Imagingthe head: functional imaging · tomography (PET), single photon emission computed tomography...

13ournalofNeurology, Neurosurgery, and Psychiatry 1995;58:132-144

NEUROLOGICAL INVESTIGATIONS

Imaging the head: functional imaging

Guy V Sawle

If asked to choose between a brain thatlooked nice, or one that functioned well, mostof us would choose the second. Furthermore,the ultimate importance of cerebral disease isthat it affects brain function, not appearance.Yet clinical neuroimaging has been builtaround the practice of imaging brain struc-ture. Why so? Because structural images are

easier to acquire, and structure and functionare so inextricably linked that one might as

well image one as the other. But is this neces-

sarily so? To what extent are structural andfunctional imaging processes independent ofone another? And are there any clinical situa-tions where structural imaging "just won'tdo"?

In this article I discuss several approachesto cerebral imaging in which the principal aimis to derive information about brain function.Specifically, I discuss positron emissiontomography (PET), single photon emissioncomputed tomography (SPECT), and func-tional magnetic resonance imaging (fMRI).Other functional imaging methods, such as

blood flow measurements with 133Xe-enhanced CT' have been of considerable his-torical importance but are now seldom usedin clinical or research practice and will not becovered further.

Principles ofthe techniquesPOSITRON EMISSION TOMOGRAPHY (PET)2In PET short lived isotopes are used to labelmolecules of biological interest. After inhala-tion or injection, they decay by positron emis-sion, each positron becoming annihilatedwithin 1-2 mm of its parent nucleus by colli-sion with an electron. This annihilation gen-erates two y rays (of 511 keV energy) thattravel apart at 180° to one another. It is. thenearly simultaneous detection of these y ray

pairs by a ring of detector crystals that ulti-mately leads to the reconstructed image of

Division of ClinicalNeurology, QueensMedical Centre,Nottingham, NG72UH, UKG V Sawle

isotope density. The theoretical limit of spa-

tial resolution is the distance that positronstravel from their parent nucleus before anni-hilation. The actual spatial resolutiondepends in part on the size of detector crys-

tals used in the camera. After positron annihi-lation deep within the brain, a percentage ofthe emitted photons fail to reach the detectorcrystals due to signal attenuation by brain tis-sue. In PET it is possible to correct for thisloss using a second set of image data col-lected before isotope injection or inhalation.For this transmission scan an external (ringor moving rod) germanium-68 source is used.Current generation PET machines have a res-

olution of around 5 mm in the reconstructedimage. Commonly employed PET isotopesinclude oxygen-i5, carbon-i 1, and fluorine-18, used to replace the naturally occurringoxygen- 16, carbon- 12, and hydrogen-1respectively in various biological molecules.This exchange of a radioactive atom for a nat-urally occurring atom results in little (if any)change in chemical behaviour. The half livesvary from two minutes (oxygen-15) to 110minutes (fluorine-18). Such short half liveshave both advantages (less radiation dose)and disadvantages (cost, dependence on an

on site cyclotron for production, and the needfor a tight time schedule). Measurements byPET take minutes to hours, depending on theparticular brain function under scrutiny.Radiation considerations preclude frequentrepeat measurements. The table gives an

overview of some of the strengths and limita-tions of the PET method, together with com-

parative data for SPECT and fMRI.

SINGLE PHOTON EMISSION COMPUTED

TOMOGRAPHY (SPECT)3The isotopes used in SPECT (such as tech-netium-99 or iodine-123) have longer halflives, obviating the need for on site produc-tion. This reduces cost and eases some of the

Strengths and limitations of imaging methods

PET SPECT fMRI

Isotopes Fluorine-1 8, carbon- 1 Technetium-99m Noneoxygen- 15 iodine- 123

Time per image Two minutes to two hours Minutes to hours 0-01 Seconds to a few minutesSpatial resolution About 5-6 mm About 8 mm 0 75 mm upwardsRepeated studies Very few Very few Yes

(limited by radiation) (limited by radiation)Able to study Metabolism, blood flow, Blood flow, some Blood flow/venous drainage

receptor-ligand receptor-ligandinteractions interactions

Useful technical references Phelps2 Lassen and Holm' Lufkin4; Stehling et all

132

on July 24, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.2.132 on 1 February 1995. D

ownloaded from

http://jnnp.bmj.com/


timing restraints of PET. On the other hand,because they emit single y rays (hence singlephoton ECT), coincidence detection cannotbe used to yield spatial information, whichmust, therefore, depend on collimation alone.Furthermore, signal attenuation by surround-ing tissues cannot be corrected by an exactsolution with transmission scan data as usedin PET. Partly for the foregoing reasons,SPECT has a lower spatial resolution thanPET. Nevertheless, because of the lower costof SPECT and the greater availability ofmachines, SPECT has found an altogetherlarger place in the clinical arena than PET.Measurements by SPECT take minutes tohours. As with PET, radiation considerationspreclude frequent repeat measurements.

FUNCTIONAL MAGNETIC RESONANCE IMAGING(FMRI)In magnetic resonance methods,4 the dividebetween structural and functional imaging isprecarious. For example, is magnetic reso-nance angiography (MRA) a functional or ananatomical measurement-because it showsthe anatomy of the major cerebral vesselsusing a sequence that is specifically sensitiveto the movement of the contained blood? Inthis review I use the term functional magneticresonance imaging (fMRI) to refer to a rangeof MR sequences designed to acquire infor-mation about brain function. These tech-niques are newer than either PET or SPECTand the scientific literature concerned withtheir use has so far been more methodologicalthan medical. Nevertheless, at least two fMRIapproaches deserve mention-namely, echo-planar imaging (EPI), and fast low angle shot(FLASH) techniques. Each is concerned withthe generation of images where a change insignal with time is most likely the conse-quence of changing neuronal function, themediator between the two being a change insmall vessel flow, or at least an increase inlocalised venous return. The present possibil-ities and imminent potentials of fMRI havebeen described in several recent reviews.5-7

Whereas the basics of PET and SPECTmethodology follow parallel processes inother techniques such as photography andx ray computerised tomography, MRI hasno easy parallel in our other experiences. Thefundamentals of magnetic resonance havebeen covered by Moseley in this Journal(1995;58:7-21); see also Luflkin4). Conven-tional MR sequences build up an image insteps, using a series of magnetic field gradi-ents to specify anatomical positions withinthe tissue of interest. In EPI an image isrecovered from the signal generated by a sin-gle free induction decay over a fraction of asecond.8 On the other hand, FLASHsequences limit the time taken for scanningby minimising the perturbation of the mag-netisation from its equilibrium so that succes-sive excitation pulses can follow each othermore quickly. Reported data from EPI athigh (3 Tesla) field strength include imagesacquired in 0-1 s with a spatial resolution of075 mm.9 Despite the large number of MRI

machines available for diagnostic purposes,few have either EPI or high (2-3 Tesla) fieldstrength. It is possible to acquire FILASHfMRI images at a lower field strength (1-5Tesla) with a "clinical" MR imaging system;although signal acquisition is in this casesomewhat slower.'0 All fMRI methods shareone advantage over PET and SPECT-namely, the avoidance of ionising radiation.

Practicalities of the techniquesBecause functional imaging studies have yetto enjoy wide use as clinical tools in routineneurological practice, the following accountdescribes elements of the basic principles, aswell as the nuts and bolts practicalities ofpatient scanning.

PETGeneral principlesThe general principle of the PET measure-ment requires a mathematical model that cor-responds to the functional system underscrutiny. This model is an approximation ofthe processes that lie between the "input"(the activity given to the patient and availableto the brain via its arterial supply or byinhalation) and the "output" (the activitymeasured regionally during the course of theexperiment).

Cerebral bloodflowOne of the simplest PET models relates tothe measurement of regional cerebral bloodflow during continuous inhalation of C'50,.After a few minutes inhalation, an equilib-rium is reached whereby the arterial supply ofradioactivity to the brain is equal to the lossof activity from venous washout and radioac-tive decay (the so called steady state condi-tion). In this situation a simple mathematicalexpression describes the regional cerebralblood flow in terms of known or measurablevalues-namely, the radioactive decay con-stant (fixed for 150), the arterial activity ofH,150 (in pCi/ml, measured in an arterialblood sample with a well counter), and theregional brain concentration of tracer inunits/ml (measured in the PET camera). Thesteady state measurement of regional cerebralblood flow takes about 15 minutes to com-plete.On the practical side, these measurements

(and all of those listed below) require that thepatient be still during image acquisition.They require a venous line for tracer adminis-tration (except when C'50", C'50, or 1502 aregiven by inhalation). Most quantitative stud-ies also require an arterial line to measure thelevel of radioactivity presented to the brainover the time course of the scan.

