Radionuclide cisternography has the unique ability to show gross abnormalities of the circulationand reabsorption of cerebrospinal fluid (CSF). Themain application of this method has been the provision of criteria distinguishing patients with dementia due to such irreversible processes as cerebralatrophy from the potentially treatable group withsymptomatic occult hydrocephalus ( 1—3). Less attention has been directed to fine structural detailsin these studies since the anatomy of the CSF compartments around the brain strains the spatial resolution of conventional radionuclide imaging (4).

Transverse-section radionuclide imaging, by separating confusing overlapping images, has alreadybeen shown to provide better information aboutfailure of the blood—brain barrier than ordinaryscanning (5,6) . The extension of these principles totransmission scanning by the introduction of computerized axial transverse scanning has clearly shownthe advantagesof this mode of brain examination(7—9).

Our project sought to determine the feasibility ofadding transverse-section scanning to ordinary radionuclide cisternography and to define the details ofanatomy of the basal cisterns, the subarachnoidspaces, and the intracerebral ventricles of normal andabnormal patients.

MATERIALS AND MFTHO[)S

In this study, 15 patients underwent both conventional and transverse-section radionuclide cisternography. Thirteen of these were being evaluated forpossible obstructive hydrocephalus and dementia,but we included two patients with suspected CSFrhinorrhea.Also, a numberof patientsproved ultimately not to have hydrocephalus.

Conventional radionuclide cisternography in thesepatients followed the intrathecal injection througha lumbar puncture of 0.5—1.0mCi of 1111n-DTPAin a volumeof 1—2ml. Multiple viewswith the scmtillation camera (Picker Dynacamera 2C or SearleRadiographics Pho/Gamma III) were obtained at 2,6, and 24 hr and, if indicated,at 24-hr intervalsupto 96 hr.

Transverse-section scans were performed at 6 and24 hr using the Mark Ill scanner (10). Each patientunderwent multiple-section studies at different levels,usually with the section plane parallel to the orbitomeatal line. Section levels were chosen eitherby direct observation of rectilinear scans on the

Received Feb. 19, 1976; original accepted April 27, 1976.For reprints contact: Howard P. Rothenberg, Dept. of

Radiology, Hospital of the University of Pennsylvania, 3400Spruce St., Philadelphia, PA 19104.

924 JOURNAL OF NUCLEAR MEDICINE

jDIIi/INSTRUMENTATIDN AND PHYSICS

Tra nsverse-Sectlon RadIo n ucllde ScannIng

In Clsternography

Howard P.Rothenberg,John Devenney, and David E.Kuhl

Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania

By applying the technique of transverse-sectionradionuclide scanning tocisternography, the structure and relationships of the basal cisterns andother subarachnoid spaces of the brain can be visualized more clearly andwith more detail than is possible with routine imaging techniques. Theability of this method to separate overlapping areas of radioactivity ensuresimproved definition of space-occupying processes within the basal cisterns.

in the evaluation by cisternography of patients with hydrocepluilus trneldementia, the transverse-section images clearly separated various normaland abnormal patterns, whereas the routine cisternogram images wereequivocal.

J Nuci Med 17: 924-929, 1976

INSTRUMENTATION AND PHYSICS

Mark III display, or by reference to the orbitomeatalline. Each Section scan required from 2.5 to 20 mm

for completion, depending on the concentration of

activity in the CSF. After processing of the transverse-Section scan data ( I 1 ) , each transverse section

produced a cross-sectional image of the radioactivityin a 1-cm-thick slice of the head at a known level.A series of these scans is then taken at I -cm intervals to examine the larger volume of brain understudy systematically.

RESULTS

Representative studies selected from the group

of patients examined during this project are shownin thissection.

Normal cisterns.Figure 1, right, illustratesthenormal appearance of section scans through thebasal cisterns at 6 hr after injection, when the activityin the basal cisterns is greatest. There is little activityin the overlying superficial CSF pathways. Notice

(A ) the clear display of the cisternal anatomy, un

hindered by superposition of adjacent radioactivity;(B) the symmetry of these normal structures; (C)the configuration and location of the related cisternsand their communicating pathways; and (D) the absence of activity in the regions of the lateral yentricles.

Section 1a, the highest level, is taken through thesuperior extent of the quadrigeminal plate cistern

and shows activity in the structures just above the

tentorial incisura. This level at this time shows theearliest ascent of radionuclide from the infratentorialto the supratentorial pathways. No significant activity

appears over the convexities; only the quadrigeminalplate region, the sylvian fissures, and the sagittalsinusregionsare seen.

