484
Direct Intracranial Sagittal and Coronal CT Scanning:
Anatomy and Pathology
Rainer G. Bluemm'
The application of whole-body scanners with a wide aperture
cone to neuroradiology permits examination of the head in the
sagittal plane. This method yields better definition of anatomic
structures such as the inner and outer supra- and infratentorial
cerebrospinal fluid spaces, the major vessels, and all structures
of the posterior fossa close to the midline with negligible artifacts . By mono-, bi-, or triplanar approaches, intracranial pathology can be demonstrated exactly with respect to the major
anatomic landmarks, without the disadvantages of multiplanar
reconstructed views. In combination with an altered positioning
for coronal scanning, optimal biplane evaluation of the sellar
region is possible. Additional cisternal enhancement studies in
variants of empty sella and brainstem tumors may not always be
necessary. The sagittal section , like the coronal , can be placed
lateral to metallic clips for early detection of tumor recurrence.
Limitations are discomfort in positioning and a degradation of
image quality caused by metal amalgam fillings.
It has recentl y been demon strated [1, 2] th at direct (positional)
computed tomographi c (CT) scanning of th e head in the sagittal
pl ane eliminates th e well known hy poden se artifact within the sella,
th e hypodense " Hounsfi eld bars " between the petrous bon es, and
th e hyperdense streaks ari sing from the internal oc c ipital protuberance . Thi s is due to a more " homogenous " sampling volum e. In
our practi ce we have observed a number of situ ation s in whic h th e
sagittal and longitudinal secti ons provid e additi onal and sometimes
invaluabl e information [3- 7] w ith respect to the axial and direct
coronal scans. This report summarizes our experience with more
th an 150 pati ents, aged 5-7 5 years, examined by thi s method to
obtain neurolog ic , neurosurgical, endocrinologic , and maxillofacial
surgica l inform ati on.
Materials and Methods
Our scanner is a Philips Tomoscan 300 . Thi s scanner ca n obtain
six to eight 6-mm-thi c k slices in 78-104 sec, at 8 sec intervals. It
has 288 xe non detectors, a prim ary geometri c enlargement fi eld of
view of 3 2 c m d iameter, 4 and 8 sec scan tim es, 600 pulses / scan,
2 msec pul se width , and operates at 120 kVp .
Th e pati ent is placed pro ne, righ t or left anteri or obliqu e, on an
inc lined plane at th e rear side of th e scann er. Thi s keeps th e
shoulders out of th e fan beam. Arm s are stretc hed alo ng th e hips,
and th e ri ght or left c heek lies again st a foam plasti c head support
I
fi xed at th e top of th e reg ular examin ati on tabl e. This helps to
parallel the longitudinal pl ane of th e head with th e scan pl ane ( +
20 ° tilt), ensuring th at th e head accurately follows th e automati c
table inde xing. To eliminate or reduce angulations between th e
anatomic fea tures and th e scan plane it is advisable to use a light
pencil in the inner ring of th e gantry, a ca p on the patient 's head
labeled with longitudinal strips, and to scan directly th e area of
interest. Because of the uncomfortable neck position , th e duration
of th e examination should be kept short by using a batc h mode
proc edure . To obtain coron al section s of th e sella nearl y perpendic ular to the sellar floor without greater hyperextension of th e
neck , th e patient' s position is modified . The head is turn ed 90 °
verti cally to the scanning pl ane, th e head holder is replaced by a
normal devi ce for coron al scanning , and th e body is put in a
co mplete pron e position .
For collimations of 6 mm or 3 mm , 300 mA or 4 28 mA setting s
were used, respectively . Contrast agents used were megluminamidotri zoate (306 mg / kg) and iopromide (300 mg I/ kg of body
weight) .