Blood flow can also be measured duringthe rise and fall of brain radioactivity sur-rounding a bolus inhalation or injection oftracer. The mathematical model required tounscramble the collected data to a measuredvalue for rCBF is in this case very much morecomplicated," but the method is faster (dataacquisition takes only two to three minutes).

133


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Sawle

Oxygen metabolismOxygen metabolism can be calculated fromcerebral blood flow after additional measure-ments of the oxygen extraction fraction (thepercentage of the available blood oxygenextracted during its passage through the brainvasculature; usually measured after inhalationof 1502) and regional blood volume (a correc-tion for the percentage of any cerebral regionthat contains blood rather than brain).'2 Sucha (triple) measurement with steady statemodels takes about 40 minutes (includingtime for radioactive decay between scans).

Glucose metabolismGlucose metabolism is measured after intra-venous injection of 2- ['8F]fluoro-2-deoxy-D-glucose (FDG), which is metabolised byhexokinase to FDG-6-phosphate. As FDG-6-phosphate can neither proceed down theglycolytic pathway nor be metabolised toglycogen (the metabolic destinies for glucose-6-phosphate) it stays trapped within thetissue for the duration of the PET measure-ment. This compartmental trapping is thecornerstone of the FDG measurement.Further interpretation may be fully quantita-tive (requiring continuous arterial andregional cerebral radioactivity measurements)or semiquantitative (using normative datafrom other subjects, and a further series ofconstants and restraints in the employedmathematical model). On the practical sidethis means that for the most accurate mea-surement, a patient must be in the cameraduring tracer injection and for the next 60 to90 minutes, with simultaneous arterial bloodsampling. For a less quantitative scan (fromwhich regional inequalities in glucose metab-olism can yet be semiquantified) the tracercan be injected out of the camera and a"snapshot" image (lasting perhaps 15 minutes)can be taken about 30 to 40 minutes later.

Neurotransmitter precursor studiesAside from blood flow and measures of tissuemetabolism, the other principal application ofPET to date has been in neuropharmacologi-cal studies; PET has a small repertoire for thestudy of neurotransmitter synthesis and stor-age, principally the decarboxylation of[18F]dopa to ['8F]dopamine. Like the FDGmethod, the premise on which most of the['8F]dopa analytical methods are based is theassumption that the injected tracer(['8F]dopa) is transported into the brain andthen specifically taken up by dopamine neu-rons where it is decarboxylated, concentrated,and then stored in nerve terminal vesicles forthe duration of the measurement. A particu-lar disadvantage of ['8F]dopa as a tracer ofthe dopamine synthetic pathway is that theconcentration of the endogenous (dopa) poolis unknown. So if anybody ever discovers theperfect mathematical model to unravel['8F]dopa scan data (an endeavour that hasattracted much energy, even some disagree-ment) they will still only have measured therate of metabolism of exogenous ['8F]dopa;the rate of endogenous dopamine production

cannot be deduced without a knowledge ofthe endogenous dopa pool-which you can'tmeasure! Despite these caveats, ['8F]dopa hasbeen an excellent work horse in the PETarmamentarium.

Neurotransmitter receptor studiesTracers of much larger variety have beenemployed as markers for particular classesof neurotransmitter receptor, includingdopamine Dl (["C]SCH23390"3) and D2(["C] raclopride'4) postsynaptic receptors,dopamine reuptake sites ([11C]nomifensine'5 16

and ["C]WIN-35,428'7) opiate receptors(mu, (["C]carfentanil'8), mu and kappa(['8F]cyclofoxy'9), mu, kappa, and delta(["C]diprenorphine20)), central (["C]flu-mazenil2') and peripheral (["C]PK111952223)benzodiazepine receptors, muscarinic cholin-ergic receptors (["C]scopolamine24), hista-mine Hi receptors (["C]pyrilamine25), andMAO-B activity (["1C]deprenyl26). These lig-ands have been used (some extensively) tostudy the changes in receptor numbers oraffinity in some disease states, includingParkinson's and other akinetic rigid syn-dromes, Huntington's disease, epilepsy, pain,and stroke. Quantitative measurementsrequire continuous measurement of blood andbrain activity during and after injection oftracer. A single PET scan may be sufficientfor semiquantification, but if a full descriptionin neuropharmacological terms is required (tocalculate, for example, Bmax, the total concen-tration of binding sites) repeat studies may benecessary in the same subject with injectionsof tracer having different specific activities, orcoinjection of unlabelled tracer.

Functional mappingAside from neurotransmitter studies, theother growth area in recent years has been thedevelopment of functional mapping studies inwhich repeat measurements of blood flow areused as a means of identifying brain regionsactive in particular cognitive or other pre-scribed tasks. If a subject is scanned twice,once at rest and the second time during rightarm movement, it is argued that any differ-ence between the two images of blood flowmay be accounted for either by noise, by arte-fact, by a general (global) effect of the activityon cerebral blood flow, or by a specific activa-tion of a responsible brain region. Variousmethods have been developed to extract thespecific information by removing the con-founding effects.2728 In large part the pub-lished base of work in this area has employedbetween subject averaging to improve the sig-nal to noise characteristics of the method.Newer scanners with MRI coregistrationallow more confident results in individualsubjects with greater accuracy in the anatomi-cal loci of activation related change.29 The useof functional mapping studies have not so farbeen reported as a clinical procedure.

SPECTGeneral principlesThe practicalities of SPECT measurements

134


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



are for the most part simpler than for PET. Inpart this is because SPECT data are bynecessity less quantitative than PET data.Another relevant factor may be the closer liai-son and relation between SPECT and clinicalmedicine, with the restraining hand of clinicalpracticality curbing the imager's urge to addcomplexity in the pursuit of accuracy.

Cerebral bloodflowThe SPECT approach to cerebral bloodflow hinges on the finding that certain tracersare irreversibly taken up into the brain ina regional pattern that reflects localiseddifferences in cerebral blood flow. After intra-venous injection, [1231] -n-isopropyl ampheta-mine (['23I]-iodoamphetamine)30 crosses theblood-brain barrier by passive diffusion with ahigh first pass extraction. It is then retained inthe brain by non-specific binding to aminereceptors. So signal intensity in a "snapshot"image taken 20 to 60 minutes after tracerinjection is proportional to the perfusiondominated distribution of tracer in thebrain.

Likewise, [99mTc]hexamethylpropylenea-mine oxime ([99mTc]HMPAO) crosses theblood-brain barrier easily by passive diffusion.It is then trapped in the brain (after decom-position to a byproduct that cannot pass backacross the barrier), uptake and trapping beingcomplete within 10 minutes. Images taken 90to 120 minutes after injection (image acquisi-tion typically taking about 20 minutes)still show the frozen image of regional cere-bral blood flow at the time of tracer adminis-tration. [99mTc]HMPAO is preferred to[1231]-iodoamphetamine on several accounts,including its optimum imaging energy andshorter half life. Unlike PET blood flow mea-surements, which require an on sitecyclotron, [99mTc]HMAAO can be producedin a hospital nuclear medicine departnentwith a molybdenum-99 generator. It must beused within about 30 minutes of productionto avoid decomposition before injection.[99mTc] labelled N,N -1,2-ethylene-diylbis-L-cysteine diethyl ester dihydrochloride([99mTc]ECD) is similar to but more stablethan [99mTc]HMPAO.3

Patients undergoing SPECT measurementof cerebral blood flow with these techniquesdo not need to be in the camera during tracerinjection. They should, however, be rested atthis time because it is blood flow around thetime of injection that is measured during thelater scan, not blood flow at the time of themeasurement, as in PET.

For the most part, the interpretation ofSPECT flow images follows the radiologicaltradition-namely, interpretation of theimage appearance by an expert in the field.As with the assessment of age related atrophyon structural images, the observer must takeinto account the known changes in cerebralblood flow that accompany the normal ageingprocess. Such images are often reportedalongside structural images to help in the dif-ferentiation between normal and pathologicalappearances.