Section lb is made through the deepest portion ofthe cistern of the quadrigeminal plate. The transversesection clearly shows the individual cisterns at thislevel, where the supratentorial and infratentorialcompartments overlap.

Section 1c is 2 cm below the previous section andshows more of the structures in the posterior fossaand along the floor of the anterior fossa. At this

level the configuration of the midline activity changes.Instead of activity in the sagittal sinus region, wesee midline activity anteriorly, in the interhemispheric region where the faix cerebri meets the floorof the anterior fossa. The posterior border of thisactivity reaches the lamina terminalis. Activity in thesylvian fissure passes into the midline cisternal struc

tures, showing the origin of these cisterns along thehorizontal portions of the middle cerebral arteries.

The cisterns of the middle cerebral arteries mergewith the parasellar cisterns, which are prominent at

ANT

dI.―

POST

F1G. 1. Transverse-sectionanatomyof normalbasalcisterns.Level of each transverse section is shown by horizontal line inaccompanying diagram. (CA) Cisterna ambiens; (CM) cisterna magna;(Cpa) cerebellopontine angle; (lH) interhemispheric; (lp) interpedun.cular; (Lcc) lateral cerebellar; (LS) lateral sinus; (LT) lamina terminalis; (MCA) cistern of middle cerebral artery; (Ps) parasellarcisterns; (OP) quadrigeminal plate cistern; (5) parasagittal; (Syl)sylvian cistern; (Va) vallecula.

this levelwhilebarelyvisiblein the levelabove.Themore posterior activity in the midline is within theinterpeduncular cistern. The weak activity extending

posteriorly and laterally from the parasellar regionis within the cistema ambiens. Finally, activity is seenin the posterior fossa in the lateral cerebellar cisterns,the cisterna magna, and a small amount of activityin the most caudal region of the quadrigeminal platecistern. A characteristic distinction between the supratentorial and infratentorial sections is seen at their

interface; this is the configuration of the midlineactivity. While the supratentorial section shows themore posteriorly located rostral cistern of the quadrigeminal plate, this disappears in the infratentorial

section, whereas the more caudal and anteriorly located parasellar and interpeduncular cisterns becomemore prominent. These features are better shown inSection 1d.

The lowest section here, Fig. I d, is made throughthe center of the posterior fossa through the activitylateral to the cerebellar hemispheres, the cisternamagna and vallecula, and the cerebellopontine angles. This section shows most clearly the union of themany basilar cisternal compartments. Here the cisternal and CSF pathways along the undersurface ofthe brain stem come together at the interpeduncular

and parasellar cisterns. This activity extends superiorly into the cisterns of the middle cerebral

arteries, the sylvian cisterns, and the circular cisterns.Communication rostrally with the olfactory sulcusand interhemispheric and sagittal cisterns can also be

seen. Finally, CSF anterior to the brain stem along

Volume 17, Number 10 925

C'

FIG.2. Normalcisternogramat24hrshows only convexity (Cony) and parasagittal (5) activity. Horizontal lines throughright lateral and anterior views of cisterno.gram show level of section. (A) Sectionviews of cisternogram through basal cistern. (B) Transverse section through lateralventricles.

- fly

ROTHENBERG, DEVENNEY, AND KUHL

A. RL ANT ! ANT

@ :iJ@@iB. RL ANT

I

RL :LL patient with normal-pressure hydrocephalus areshown in Fig. 3. Almost all of the activity is locatedwithin the ventricularsystem,althoughsomeis seeninferiorly in the basal cisterns. The persistence ofthis pattern beyond 24 hr indicates the presence ofa block in the normal pathways of CSF migration andabsorption and shows the alternate compensatorypathway of reabsorption through the ventricles.

Figure 4 showsthe transversesectionsmade at6 hr in the samepatientas in Fig. 3. Thesesectionsare keyed in level to the posterior rectilinear scansat the left. Section 4a passes through the lateralrecesses of the fourth ventricle. Even at this level

ANT POST

a. POST ANT

MCA

:4@;ANT

F1G.3. In conventional6-hrcisternogrampenetrationisseeninto right and left lateral ventricles (RLV and LIV) and third andfourth ventricles. Some activity is present in basal cisterns (BC).

b. POST

LV±

(bo d

- ANTC. POST

POST

F1G.4. Transverse-sectionappearanceof activitywithinyentricular system. (a) Transverse section at level of fourth ventricleshows activity in fourth ventricle, parasellar cisterns (PS), and cistern of middle cerebral artery (MCA). (b) Transverse section scanthrough inferior portion of lateral ventricle shows most of activitywithin ventricular system. (c) Transverse section through superiorportion of lateral ventricles.