Results
Th e sagittal proj ection obtained approximates a true sagittal
section through th e intracranial and facial stru ctures, and is co mparabl e to a longitudinally reconstructed nuc lear magneti c resonance (NMR) image. Th e most interesting anatomy is located in the
midline region (fig . 1). Th e soft tissue above and below th e tentorium, in relation to th e inner and outer cerebrospin al fluid (CSF)
spaces, and in th e vici nity of th e ca lvarium and th e skull base
(including th e paran asal sinu s), is easily visualized on a 3-mm-thi ck
preco ntrast midsagittal sca n. Th e scan is tangential to th e falx
cerebri and the gyri of either sid e are not cl early seen . Th e third
ventricle, the sylvian aqueduc t, and th e tri angular fo urth ventri c le
are well delineated . The brainstem (co mposed of th e mesence phalon, pons, and th e medulla obl ongata) is separated from th e c livus
by th e interpeduncul ar, pontine, and medullary c isterns . Th e pontomedullary notc h is often well visualized . Th e c istern a cerebellomedullari s separates th e inferior vermis from the occ ipital bone; th e
superior vermi s is well outlined by th e vermi an c istern . Th e normal
quadrigeminal cistern is perfectly delineated. Th e major vascular
stru ctures of the posterior fossa are easily identified with or without
co ntrast enh ancement. The entire length of the basilar artery is
seen in th e subarachnoid c istern s in front of the brainstem . Th e
Departm ent of Radiolog y, Knappschafts-Krankenhaus, University Cli ni c , In der Sc horn au 23 / 25 , 4630 Boc hum 7 , West Germ any .
AJNR 4:484-487, M2Y/ June 1983 0 195-6108 / 83 / 0403 -0484 $00 .00 © Ameri can Roe ntgen Ray Soc iety
AJNR:4, May / June 1983
CT OF THE HEAD
48 5
Fig . 1.-3 mm midsagittal section without contrast enhancement in 21-year-old man . Aquedu ct (arrow) . vg = vein of Galen; gv = galenic supravermi an
cistern ; IV = fourth ventric le; pm = pontomedullary ci stern ; ic = interpeduncul ar c istern; III = third ventric le.
Fig. 2.-Partly enlarged branches (arrow) of postcommunicating anterior cerebral art ery leading to frontopari etal angi oma (not shown) near convex ity.
Reduced perfusion of pericallosal artery (pa) (steal effect). icv = intercerebral vein, III = third ventricle ; IV = choroid plex us of third and fourth ventricle.
Fig . 3.-Pituitary gland with infundibulum. III = third ventricle; fm = foramen of Monro; ba = basilar art ery; ss = sphenoid sinus. CSF in the se ll a secondary
to internal hydrocephalus (arrow) .
straight sinus and the great vein of Galen form the superior border
of the posterior fossa ; thus, the free edge of the tentorium is clear.
An angio-CT study of another patient (fig. 2) a few millimeters
lateral to the midline demonstrates branches of the postcommunicating anterior cerebral artery and inner cerebral veins. The superior surface is demarcated by the thin callosal artery (steal effect of
the enlarged proximal branches, leading to a frontoparietal angioma, not shown in the figure), and parts of the corpus callosum
are therefore better visualized . Note also the enhancement of the
choroid plexus of the third and fourth ventricles . The gyri of the
ipsilateral hemisphere can be better visualized because the fal x
cerebri is medial to the section .
The relative position or accurate site of intracranial lesions (especially those that are medial or parasagittal with respect to the
major external anatomic landmarks) cannot be determined on axial
scans alone . The longitudinal view of a 62-year-old patient (fig . 3)
indicates where a calvarial meningioma is located with regard to
shape , size , and the neighboring brain tissue . There is no substantial loss in density resolution as confirmed by the CT values of the
tumor or the gray and white matter. On the ax ial scans (not illustrated) the tumor could not be defined as clearly, because the
sections were tangential to it, and hyperdense streaks of metallic
clips of the opposite parietal side (secondary to an earlier operation)
degraded the image as well as the coronal slice s.
Before the advent of CT in conventional neuroradiology , the
lateral plane prayed an important role in evaluating the sellar region .
Sagittal CT offers the additional possibility of delineation with negligible artifacts of the pituitary stalk , which extends from the infundibular recess of the third ventricle inferiorly and slightly anteriorly
to reach the pituitary gland (fig . 4). It is also useful in patients
lacking obviou s sellar enlargement. By means of sagittal sec tion s
the curved anterior / posterior walls of th e sella can be c learly
defined (figs. 4 and 5A). This is not possible with c oron al sli ces.
Partial anterior herniation of th e diaphragm into a norm al-sized sell a
due to occ lusive hydrocephalus can be vi suali zed on a 6-mm-thick
section .
It is possible to compare positional scans with reform atted vi ews.