Neurotransmitter receptor studiesSeveral neurotransmitter systems have beenstudied with SPECT. Specifically thedopamine D2 ligand ['23I]-(S-)-2-hydroxy-3-iodo-6-methoxy-N[( 1 -ethyl-2-pyrrolidinyl)methyl]-benzamide (['23I]-iodobenzamide)has been extensively used in neurological dis-orders. This ligand binds reversibly todopamine D2 receptors. The amount of spe-cific striatal binding increases over about 40minutes and then remains stable for up totwo hours. Typically data are acquired duringthe period 60 to 120 minutes after tracerinjection, image acquisition taking about 50minutes. Because of the limitations of mea-surement, SPECT ["23I]-iodoamphetaminedata are reported as a specific: non-specificratio, such as striatum:cerebellar counts.

fMRIGeneral principlesProcedures for fMRI are very different fromeither PET or SPECT, being independent ofionising radiation. Aside from any activitythat the patient might be asked to performwhile in the magnet, the patient's experienceof fMRI is unlikely to differ greatly from anyother MR procedure (loud noises in a darktunnel), although EPI imaging is presentlyparticularly noisy. Although rapid MRsequences (such as EPI) yield clear images ofmoving structures (such as a beating heart ora waving head) fMRI methods rely on a com-parison of successive scans of the same area.In this case, the head position must be identi-cal for the acquisition of each of the imagescontributing to data analysis. Even the tiniesthead movements can wreck havoc with fMRIanalysis; indeed it has even been possible tocreate striking "functional" data as a result ofhead movement artefact alone.'2 As withPET, one approach to the problem of headmovement between scans may be to realignthe data in software after image acquisition.3'4

What has been learned with thesetechniques, and to what extent may theybe used in clinical practice?Both PET and SPECT have been used inneuroscience research to examine brain func-tion in health and in disease. Thus far fMRIhas been most closely applied to the study ofhealthy subjects, although this will certainlychange. This review will concentrate on thosestudies designed primarily to answer ques-tions about disease. When applicable, men-tion will be made of clinical situations inwhich functional imaging could provide valu-able additional information. The UnitedKingdom has a single PET research institu-tion (the Medical Research CouncilCyclotron Unit at the HammersmithHospital) and a single dedicated clinical PETfacility (at the St Thomas' and Guy'sHospitals). The first has a broad range ofchemistry facilities, whereas the available lig-ands in the clinical PET centre are fewer.This pattern is generally true elsewhere-most PET centres with clinical services

135


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


136

chiefly offer [18F]fluorodeoxyglucose scans;other more closely research based units typi-cally have a more extended repertoire. In theUnited States there is now an Institute forClinical PET. Furthermore, some UnitedStates health insurance schemes haveapproved a variety of clinical PET proceduresfor reimbursement. Presently, SPECT scan-ners are more widely distributed, both in theUnited Kingdom and elsewhere. Althoughmany hospital radiology departments areequipped with magnetic resonance machines,few have the hardware and software on handfor the acquisition of an fMRI signal.

CEREBROVASCULAR DISEASEEarly PET studies measured regional cerebralblood flow, blood volume, oxygen extraction,and oxygen metabolism to examine thepathophysiology of stroke, particularly themechanisms of cerebrovascular compensationin the face of falling and failing arterialperfusion pressure.3536 Currently PET andSPECT can both detect cerebral ischaemiain acute stroke at a stage when CT imagesare still normal. It has been shown that PETalso has some ability to predict the extentof functional recovery from stroke37 and inrecovered patients (using functional map-ping) it can show the anatomical andfunctional substrate of recovered function.38 39Likewise PET and SPECT can showevidence of hypoperfusion ("misery perfu-sion") in the absence of infarction,40 andhyperperfusion ("luxury perfusion") at a siteof previous infarction. Haemodynamicchanges can be shown by PET in patientsafter extracranial-intracranial bypass opera-tions.4' Although this operation has not beenshown to be of benefit in large interventionalstudies42 it is possible that preoperativefunctional imaging could be used to

identify patients more likely to gain fromoperation.

Basic research in cerebrovascular diseasehas given cause for guarded optimism overthe possible use of cerebral protective agentssuch as glutamate antagonists and free radicalscavengers; PET is waiting in the wings to beused to increase the pathophysiological andtherapeutic gains from patient trials withthese agents. Whether it will be called onremains to be seen.

CLINICAL INDICATIONSThere are presently no consensus indicationsfor the clinical use of functional imaging incerebrovascular disease (but then there are noclinically proved treatments for acute strokeeither43; both could change).

DEMENTIA

Early PET studies showed regional metabolicchanges in Alzheimer's disease and someother degenerative conditions, and these find-ings led to the notion that particular diseasesmight be recognisable by specific regionalchanges in cerebral function. As in the studyof several other disease states, many of theearly reports of functional imaging inAlzheimer's disease included small numbersof patients and employed loose diagnostic cri-teria. Results were sometimes contradictory.Consensus has now been reached in a num-ber of areas as follows. In Alzheimer's dis-ease, the brunt of the early PET changes arecentred around the posterior temporal andparietal cortices (fig 1).44 47 Regional betweenpatient differences may correlate with differ-ences in neuropsychological test scores.4548 49Changes on PET may antedate clinicaldementia in patients presenting with mildmemory deficits.50 Studies with SPECT inAlzheimer's disease have also shown reduced

Figure 1 PET image of ['8F]fluorodeoxyglucose metabolism in Alzheimer's disease. Note reduced tracer uptake inposterior temporal and parietal cortex. (Picture courtesy ofDrA Kennedy.)

Sawle


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



flow in posterior temporal and other corticalregions.5' In patients with familialAlzheimer's disease ['8F]fluorodeoxyglucosePET in affected family members shows thesame pattern of parietotemporal hypometabo-lism. Scans in asymptomatic at risk relativesshow a similar (but less severe) pattern.54Most PET scans in patients with Pick's

disease established by necropsy or biopsyhave shown predominantly frontal hypome-tabolism.5556 This finding is not specific toPick's disease, however, as it has also beenreported in progressive supranuclear palsy57-59and SPECT studies have shown a reductionin frontal [99mTc]HMPAO uptake in patientspresenting with dementia of the frontal lobetype.60

Patients with focal cortical degenerationspresenting with slowly progressive apraxia oraphasia have been studied with PET and havebeen shown to have appropriate areas of cor-tical hypometabolism at a stage when struc-tural imaging studies have been normal.6162The pathology in these patients turns out tobe variable. Much has been written about theidentification of clinical and preclinicalchanges in PET metabolic indices inHuntington's disease.63-6 Both striatal andcortical hypometabolism have been reported.After the demonstration of low caudate[18F]fluorodeoxyglucose metabolism in someat risk patients, considerable efforts weredirected towards the development of PET asa preclinical disease marker.6566 Althoughgenetic testing now provides a generally reli-able means of making a positive diagnosis ofHuntington's disease based on a blood sam-ple alone,6768 it may be that PET still has apart to play in these patients. If, for example,neurotransplantation procedures become apractical proposition in this disorder, it maybe that PET will provide a crucial means ofidentifying an appropriate preclinical or early

clinical stage of disease for intervention. Genepositive at risk patients have lower striatal andpallidal volumes (a structural MR measure-ment) than gene negative at risk patients69; ithas yet to be shown that either MRI or PETcan indicate when an at risk subject willdevelop clinical problems. Neurotransmitterstudies in the foregoing conditions are dis-cussed in the next section.

CLINICAL INDICATIONSAs with cerebrovascular disease, there arepresently few if any treatment options thatcould reasonably be said to depend on diag-nostic information that could only be gleanedfrom fimctional imaging studies. Both PETand SPECT may assist in the accurate diag-nosis of these conditions presenting as adementing illness; but such data cannot yet beregarded as mandatory in good patient care.

MOVEMENT DISORDERSIn Parkinson's disease, early blood flow andmetabolic studies7072 were soon upstaged bythe demonstration of reduced striatal uptakeof [18F]-6-L-fluorodopa (['8F]dopa) inaffected patients.73 Many [18F]dopa studieshave now been reported in this disorder, con-sidering issues such as the role of ageing(most centres,74 75 but not all,76 have shown noeffect of age on ['8F]dopa uptake), the detec-tion of presymptomatic disease (fig 2), therate of progression of clinically evident dis-ease,78-80 and the efficacy of neurotransplanta-tion procedures (fig 3).8182

Various other akinetic rigid conditionshave been studied by [18F]dopa PET, includ-ing multiple system atrophy,83-85 progressivesupranuclear palsy,578384 corticobasal degen-eration,86 neuroacanthocytosis,87 and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)parkinsonism. 82 88 Some of these disordershave been reported to show characteristic

Figure 2 PET image to show /'8F] dopa uptake in a normal subject (left), a patient with idiopathic (sporadic)Parkinson's disease, and two members of a sibship with familial parkinsonism. The symptomatic patient shows profoundlyimpairedfluorodopa uptake whereas the presymptomatic subject (who became clinically affected within months of thescan) shows fluorodopa uptake at a level internediate between normal and parkinsonian values.