926 JOURNAL OF NUCLEAR MEDICINE

@‘@:@‘

LLV

the clivus communicates with the CSF activity overthe cerebellar hemispheres.

Normalcisternsin delayedscans.Sectionscansat24 hr after injection differ considerably from theearlier versions. The major activity is now distributedabout the convexity and in the sagittal sinus region,

and no significant activity is noted within the basalcisterns (Fig. 2) . In the normal patient, no activitywill be seen in the regions of the lateral ventricles.

Ventricular penetration. In adult patients withhydrocephalus and dementia, the most common abnormality observed on the cisternogram is the appearance of radioactivity in the cerebral ventricles.This is almost never seen in normal patients (12).The following examples illustrate the value of section scanning in distinguishing those patterns associated with ventricular penetration of the tracer.

Early scans. The conventional 6-hr scans of a

INSTRUMENTATION AND PHYSICS

most of the activity is in the ventricular system, sepa

rated from activity in the more anterior and lateral

cisterns. Section 4b is made at a higher level andshows the inferior portions of the lateral ventricles,

distinguishable from the basal cisterns at this level.Section 4c shows activity in the uppermost portions

of the lateral ventricles and its failure to ascend into

the adjacent cisterns, sagittal sinus, and the sylvian

fissures. Compare the abnormal appearance of sec

tions in this figure with the normal 6-hr appearancein Fig. 1.

The sections of Fig. 4 emphasize what is alreadyobvious in the ordinary rectilinear views of Fig. 3.More commonly, the section study clarifies what isobscure in the rectilinear views; a vertex view can

also help to visualize ventricular penetration. Forexample, in the cisternogram illustrated in Fig. 5,a diagnosis of ventricular penetration, inconclusive

in all ordinary views, was clarified by the transversesection scans. Note the difficulty of separating theventricular radioactivity from other images in thevertex view, where all levels of structures are superimposed in one picture. The four transverse sectionspermit positive identification of ventricular filling,since the otherwise overlapping images of adjacentcisterns and ventricles are separated in the sections.

Compare these sections to the normal anatomyshownin Fig. 1. Section5a passesthroughthe levelof the caudal medullary structures. It clearly showsactivity (A) within the posterior fossa along theclivus; (B) in region of the cerebellopontine angle;(C) in the cisterna magna; and (D) in the centralarea, which suggests activity either in the fourth yentricle or along the anterior surfaceof the medullaand pons. Section Sb is at a higher level and showsactivity in the posterior fossa, the more rostral basal

cisterns,and within the caudalextent of the quadrigeminal plate cistern. Section Sc is at the junctionof the supra- and infratentorial compartments. Thissection clearly shows ventricular activity. The highestsection, Sd, shows considerable ventricular activityand some additional activity in the sylvian fissures,but it is free from the otherwise interfering imagesof the basal cisterns.

Figure 6 compares 24-hr rectilinear and sectionalviews showing stasis of tracer in the ventricular system. As is often the case in delayed scans, the yentricular activity is somewhat obscured because ofdiffusion of @‘In-DTPAinto the brain (13) and theconfusing overlap of activity in the sylvian fissuresand convexities. The transverse sections clearly identify the ventricularpenetrationand stasis of radionuclide, which might have been misinterpreted inthose two cases.

Tumor-deformed cisterns. Less commonly, cisternography is usedto identify tumor location by notingdistortion of adjacent cisterns. These distortions aresubtle at best when one is restricted to ordinary brainimaging (14,15) . Section scanning can improve delineation of the abnormal cisternal boundaries, asillustrated in the following case.

A 60-year-old man underwent a right suboccipitalcraniotomy 1.5 years before for removal of a

S right acoustic neuroma. His presenting symptoms

. (A) suggested recurrence. Pertechnetate brain imaging

@@ (Post)showedabnormalactivityintherightposteriorparamedian region (Figs. 7a and 7b) , but there was

S confusion as to whether the abnormal uptake was

due to recurrent tumor, to persistent postoperativeartifact, or to a combination of both. A subsequentcisternogram (Fig. 7c) suggested decreased activityin the cistern of the right cerebellopontine angle. Fig

ures 7a—7cshow what can be accomplished with rou

a. ANT C. ANT

b. d.

