:935—948. Unilateral MPTP lesion in a rhesus monkey: effects 160. Engel J Jr, Lubens P, Kuhl D, et at. Local cerebral on thestriataldopaminergicsystemmeasuredin vivo metabolic rate for glucose during petit mat absences. with PET usingvarious noveltracers. Brain Res 1988; Ann Neurol1985;17:121—128.

1606 Jolles,Chapman,Alavietal TheJournalof NuclearMedicine