137


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Sawle

Figure 3 Combined MRI(Tl weighted) and PET(P78F] dopa) image from aparkinsonian patient whohas undergoneimplantation offetalmaterial into the putamen.The MRI image shows thesite of the burr hole andthree needle tracks reachingdown through the cortexand subcortical whitematter into the putamen.The /'8F] dopa imageshows uptake at the graftszte.

patterns of striatal ['8F]dopa uptake, such assevere asymmetric loss of caudate and puta-men uptake (in corticobasal degeneration86)or severe bilateral early loss of both caudateand putamen signal (in progressive supranu-clear palsy84). The difficulty in translatingthese patterns from research to clinical prac-tice is that these studies have of course beenundertaken in patients who carry a (fairly)confident clinical diagnosis. Even in suchpatients, if we move from a group to individ-ual subjects and study their [18F]dopa PETdata, it may be impossible to ascribe a partic-ular diagnosis (for example Parkinson's dis-ease v multiple system atrophy) with absolutecertainty.89

['8F]Dopa studies have also been per-formed in possibly less obvious disorders. Ithas been shown that the extrapyramidalsymptoms in clinically diagnosed Alzheimer'sdisease seem not to be due to nigral degener-ation.90 The motor disorders in obsessionalslowness9' and manganese toxicity92 are like-wise unaccompanied by changes in [18F]dopauptake, whereas in patients poisoned bycyanide ['8F]dopa uptake is reduced, suggest-ing (direct or hypoxic induced) nigral toxic-ity.93 Patients with parkinsonism resultingfrom neuroleptic or other dopamine blockingdrugs may have either normal ['8F]dopauptake (suggesting a likely return to clinicalnormality after cessation of the offendingagent) or low uptake (suggesting unmaskingof otherwise subclinical parkinsonism.94

[18F]dopa is not the only tracer to provideinformation about the presynaptic dopaminesystem. ["C]Nomifensine has also been used(as a marker of catecholaminergic presynapticreuptake sites) but in most cases the results of[11C]nomifensine studies have closely paral-leled those with ['8F]dopa.1695 Cocaine ana-

logues such as ['IC]CFT (also known as WIN35 428) have also been used to study thedopamine fibre system. In a primate model ofparkinsonism, [1"C]CFT uptake was reducedin the striatum.96 This ligand has the advan-tage of a substantially higher striatal to back-ground signal than ['8F]dopa, but the kineticsare, if anything, a little slow for use as a PETtracer. The related compound, CIT, has beenused in SPECT as [123I]f-CIT. Striataluptake of this tracer was reduced inParkinson's disease.97

In part, the answer to the problem of dif-ferential diagnosis by PET in individualpatients might be helped by multiple tracerstudies to examine receptor status as well as(or in place of) ['8F]dopa uptake. There aremany publications concerning dopamine D2receptors in akinetic-rigid syndromes. On bal-ance PET and SPECT studies suggest rela-tive upregulation of D2 receptors in patientswith early Parkinson's disease, with normal oreven lower levels later in disease; perhaps inpart as an effect of treatment with dopamin-ergic drugs (PET148798 101; SPECT'02>'04).Patients with multiple system atrophy aremore likely to have low D2 ligand binding(PET'4 84) and low tracer uptake in dopanaive akinetic-rigid patients may predict sub-sequent evolution to multiple system atrophyrather than Parkinson's disease (PET'4;SPECT'05) (fig 4). Unlike the D2 system,PET studies of dopamine Dl receptors haveshown no evidence of up regulation in earlylevodopa naive patients.'3 Other neurotrans-mitter receptors have also been studied in aki-netic-rigid syndromes. Specifically, striatal["lC]diprenorphine binding has been shownto be impaired in patients with multiple sys-tem atrophy, but not in patients withParkinson's disease. 106 In a study using

138 on July 24, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.58.2.132 on 1 F

ebruary 1995. Dow

nloaded from


139Imaging the head: functional imaging

Figure 4 SPECTf23I]iodobenzamide binding ina patient with Parkinson'sdisease (left) and a patientwith a non-dopa responsiveparkinsonian syndrome(right). (Picture courtesyofDrs K Tatsch and JSchwarz, Department ofNuclear Medicine andNeurology, University ofMunich.)

["C]flumazenil as tracer to measure cerebel-lar GABA-A/benzodiazepine receptors,increased tracer binding was reported inpatients with multiple system atrophy, where-as patients with sporadic and dominantlyinherited olivopontocerebellar atrophy hadincreased and unchanged binding respec-

tively.'07Although neurotransmitter studies occupy

most of the literature on functional imagingin akinetic-rigid syndromes, specific focusedmetabolic studies and more recent functionalmapping papers are starting to redress thebalance. In depressed parkinsonian patients,for example, a particular pattern of frontalhypometabolism has been found.'08 Also, withmore sophisticated statistical techniques,some correlations have been reportedbetween regional metabolic changes, motorasymmetries, and fluorodopa uptake con-

stants.109 Functional mapping studies haveshown failure of supplementary motor cortexactivation during internally generated move-

ments in parkinsonian patients"0 with resolu-tion towards the normal after dopamineagonist treatment (shown both with PET"'and SPECT"'2). More recently, fMRI studieshave begun to consider this same issue-namely, the identification of cortical areas

responsible for the control of human move-

ment."3 114 Reported fMRI findings in thisarea include a positive relation betweenmovement rate and the fMRI signal in theprimary motor cortex,"14 and a greater signalintensity in supplementary motor areas dur-ing complex self paced movements than dur-ing externally paced movements."13

Although most PET studies inHuntington's disease have used ['8F]fluo-rodeoxyglucose as the tracer, more recentattention has focused on the dopamine sys-

tem. Dopamine D1 and D2 receptors haveeach been studied in affected patients (with["1C]SCH23390 and ["C]raclopride as trac-ers). Affected patients show a reduction ofboth Dl and D2 binding potentials (fig 5).115With ["C]raclopride PET, asymptomaticgene positive subjects may show an interme-diate reduction in binding potential (RWeeks, personal communication).

At the time of writing, fMRI studies haveyet to be reported in patients with Parkin-son's disease or other movement disorders.

CLINICAL INDICATIONSWhat of the clinical utility of functional imag-ing in movement disorders? Despite the sci-entific findings discussed, no clear necessityfor such imaging studies has yet been shown.In part, this relates to the lack of availabletreatment for many of these disorders. In aki-netic-rigid syndromes for example, some, butnot all, patients respond to levodopa ordopamine agonists. Although there is littleeffective treatment for those patients who areresistant to such treatment, informed trial anderror seems as good a therapeutic approach asfunctional imaging. A special exception tothis rule might be argued for dopa responsivedystonia. Patients with this disorder gain longterm benefit from levodopa without develop-ing the side effects and complications thatbedevil patients with Parkinson's disease,"16particularly those with onset early in life."17Fluorodopa uptake in dopa responsive dysto-nia is close to normal,"8 1"9 but it is pro-foundly impaired in patients with young onset(age 20-40) or juvenile onset (before age 20)Parkinson's disease. 120 A highly abnormal['8F]dopa PET scan in a young patient withan akinetic-rigid syndrome would thereforeimply a likelihood of early problems with


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Sawle

Figure 5 PET [IIC] raclopride images from a normal subject (left) and a patient with Huntington's disease (right). Noteprofoundly reduced tracer uptake in the striatum of the patient with Huntington's disease. (Picture courtesy ofDr R A Weeks.)

levodopa treatment, whereas early treatmentwith levodopa would be appropriate in apatient in whom [18F]dopa PET showeduptake close to or in the lower normal range.

Patients undergoing experimental neuro-transplantation procedures may also gainsome direct personal benefit (graft site selec-tion, for example)81 but otherwise for now,the principal promise of functional imaging inmovement disorders is in the advancement ofour understanding of disease, causation, andtreatment.

Figure 6 SPECTf'99,Tc] HMPAO images from a fouryear old child withpartial motor seizures. The interictal image shows hypometabolism in the lefthemisphere. The ictal image shows an area of hypermetabolism correspondingto the seizure focus. (Picture courtesy ofDr H Cross.)

EPILEPSYIn patients with focal epilepsy, functionalimaging studies (PET and SPECT) haveshown evidence of interictal focal hypometab-olism (fig 6).121123 In some cases it has beenpossible to scan patients during seizures, inwhich case areas that are interictallyhypometabolic may become ictally hyperme-tabolic or show high flow.122 124 125 It has beenreasonably argued that such regions representepileptic foci, even in the absence of corrobo-rative findings from structural imaging orEEG. In patients with intractable epilepsy,surgical excision of a definite seizure focusmay radically improve clinical status. CurrentMRI techniques are able to identify structuralabnormalities in an increasing number ofsuch patients. There are, nevertheless, a sig-nificant number of patients in whom non-invasive means fail to clearly identify a seizurefocus. Options in these patients include theplacement of depth electrodes and functionalimaging studies.