FIG. 5. Scinticisternogramvertexview(top)iscontrastedwithtransverse sections done at approximately same time. Sections athrough d are made at increasingly higher levels. (Cl) clivus; (M)medulla; (OG) olfactory groove; (OP Vent) quadrigeminal plate andfourth ventricle activity.

Volume 17, Number 10 927

POET

ROTHENBERG, DEVENNEY, AND KUHL

ANTA. RL ANT

POST

B.

F1G.6. Transverse.sectionscansre.veal activity within region of ventricles,which is suspected from rectilinear viewsin patient A but obscured by adjacentactivity in patient B.

FIG.7. Recurrenttumorofrightcer•bellopontine angle. (A,B) Intravenous p.rtechnetate brain scans show abnormal activity in right cerebellopontins angle (C)Posterior scinticisternogram suggests decreased activity in cistern of right cerebellopontine angle. (D) Transverse.ssctionpertechnetate brain scan shows increaseduptake in right cerebellopontine angle.(E,F) Transverse.section cisternogram andaccompanying diagram show relationshipof tumor to deformed and displaced cere•bellopontine angle cistern and anteriorsubarachnoid space.

A B

ED F

tine brain and cisternal imaging alone, but each oftheseexaminations was complemented by transversesection scanning. In Fig. 7d the transverse-sectionpertechnetate scan clearly identifies a round regionof activity in the right cerebellopontine angle, consideredto be recurrenttumor.The sectionscanaftercisternography (Fig. 7e) shows the correspondingdisplacement of activity from the cistern of the rightcerebellopontine angle, with lateral displacement ofthe basal cisterns along the floor of the posteriorfossa. This characterizes the lesion as being extraaxial and within the cerebellopontine angle.

This study shows the advantage of imaging boththe tumor and its effects on adjacent structures andthe use of the section scan to isolate the relevantimages.

DISCUSSION

Computerized axial tomography of the brain (CTscanning) has displaced pneumoencephalographyand cerebral angiography in most neuroradiologic

investigations of central nervous system disease, including the evaluation of hydrocephalus. However,radionuclidecisternographyremains unique in itsability to show graphicallythe physiologyof cerebrospinal fluid circulation, and it remains a valuableprocedure in the preoperative assessment of patientswith dementiaandhydrocephalus(16,17). The ability of the radionuclide transverse-section scan toovercome some of the difficulties in the interpretationof the cisternogram has been dealt with in this project. With further refinements in technique and increasing clinical experience, and in conjunction withother tests of CSF function and clinical criteria ofnormal-pressurehydrocephalus,betterpatientselection can be expected and fewer patients subjectedto the recognizedmorbidityof CSF shunting(18).

In patients with hydrocephalus and dementia whoshow ventricular penetration on cisternography, thetransverse-sectionscanwill providebetter visualization of the ventricular activity. This has beenachieved when the standard rectilinear scans and

928 JOURNAL OF NUCLEAR MEDICINE

ANT

POST

INSTRUMENTATION AND PHYSICS

scinticisternograms have produced equivocal results.The transverse-sectionviews have provided studiesof diagnostic quality, particularly in such low-countsituations as occur with leakage of radionuclide atthe injection site, with expandedCSF volume leadingto low radionuclide concentration, and with otherfactors related to technique and patient idiosyncrasy.In delayed scans detail is often lost due to diffusionof the chelated radionuclide out of the CSF and intothe paraventricular tissues and dural spaces, but thishas not been a problem when the transverse-section

scans are viewed because of the improved separation

of adjacent areas of radioactivity. Patients will benefit from the decreasing numbers of repeated procedures.

Routine radionuclide cisternography has generatedoccasional interest as a method of examining the subarachnoid space for space-occupying lesions. Reportsof this application have been concerned with theinfrequent abnormal collections of CSF and grossdisplacement or obstruction of adjacent subarachnoid spaces by local processes. These have includedthe demonstration of porencephaly (19), arachnoidcysts and other cystic lesions of the posterior fossain children (20,21 ) , the effect of cerebrovascularocclusive disease on the adjacent subarachnoid space(22), subdural hematoma (23), and lesions of thespinal canal (24).

Recently the application of the cisternogram forthe diagnosis of tumors of the cerebellopontine anglehas been evaluated ( 14,15) . Transverse-section radionuclide cisternography has provided superiordefinition of theselesionsbecauseof its ability toseparate the overlapping CSF compartments aboutthe base of the brain, thus offering a relatively Unobstructed view of the affected cisterns. The postenor fossa,the brain stem, the superficialregionsof the cerebellarand cerebralhemispheres,and theadjacent subarachnoid spaces are difficult areas bothfor nuclear imaging and for radiographic contrastexamination. This seems to be true also of computerizedaxial transverse-sectiontransmissionscanning. Transverse-section radionuclide cisternographyshould provide an additional benign method of obtaming useful anatomic information about theseareas, similar in some respects to the informationobtained with pneumoencephalography.