In a comparative study of [18F]fluo-rodeoxyglucose PET and [99mTc]HMPAOSPECT in patients undergoing investigationbefore surgery for temporal lobe epilepsy, dif-ferent sensitivities were reported for the twotechniques. In patients who had a normalMRI, PET with ['8F]fluorodeoxyglucoseshowed focal hypometabolism in 80% v 20% forSPECT with [99mTc]HMPAO. 126 The authorsattributed this difference to the greater spatialresolution of the PET technique. AlthoughPET is a potentially quantitative technique, ithas been argued that for clinical epileptologypurposes, image inspection by experiencedeyes is generally sufficient.127

In another study of patients withintractable epilepsy, SPECT detected lateral-ising abnormalities in 19 of 30 patients; onlytwo further lateralised abnormalities were



j.com/

J Neurol N


ebruary 1995. Dow

nloaded from



found with CT or MRI.128 As in many otherareas of imaging the ground is shiftingrapidly. With increasing structural resolutionin MRI (including hippocampal volume mea-surements) the balance is swinging in favourof MRI having a greater chance of correct lat-eralisation than SPECT.69A recent fMRI study of a four year old boy

showed changes in image signal restricted toan area of structural abnormality during fiveseizures over a 25 minute period. InterictalSPECT showed reduced [99mTc]HMPAOuptake in the same region, whereas increaseduptake was found during a seizure.129

CLINICAL INDICATIONSBoth interictal and ictal functional imagingstudies may show areas of abnormal signal inpatients with focal epilepsy. This localisationis appropriately used to confirm or refute col-lateral evidence from structural MRI andEEG examinations in the assessment ofpatients with refractory epilepsy who arebeing considered for surgical treatment (usu-ally a partial temporal lobectomy).

ONCOLOGYMost cerebral tumours are easily seen bystructural imaging, which typically shows thelesion location, morphological details, evi-dence of damage to the blood-brain barrier,and induced cerebral oedema. From thesedata it is often possible to reach an accurateprediction of tumour type and likely histol-ogy. Functional imaging studies can add fur-ther information. In patients with gliomas,PET [18F]fluorodeoxyglucose studies haveshown a relation between glucose metabolismand both histological gradel30 and survival.'3'It should not be assumed, however, that alesion with high ['8F]fluorodeoxyglucose uptakeis necessarily a tumour, as cerebral abscessesmay also show increased uptake."32 Use hasalso been made of SPECT in an effort to dif-

ferentiate high from low grade gliomas.Thallium-201 (a tracer more familiarly usedin myocardial studies) exhibits increaseduptake in some tumours. In gliomas, uptakeis greater in high grade lesions.I"'A particular clinical problem in neuro-

oncology is the management of patients pre-senting with recurrent lesions afterradiotherapy for tumour. It can be difficult todifferentiate recurrent tumour from radiationinduced necrosis on the basis of clinicalassessment and structural imaging alone.['8F]Fluorodeoxyglucose PET is of clinicaluse in this situation,"34 as recurrent tumourhas a high metabolic rate (fig 7), whereas low['8F]fluorodeoxyglucose uptake suggestsradionecrosis."35 The measurement is notaffected in the early postoperative period, norby steroid treatment."35

Other aspects of tumour biochemistry havealso been explored with PET, including mea-surements of amino acid uptake and proteinsynthesis. ["C]Methionine accumulates read-ily in gliomas,"36 higher uptake usually occur-ring in high grade tumours.'37 DNA synthesiscan also be followed with nucleosides such asdeoxyuridine labelled with fluorine-18."38Peripheral benzodiazepine (w3) receptors areexpressed on human glioma cells; the pres-ence of this tumour marker may be recog-nised with PET and the specific marker["GC]-PK1 195.23

CLINICAL INDICATIONThe principal consensus use of functionalimaging in oncology is in the differentiationof recurrent cerebral glioma from postradia-tion necrosis.

Summary of clinical indications forfunctional imaging studiesAs mentioned at the outset of this review,the cornerstone of clinical neuroimaging

Figure 7 PET f'8F]fluorodeoxyglucose imagesfrom a patient with a recurrent glioma. The PET images (to the left)show increased metabolism indicative of recurrent tumour. (IUlustration courtesy ofProfessorM Maisey.)



j.com/

J Neurol N


ebruary 1995. Dow

nloaded from


Sawle

procedures has been the identification ofbroadly "structural" changes in neural tissue.The tools for such image acquisition (x rayCT and MRI) are widely available and ofincreasingly high quality and resolution. Aswe have so few functional imaging facilities,clinical indications for functional imagingstudies must be restricted to situations whereCT and MRI fail to answer the clinical ques-tion. All of these techniques are presentlydeveloping rapidly, with new MRI sequencesblurring the structure and function divide andnewer PET and SPECT ligands pushing for-ward the capabilities of functional studies.

At the present time, I list the following sen-sible clinical indications for functional imag-ing. To my mind, these are clinical situationsin which functional imaging studies can pro-vide clinical information with important ther-apeutic implications.* Differentiation of tumour recurrence from

radionecrosis (PET/FDG)* Contribution to presurgical assessment of

patients with refractory epilepsy (PET/FDG, SPECT/flow tracers)

* Differentiation of juvenile Parkinson's dis-ease from dopa responsive dystonia(PET/[18F]dopa).Further possibilities of substantial clinical

use include:* Identification of critical gyri or sulci before

neurosurgical or neuroradiological proce-dures (fMRI)

* Neurochemical monitoring of patientsundergoing neurotransplantation proce-dures (PET).Functional imaging studies can also pro-

vide precise information contributing to diag-nostic precision in, for example, thedementias and akinetic-rigid syndromes. Butwhereas functional imaging touches on clini-cal practice, to my mind its principal strengthlies in its position as one of our most power-ful instruments for clinical research.

1 Mallett BL, Veall N. Investigation of cerebral blood flowin hypertension, using radioactive-xenon inhalationand extracranial recording. Lancet 1963;i: 1081-2.

2 Phelps ME. Positron emission tomography (PET). In:Mazziotta JC, Gilman S, eds. Clinical brain imagingprinciples and applications (contemporary neurology seriesNo 39). Philadelphia: F A Davis, 1992:71-107.

3 Lassen NA, Holm S. Single photon emission computedtomography (SPECT). In: Mazziotta JC, Gilman S,eds. Clinical brain imaging principles and applications(contemporary neurology series No 39). Philadelphia: F ADavis, 1992:108-34.

4 Lufkin RB. Magnetic resonance imaging. In: MazziottaJC, Gilman S, eds. Clinical brain imaging principles andapplications (contemporary neurology series No 39).Philadelphia: F A Davis, 1992:39-70.

5 Prichard JW, Rosen BR. Functional study of the brain byNMR. Jf Cereb Blood Flow Metab 1994;14:365-72.

6 Cohen MS, Bookheimer SY. Localization of brain func-tion using magnetic resonance imaging. Trends Neurosci1994;17:268-77.

7 Turner R. Magnetic resonance imaging of brain func-tion. Ann Neurol 1994;35:637-8.

8 Stehling MK, Turner R, Mansfield P. Echo-planar imag-ing: magnetic resonance imaging in a fraction of a sec-ond. Science 1991;254:43-50.

9 Mansfield P, Coxon R, Glover P. Echo planar imaging ofthe brain at 3-0T: first normal volunteer results.Jf Comput Assist Tomogr 1994;18:339-43.

10 Haase A, Frahm J, Matthaei D, Hanicke W, MerboldtKD. FLASH imaging: rapid NMR imaging using lowflip-angle pulses. J Magn Reson Imaging 1986;67:258-66.

11 Lammertsma AA, Frackowiak RSJ, Hoffman JM, et al.The C'505 build-up technique to measure regionalcerebral blood flow and volume of distribution ofwater. _7 Cereb Blood Flow Metab 1989;9:461-70.

12 Lammertsma AA, Jones T. Correction for the presenceof intravascular oxygen-15 in the steady-state tech-nique for measuring regional oxygen extraction ratio inthe b. _7 Cereb Blood Flow Metab 1983;3:416-24.

13 Rinne JO, Laihinen A, Nagren K, et al. PET demon-strates different behaviour of striatal dopamine D-1and D-2 receptors in early Parkinson's disease._JNeurosci Res 1990;27:494-9.

14 Sawle GV, Playford ED, Brooks DJ, Quinn N,Frackowiak RSJ. Asymmetrical presynaptic and post-synaptic changes in the striatal dopamine projection indopa-naive parkinsonism: diagnostic implications ofthe D2 receptor status. Brain 1993;116:853-67.

15 Tedroff J, Aquilonius S-M, Hartvig P, et al. Monoaminere-uptake sites in the human brain evaluated in vivo bymeans of "'C-nomifensine and positron emissiontomography: the effects of age and Parkinson's disease.Acta NeurolScand 1988;77:192-201.

16 Salmon E, Brooks DJ, Leenders KL, et al. A two-compartment description and kinetic procedure formeasuring regional cerebral [l C]nomifensine uptakeusing positron emission tomography. .7 Cereb BloodFlow Metab 1990;10:307-16.

17 Frost JJ, Rosier AJ, Reich SG, et al. Positron emissiontomographic imaging of the dopamine transporter with'C-WIN 35,428 reveals marked declines in mildParkinson's disease. Ann Neurol 1993;34:423-31.

18 Mayberg HS, Sadzot B, Meltzer CC, et al.Quantification of Mu and non-Mu opiate receptors intemporal lobe epilepsy using positron emission tomog-raphy. Ann Neurol 199 1;30:3-1 1.

19 Carson RE, Channing MA, Blasberg RG, et al.Comparison of bolus and infusion methods for recep-tor quantitation: application to ["8F] cyclofoxy andpositron emission tomography. .7 Cercb Blood FlowMetab 1993;13:24-42.

20 Bartenstein PA, Duncan JS, Prevett MC, et al.Investigation of the opioid system in absence seizureswith positron emission tomography. 7 Neurol NeurosurgPsychiatry 1993;56:1295-302.

21 Price JC, Mayberg HS, Dannals RF, et al. Measurementof benzodiazepine receptor number and affinity inhumans using tracer kinetic modelling, positron emis-sion tomography, and ["C]flumazenil. _7 Cereb BloodFlow Metab 1993;13:656-67.

22 Ramsay SC, Weiller C, Myers R, et al. Monitoring byPET of macrophage accumulation in brain afterischaemic stroke. Lancet 1992;339:1054-5.

23 Junck L, Olson JMM, Ciliax BJ, et al. PET imaging ofhuman gliomas with ligands for the peripheral benzodi-azepine binding site. Ann Neurol 1989;26:752-8.

24 Frey KA, Koeppe RA, Mulholland GK, et al. In vivomuscarinic cholinergic receptor imaging in humanbrain with ["C]scopolamine and positron emissiontomography. .7 Cereb Blood Flow Metab 1992;12:147-54.

25 Yanai K, Watanabe T, Yokoyama H, et al. Mapping ofhistamine H, receptors in the human brain using["C]pyrilamine and positron emission tomography._7 Neurochem 1992;59: 128-6.

26 Lammertsma AA, Bench CJ, Price GW, et al.Measurement of cerebral monoamine oxidase B activ-ity using L-["C]Deprenyl and dynamic positron emis-sion tomography. I. Cereb Blood Flow Metab 1991;11:545-56.

27 Friston KJ, Frith CD, Liddle PF, Dolan RJ,Lammertsma AA, Frackowiak RSJ. The relationshipbetween global and local changes in PET scans. . CerebBlood Flow Metab 1990;10:458-66.

28 Friston KJ, Frith CD, Liddle PF, Frackowiak RSJ.Comparing functional (PET) images: the assessment ofsignificant change. _7 Cereb Blood Flow Metab 1991;11:690-9.

29 Watson JDG, Myers R, Frackowiak RSJ, et al. Area V5of the human brain: evidence from a combined studyusing positron emission tomography and magnetic res-onance imaging. Cereb Cortex 1993;3:79-94.

30 Winchell HS, Baldwin RM, Lin TH. Development ofI-123 labelled amines for brain studies: Localization ofI-123 iodophenylalkylamines in rat brain. _7 Nucl Med1980;21:940.

31 Greenberg JH, Lassen NA. Characterization of '9-Tc-Bicasate as an agent for the measurement of cerebralblood flow with SPECT. .7 Cereb Blood Flow Metab1994;14(suppl 1):S1-S3.

32 Hajnal JV, Myers R, Oatridge A, Schwieso JE, YoungIR, Bydder GM. Artifacts due to stimulus correlatedmotion in functional imaging of the brain. Magn ResonMed 1994;31:283-91.

33 Woods RP, Cherry SR, Mazziotta JC. Rapid automatedalgorithm for aligning and reslicing PET images._7 Comput Assist Tomogr 1992;115:565-87.

34 Tyszka JM, Grafton ST, Chew W, Woods RP, CollettiPM. Parceling of mesial frontal motor areas duringideation and movement using functional magnetic res-onance imaging at 1-5 Tesla. Ann Neurol 1994;35:746-9.

35 Wise RJS, Bemardi S, Frackowiak RSJ, et al. Serialobservations on the pathophysiology of acute stroke:the transition from ischaemia to infarction as reflectedin regional oxygen extraction. Brain 1983;106:197-222.

36 Gibbs JM, Wise RJS, Leenders KL, Jones T. Evaluationof the cerebral reserve in patients with carotid arteryocclusion. Lancet 1984;i:310-4.

37 Kushner M, Reivich M, Fieschi C, et al. Metabolic and

142


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



clinical correlates of acute ischaemic infarction.Neurology 1987;37:1103-10.

38 Chollet F, DiPiero V, Wise RJS, Brooks DJ, Dolan RJ,Frackowiak RSJ. The functional anatomy of motorrecovery after stroke in humans: a study with positronemission tomography. Ann Neurol 1991;29:63-71.

39 Weiller C, Chollet F, Friston KJ, Wise RJS, FrackowiakRSJ. Functional reorganization of the brain in recoveryfrom striatocapsular infarction in man. Ann Neurol1992;31:305-14.

40 Baron JC, Bousser MG, Rey A, et al. Reversal of focal"misery-perfusion" by extra-intracranial artery bypassin haemodynamic cerebral ischaemia. Stroke 1981;12:454-9.

41 Gibbs JM, Wise RJS, Thomas DJ, Mansfield AO, RossRussell RW. Cerebral haemodynamic changes afterextracranial-intracranial bypass surgery. NeurolNeurosurg Psychiatry 1987;50: 140-50.

42 EC-IC bypass study group. Failure of extracranial-intracranial bypass to reduce the risk of ischaemicstroke: Results of an international randomised trial. NEnglJ Med 1985;313:1 191-200.

43 Humphrey P. Stroke and transient ischaemic attacks.Neurol Neurosurg Psychiatry 1994;57:534-43.

44 Frackowiak RSJ, Pozzilli C, Legg NJ, et al. Regionalcerebral oxygen supply and utilization in dementia. Aclinical and physiological study with oxygen-15 andpositron tomography. Brain 1981;104:753-78.

45 Foster NL, Chase TN, Fedio P, Patronas NJ, BrooksRA, DiChiro G. Alzheimer's disease: focal corticalchanges shown by positron emission tomography.Neurology 1983;33:961-5.

46 Herholz K, Adams R, Kessler J, Szelies B, Grond M,Heiss W-D. Criteria for the diagnosis of Alzheimer'sdisease with positron emission tomography. Dementia1990;1:156-64.

47 Kumar A, Schapiro M, Grady C, et al. High-resolutionPET studies in Alzheimer's disease. Neuropsycho-pharnacology 1991;4:35-46.

48 Haxby JV, Grady CL, Koss E, et al. Heterogeneous ante-rior-posterior metabolic patterns in dementia of theAlzheimer-type. Neurology 1988;38: 1853-63.

49 Foster NL, Chase TN, Patronas NJ, Gillespie MM,Fedio P. Cerebral mapping of apraxia in Alzheimer'sdisease by positron emission tomography. Ann Neurol1986;19: 139-43.

50 Kuhl DE, Small GW, Reige WH. Cerebral metabolicpatterns before the diagnosis of probable Alzheimer'sdisease. Jf Cereb Blood Flow Metab 1987;7:S406.

51 Montaldi D, Brooks DN, McColl JH, et al.Measurements of regional cerebral blood flow and cog-

nitive performance in Alzheimer's disease. NeurolNeurosurg Psychiatry 1990;53:33-8.

52 Burns A, Philpot MP, Costa DC, Ell PJ, Levy R. The

investigation of Alzheimer's disease with single photonemission tomography. Jf Neurol Neurosurg Psychiatry1989;52:248-53.

53 Waldemar G, Bruhn P, Kristensen M, Johnsen A,Paulson OB, Lassen NA. Heterogeneity of neocorticalcerebral blood flow deficits in dementia of theAlzheimer type: a [99-Tc]-d,l-HMPAO SPECT study.Neurol Neurosurg Psychiatry 1994;57:285-95.

54 Kennedy AM, Frackowiak RSJ, Newman S, Roques P,Rossor MN. Presymptomatic deficits in individuals at

risk of familial Alzheimer's disease: a PET study. AnnNeurol 1995 (in press).

55 Kamo H, McGeer PL, Harrop R, et al. Positron emis-sion tomography and histopathology in Pick's disease.Neurology 1987;37:439-45.

56 Salmon E, Franck G. Positron emission tomographicstudy in Alzheimer's disease and Pick's disease.Archives of Gerontology and Geriatrics 1989;8:241-7.

57 Leenders KL, Frackowiak RSJ, Lees AJ. Steele-Richardson-Olszewski syndrome. Brain energy metab-olism, blood flow and fluorodopa uptake measured bypositron emission tomography. Brain 1988;111:615-30.

58 Karbe H, Grond M, Huber M, Herholz K, Kessler J,Heiss W-D. Subcortical damage and cortical dysfunc-tion in progressive supranuclear palsy demonstrated bypositron emission tomography. Neurol 1992;239:98-102.