REFERENCES

I . BENSON DF, LE MAY M, PATrEN DH, et al. : Diagnosis of normal-pressure hydrocephalus. N Engi I Med283:609—615,1970

2. HARBERT JC: Radionuclide cisternography. Semin NuciMed I : 90—106,1971

3. KIEFFER SA, WOLFF JM, WESTREICH G : The borderline scinticisternogram. Radiology 106: 133—140,1973

4. JAMES AE, DELAND FH, HODGES Fl, et al. : Cerebro

spinal fluid (CSF) scanning: Cisternography. Am I Roentgenol Radium Titer Nuci Med 110: 74—87, 1970

S. KUHL DE, SANDERSTP: Comparison of rectilinearvertex and transverse section views in brain scanning. I NuciMed 11: 2—8,1970

6. KUHL DE, SANDERSTP: Characterizing brain lesionswith the use of transverse section scanning. Radiology 98:317—328,1971

7. HOUNSFIELDGN : Computerized transverse axial scanning. Part 1. Description of system. Br I Radio! 46: 1016-.1022,1973

8. AMBROSE J: Computerized transverse axial scanning(tomography). Part 2. Clinical application. Br I Radio! 46:1023—1047,1973

9. NEW PFJ, Scorr WR, SCHNuRJA, et al.: Computerized axial tomography with the EM! scanner. Radiology 110:109—123,1974

10. KURL DE, EDWARDSRQ: The Mark II! scanner: Acompact device for multiple view and section scanning ofthe brain. Radiology 97: 563—570,1970

11. KUHL DIE.,E@wuws RQ, RICcI AR: Quantitativesection scanning using orthogonal tangent correction. I NuciMed 14:196-200,1973

12. KIEFFER SA, WoLFF JM, PRENTICE WB, et al. : Scmticisternography in individuals without known neurologicaldisease. Am I Roenigenol Radium Ther Nuc! Med I 12:225—236,1971

13. HARBERT JC, REED V, MCCULLOUGHDC: Comparison between @9-HISAand “Yb-DTPAfor cisternography.I Nuc! Med 14: 765—768,1973

14. MAMO L, HOUDART R: Radioisotope cisternographycontribution to the diagnosis of cerebellopontine angle tumors.INeurosurg37:325—331,1972

15. Orro H, FIEBACH0, SAuER J, et al. : Cerebral wintigraphy in relation to roentgenological methods for detection of tumors situated in the sellar region and the posteriorfossa. Neuroradio!ogy 4: 30—35,1972

16. SALMONJH, Goi@iENJY, BROWNL: Ventriculoatrialshunt for hydrocephalus ex-vacuo psychological and clinicalevaluation. Dis Nerv Syst 32: 299—307, 1971

17. TRO1TER.JL, LUZECKY M, Swar.i. B, et al. : Cerebrospinal fluid infusion test: Identification of artifacts andcorrelation with cisternography and pneumoencephalography.Neurology24: 181—186,1974

18. SAMUELSON S, LONG DM, Cnou SN: Subdural hematomas as a complication of shunting procedures for normalpressure hydrocephalus. I Neurosurg 37: 548—551,1972

19. FRONT D, BEKS JWF, PENNING L: Porencephaly diagnosed by isotope cisternography. I Neurol Neurosurg Psychiatry35:669—675,1972

20. GOLUBOFFLG : Arachnoid cyst of the posterior fossademonstrated by isotope cisternography. I Nuc! Med 14:61-62,1973

21. HARBERTIC, JAMESAE: Posterior fossa abnormalities demonstrated by cisternography. I Nuci Med 13: 73—80,1972

22. HALPERNSE, ALAZRAICIN, HuRwrrz 5, et al.:Changes in radioisotope cisternogram in cerebrovascularocclusive disease. I Nuci Med I 3: 493—497,1972

23. RINALDI I, HARRIS WO, DiCsmto G : Radionuclidecisternography in subdural hematomas. Radiology 105: 597—602,1972

24. O'r'ro H, SAUER J, FIEBACH 0, et al. : Myelo-scintigraphy in preoperative evaluation of space occupying lesionsin the spinal canal. Fortschr Geb Roentgenstr Nuklearmed116: 766—772,1972

929Volume 17, Number 10