59 Foster NL, Gilman S, Berent S, Morin EM, Brown MB,Koeppe RA. Cerebral hypometabolism in progressivesupranuclear palsy studied with positron emissiontomography. Ann Neurol 1988;24:399-406.

60 Neary D, Snowden JS, Shields RA, et al. Single photonemission tomography using 99mTc-HM-PAO in theinvestigation of dementia. Jf Neurol Neurosurg Psychiatry1987;50:1101-9.

61 Tyrrell PJ, Warrington EK, Frackowiak RSJ, RossorMN. Progressive degeneration of the right temporallobe studied with positron emission tomography.Jf Neurol Neurosurg Psychiatry 1990;53: 1046-50.

62 Tyrrell PJ, Warrington EK, Frackowiak RSJ, RossorMN. Heterogeneity in progressive aphasia due to focalcortical atrophy: A clinical and PET scan study. Brain1990;113:1321-36.

63 Kuhl DE, Phelps ME, Markham CH, Metter EJ, RiegeWH, Winter EJ. Cerebral metabolism and atrophy inHuntington's disease determined by 18FDG and com-

puted tomographic scans. Ann Neurol 1982;12:425-34.64 Hayden MR, Martin WRW, Stoessl AJ, et al. Positron

emission tomography in the early diagnosis ofHuntington's disease. Neurology 1986;36:888-94.

65 Grafton ST, Mazziotta JC, Pahl JJ, et al. A comparison ofneurological, metabolic, structural, and genetic evalua-tions in persons at risk for Huntington's disease. AnnNeurol 1990;28:614-21.

66 Mazziotta JC, Phelps ME, Pahl JJ, et al. Reduced glucosemetabolism in asymptomatic subjects at risk forHuntington's disease. N EnglJ Med 1987;316:357-62.

67 The Huntington's disease collaborative research group.A novel gene containing a trinucleotide repeat that isexpanded and unstable in Huntington's disease chro-mosomes. Cell 1993;72:971-83.

68 MacMillan JC, Snell RG, Tyler A, et al. Molecularanalysis and clinical correlations of the Huntington'sdisease mutation. Lancet 1993;342:954-8.

69 Aylward EH, Brandt J, Codori AM, Mangus RS, BartaPE, Harris GJ. Reduced basal ganglia volume associ-ated with the gene for Huntington's disease in asymp-tomatic at-risk persons. Neurology 1994;44:823-8.

70 Kuhl DE, Metter EJ, Riege WH. Patterns of local cere-bral glucose utilisation determined in Parkinson's dis-ease by the 18F-fluorodeoxyglucose method. AnnNeurol 1984;15:419-24.

71 Raichle ME, Perlmutter JS, Fox PT. Parkinson's disease:metabolic and pharmacological approaches withpositron emission tomography [abstract]. Ann Neurol1984;15:S131-2.

72 Martin WRW, Beckman JH, Calne CB, et al. Cerebralglucose metabolism in Parkinson's disease. Can JNeurol Sci 1984;11:169-73.

73 Leenders KL, Palmer AJ, Quinn N, et al. Braindopamine metabolism in patients with Parkinson's dis-ease measured with positron emission tomography.J Neurol Neurosurg Psychiatry 1986;49:853-60.

74 Sawle GV, Colebatch JG, Shah A, Brooks DJ, MarsdenCD, Frackowiak RSJ. Striatal function in normalaging: Implications for Parkinson's disease. Ann Neurol1990;28:799-804.

75 Eidelberg D, Takikawa S, Dhawan V, et al. Striatal '8F-DOPA uptake: absence of an aging effect. J Cereb BloodFlow Metab 1993;13:881-8.

76 Martin WRW, Palmer MR, Patlak CS, Calne DB.Nigrostriatal function in humans studied with positronemission tomography. Ann Neurol 1989;26:535-42.

77 Sawle GV, Wroe SJ, Lees AJ, Brooks DJ, FrackowiakRSJ. The identification of presymptomatic parkinson-ism: clinical and [18F]Dopa PET studies in an Irishkindred. Ann Neurol 1992;32:609-17.

78 Sawle GV, Turjanski N, Brooks DJ. The rate of progres-sion of clinical and subclinical Parkinson's disease.J Neurol Neurosurg Psychiatry 1992;55: 1215.

79 Bhatt MH, Snow BJ, Martin WRW, Pate BD, Ruth TJ,Calne DB. Positron emission tomography suggests thatthe rate of progression of idiopathic parkinsonism isslow. Ann Neurol 1991;29:673-7.

80 Sawle GV. The rate of progression of Parkinson's dis-ease. Ann Neurol 1992;31:229.

81 Sawle GV, Myers R. The role of positron emissiontomography in the assessment of human neurotrans-plantation. Trends Neurosci 1993;16: 172-6.

82 Widner H, Tetrud J, Rehncrona S, et al. Bilateral fetralmesencephalic grafting in two patients with parkinson-ism induced by l-methyl-4-phenyl-1,2,3,6-tetrahy-dropyridine (MPTP). N Engl J Med 1992;327:1556-63.

83 Brooks DJ, Salmon EP, Mathias CJ, et al. The relation-ship between locomotor disability, autonomic dysfunc-tion, and the integrity of the striatal dopaminergicsystem in patients with multiple system atrophy, pureautonomic failure, and Parkinson's disease, studiedwith PET. Brain 1990;113:1539-52.

84 Brooks DJ, Ibafiez V, Sawle GV, et al. Differing patternsof striatal 18F-dopa uptake in Parkinson's disease,multiple system atrophy and progressive supranuclearpalsy. Ann Neurol 1990;28:547-55.

85 Bhatt MH, Snow BJ, Martin WRW, Cooper S, CalneDB. Positron emission tomography in Shy-Drager syn-drome. Ann Neurol 1990;28:101-3.

86 Sawle GV, Brooks DJ, Marsden CD, Frackowiak RSJ.Corticobasal degeneration: a unique pattern of regionalcortical oxygen metabolism and striatal fluorodopauptake demonstrated by positron emission tomogra-phy. Brain 1991;114:541-56.

87 Brooks DJ, Ibanez V, Playford ED, et al. Presynaptic andpostsynaptic striatal dopaminergic function in neuro-acanthocytosis: a positron emission tomographic study.Ann Neurol 199 1;30:166-71.

88 Calne DB, Langston JW, Martin WRW, et al. Positronemission tomography after MPTP: observations relat-ing to the cause of Parkinson's disease. Nature1985;317:246-8.

89 Burn DJ, Sawle GV, Brooks DJ. Differential diagnosis ofParkinson's disease, multiple system atrophy, andSteele-Richardson-Olszewski syndrome: discriminantfunction analysis of striatal 18F-dopa PET data.J NeurolNeurosurg Psychiatry 1994;57:278-84.

90 Tyrrell PJ, Sawle GV, Bloomfield PM, et al. Clinical andPET studies in the extrapyramidal syndrome ofdementia of the Alzheimer type. Arch Neurol 1990;47:1318-23.

91 Sawle GV, Hymas NF, Lees AJ, Frackowiak RSJ.Obsessional slowness: functional studies with positronemission tomography. Brain 1991;114:2191-202.

92 Wolters ECh, Huang C, Clark C, et al. Positron emissiontomography in manganese intoxication. Ann Neurol1989;26:647-51.



j.com/

J Neurol N


ebruary 1995. Dow

nloaded from


144

93 Rosenberg NL, Myers JA, Martin WR. Cyanide-inducedparkinsonism: clinical, MRI, and 6-fluorodopa PETstudies. Neurology 1989;39:142-4.

94 Burn DJ, Brooks DJ. Nigral dysfunction in drug-inducedparkinsonism: An "8F-dopa PET study. Neurology1993;43:552-6.

95 Tedroff J, Aquilonius S-M, Laihinen A, et al. Striatalkinetics of [1'C]-(+)-nomifensine and 6-['8F]fluoro-L-dopa in Parkinson's disease measured with positronemission tomography. Acta Neurol Scand 1990;81:24-30.

96 Hantraye P, Brownell AL, Elmaleh D, et al. Dopa-mine fiber detection by ["C]-CFT and PET in a pri-mate model of parkinsonism. NeuroReport 1992;3:265-8.

97 Marek KL, Seibyl JP, Sandridge B, et al. SPECT imag-ing with [I-123]f3-CIT demonstrates striatal dopaminetransporter loss in parkinsonism. Neurology 1994;44(suppl 2):A352.

98 Hagglund J, Aquilonius S-M, Eckernas S-A, et al.Dopamine receptor properties in Parkinson's diseaseand Huntington's chorea evaluated by positron emis-sion tomography using 'C-N-methyl-spiperone. ActaNeurol Scand 1987;75:87-94.

99 Wienhard K, Coenen HH, Pawlik G, et al. PET studiesof dopamine receptor distribution using [(18)F]fluo-roethylspiperone: findings in disorders related to thedopaminergic system. J Neural Transm 1990;81:195-213.

100 Rinne UK, Laihinen A, Rinne JO, Nagren K, Bergman J,Ruotsalainen U. Positron emission tomographydemonstrates dopamine D-2 receptor supersensitivityin the striatum of patients with early Parkinson's dis-ease. Mov Disord 1990;5:55-9.

101 Rinne JO, Laihinen A, Rinne UK, Nagren K, Bergman J,Ruotsalainen U. PET study on striatal dopamine D2receptor changes during the progression of earlyParkinson's disease. Mov Disord 1993;8:134-8.

102 Brucke T, Podreka I, Angelberger P, et al. Dopamine D2receptor imaging with SPECT: studies in differentneuropsychiatric disorders. J Cereb Blood Flow Metab1991;11:220-8.

103 Tatsch K, Schwarz J, Oertel WH, Kirsch C-M. SPECTimaging of dopamine D2 receptors with 123I-IBZM:initial experience in controls and patients withParkinson's syndrome and Wilson's disease. Nucl MedCommun 1991;12:699-707.

104 Pizzolatto G, Rosatto A, Briani C, et al. Alterations ofstriatal D2 receptors contribute to deterioratedresponse to L-dopa in Parkinson's disease (PD): a 123I-IBZM study. Neurology 1993;43(suppl):A271.

105 Schwarz J, Tatsch K, Arnold G, et al. '23I-iodobenza-mide-SPECT predicts dopaminergic responsiveness inpatients with de novo parkinsonism. Neurology 1992;42:556-61.

106 Burn DJ, Mathias CJ, Quinn N, Marsden CD, BrooksDJ. Striatal opiate receptor binding in Parkinson's dis-ease and multiple system atrophy: 'IC-diprenorphinestudy. Neurology 1993;43(suppl):A270.

107 Gilman S, Koeppe RA, Junck L, Kluin KJ, Lohman M,St Laurent RT. PET studies of cerebellar bensodi-azepine receptors with [1 C]flumazenil show increasedbinding in MSA and decreased binding in OPCA.Neurology 1994;44: (suppl 6) :A353.

108 Mayberg HS, Starkstein SE, Sadzot B, et al. Selectivehypometabolism in the inferior frontal lobe indepressed patients with Parkinson's disease. AnnNeurol 1990;28:57-64.

109 Eidelberg D, Moeller JR, Dhawan V, et al. The meta-bolic anatomy of Parkinson's disease: Complementary[(18)F]fluorodeoxyglucose and [(18)]fluorodopapositron emission tomographic studies. Mov Disord1990;5:203-13.

110 Playford ED, Jenkins IH, Passingham RE, Nutt J,Frackowiak RSJ, Brooks DJ. Impaired mesial frontaland putamen activation in Parkinson's disease: Apositron emission tomography study. Ann Neurol 1992;32:151-61.

111 Jenkins IH, Fernandez W, Playford ED, et al. Impairedactivation of the supplementary motor area inParkinson's disease is reversed when akinesia is treatedwith apomorphine. Ann Neurol 1992;32:749-57.

112 Rascol 0, Sabatini U, Chollett F, et al. Supplementaryand primary sensory motor area activity in Parkinson'sdisease. Arch Neurol 1992;49: 144-8.

113 Rao SM, Binder JR, Bandettini PA, et al. Functionalmagnetic resonance imaging of complex human move-ments. Neurology 1993;43:2311-8.

114 Rao SM, Binder JR, Bandettini PA, et al. Relationshipbetween movement rate and functional magnetic reso-nance signal change in primary motor cortex. Neurology1994;44(suppl 2) :A262.

115 Turjanski N, Burn DJ, Lammertsma AA, et al. PETstudies on DI and D2 receptor status in chorea.Neurology 1993;43:A333.

116 Nygaard TG, Marsden CD, Fahn S. Dopa-responsivedystonia: long term treatment, response and prognosis.Neurology 1991;41:174-81.

117 Quinn N, Critchley P, Marsden CD. Young onsetParkinson's disease. Mov Disord 1987;2:73-91.

118 Sawle GV, Leenders KL, Brooks DJ, et al. Dopa-respon-sive dystonia: [18F]dopa positron emission tomo-graphy. Ann Neurol 199 1;30:24-30.

119 Snow BJ, Nygaard TG, Takahashi H, Calne DB.Positron emission tomographic studies of dopa-respon-sive dystonia and early-onset idiopathic parkinsonism.Ann Neurol 1993;34:733-8.

120 Sawle GV, Morrish PK, Playford ED, Burn DJ, BrooksDJ. Young-onset parkinsonism: clues from ['8F]dopaPET studies. Neurology 1994;44(suppl 2):A353-4.

121 Engel J, Kuhl DE, Phelps ME, Mazziotta JC. Interictalcerebral glucose metabolism in partial epilepsy: its rela-tion to EEG changes. Ann Neurol 1982;12:510-7.

122 Theodore WIH, Newmark ME, Sato S, et al. '"F-fluoro-deoxyglucose positron emission computed tomographyin refractory complex partial seizures. Ann Neurol1983;14:429-37.

123 Abou-Khalil BW, Siegel GJ, Sackellares JC, Gilman S,Hichwa RD, Marshall R. Positron emission tomo-graphy studies of cerebral glucose metabolism inchronic partial epilepsy. Ann Neurol 1987;22:480-6.

124 Theodore WH, Jabbari B, Leiderman D, McBurney J,Van Nostrand D. Positron emission tomography andsingle photon emission tomography in epilepsy: com-parison on cerebral blood flow and glucose metabo-lism. Ann Neurol 1990;28:262-3.

125 Engel J, Kuhl DE, Phelps ME. Patterns of human localcerebral glucose metabolism during epileptic seizures.Science 1982;218:64-6.

126 Ryvlin P, Philippon B, Cinotti L, Froment JC, Le BarsD, Mauguiere F. Functional neuroimaging strategy intemporal lobe epilepsy: a comparative study of '8FDG-PET and 99'Tc-HMPAO-SPECT. Ann Neurol 1992;31: 650-6.

127 Sadzot B, Debets R, Maquet P, Comar C, Franck G.PET studies of patients with partial epilepsy: visualinterpretation vs. semi-quantitation/quantitation. ActaNeurol Scand Suppl 1994;152:175-8.

128 Duncan R, Patterson J, Hadley DM, et al. CT, MR andSPECT imaging in temporal lobe epilepsy. _7 NeurolNeurosurg Psychiatry 1990;53: 11-5.

129 Jackson GD, Connelly A, Cross JH, Gordon I, GadianDG. Functional magnetic resonance imaging of focalseizures. Neurology 1994;44:850-6.

130 Di Chiro G, DeLaPaz RL, Brooks RA, et al. Glucose uti-lization of cerebral gliomas measured by ['8F]fluoro-deoxyglucose and positron emission tomography.Neurology 1982;32: 1323-9.

131 Patronas NJ, Di Chiro G, Kufta C, et al. Prediction ofsurvival in glioma patients by PET. _7 Neurosurg 1985;62:816-22.

132 Sasaki M, Ichiya Y, Kuwabara Y, et al. Ringlike uptakeof [(18)FIFDG in brain abscess: A PET study..7 Comput Assist Tomog 1990;14:486-7.

133 Kim KT, Black KL, Marciano D, et al. Thalium-201SPECT imaging of brain tumours. .7 Nucl Med1990;31:965-9.

134 Mazziotta JC. The continuing challenge of primary braintumour management: the contribution of positronemission tomography. Ann Neurol 1991;29:345-6.

135 Glantz MJ, Hoffman JM, Coleman RE, et al.Identification of early recurrence of primary centralnervous system tumours by ["8F]fluorodeoxyglucosepositron emission tomography. Ann Neurol 1991;29:347-55.

136 Hatazawa J, Ishiwata K, Itoh M, et al. Quantitative eval-uation of L-[methyl-C-l l]methionine uptake in tumorusing positron emission tomography. _7 Nucl Med 1989;30:1809-13.

137 Bustany P, Chatel M, Derlon M, et al. Brain tumourprotein synthesis and histological grades: a study bypositron emission tomography (PET) and C-i1-L-methionine. _7 Neurooncol 1986;3:397-404.

138 Tsurumi Y, Kameyama M, Ishiwata K, et al. (18)F-flu-oro-2'-deoxyuridine as a tracer of nucleic acid metabo-lism in brain tumors. _Neurosurg 1990;72:110-3.

Sawle


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Imagingthe head: functional imaging · tomography (PET), single photon emission computed tomography...

Documents

Transcript of Imagingthe head: functional imaging · tomography (PET), single photon emission computed tomography...