Figure 5B illustrates a nearly midsagittal longitudinal vi ew of th e
same patient as in figure 5A . The view is a computer rearrangement
of 13 partly overlapping 5-mm-thic k axial sections. A posterior
fossa tumor is visible, although th e tum or contours appear blurred ,
and further details of th e surrounding stru ctures are diffic ult to
analyze. Figure 5A represents a direct sagittal scan a few millim eters medial to this projection. Apart from th e difference in co ntrast
agent kinetics (nearly homogenou s enhance ment in th e direct scan)
the image clearly demonstrates the bord ers of th e tum or with
respect to the skull base. The longitudinal and verti cal dimensions
between the tentorium and the foramen magnum are c lea rl y visualized .
The supra- / infratentori al parts of mesence ph alic masses ca n be
outlined more precisely th an on coronal sc an s (fig . 6). In thi s case
the cystic tumor compresses th e aqueduct and th e quadrigemin al
c istern but not th e interpeduncular and pontine c istern s. Th e upper
part of the fourth ventricl e is displaced inferi orl y.
Even the so-called half-a xial sec tion +2 5 0 maxim all y to th e
canthomeatalline is limited by interfering metallic dental fillin gs and
degraded by interpetrosal artifacts. In a 43 -yea r-old pati ent w ith
left-sided central hemiparesis a hypodense lesion second ary to
infarction could be c learly defin ed only in a ri ght parasagittal slice
c lose to the midlin e (fig . 7) .
Discussion
Th e clinica lly useful appli cati on of direc t sag ittal CT of th e b rain
extend s from th e midlin e to th e lateral bord er of th e basal gangli a.
Stru ctures d escribed in fig ures 1 -3 ca n be more easil y discriminated an d path olog ic in vo lve ment better delin eated in th e sag ittal
projecti on, according to th e prin ciple of sec ti onin g perp endicul ar to
th e pl ane of th e obj ec t of interest. Thi s also avoid s density averag in g
of rath er small soft-ti ss ue stru ctures (e. g ., th e pituitary gl and) with
th e adjacent bone. Th e indi vid ual co mbin ati on o f scans in th e
sagittal, co ron al, and axial pl anes all ows a mono-, bi-, and tripl anar
approach to co mpl ex lesions-intra- / ex traorbital [5 , 6 ]; intra-/suprasell ar; intra- / ex traventric ular; supra- / infratentori al (fi g . 6); intra/ extraax ial (fi g. 7)-adjacent to th e skull base [5 , 6], ari sing from
th e ca lva rium or dura, and co ngeni ta l anomali es [4 , 7], eve n in
areas of high attenu ati on differences [6 , 8 , 9 ]. In thi s way the
48 6
CT OF THE HEAD
AJNR:4 , May / June 1983
A
Fig . 4.- Accurate tocati on, co nfi guration, and
size of ca tvarial menin giom a wit h respect to adjacent soft ti ssue and bone in longitudin al plane
lateral to metallic clips in 62-year-old patient.
Crani otomy was perform ed due to parietal tumor
in opposite side.
Fig. 5.-Direct sagittal section, 6 mm thick, in 40-year-old patient with progressive loss of hearing
and sympto ms of increased intracranial pressure. A, Exact visualization of tumor size, contours, and
relation to neighboring structures. Difference in contrast agent kin etics. Partial anterior herni ation of
sellar diaphrag m (arrow ) due to occlusive hydroceph alu s. B, Reform atted longitudinal view of extraaxial
posteri or fossa tumor, computer-rearranged from 5 mm thick partly overlapping ax ial slices. Blurred
tumor co ntours. Poorer spati al resolution and inferior topography compared with A.
Fig. 6.-Cystic mesencephalic tumor. Displacement of roof of fourth ventricle inferiorly.
Exact determination of that part of tumor extending below tentorial gap. Interpeduncular and pontine cistern not compressed. Occ ipital hypodense
lesion consistent with secondary ischemic infarction of posterior cerebral artery .
Fig . 7. -Centralleft hemiparesis in 43-year-old
patient. Cl early defined round intraaxial upper
brainstem lesion a few millimeters right lateral to
midline, secondary to ischemic infarcti on. This
could not be outlined exactly even on axial scans
+ 25 0 to baseline (interpetrosal artifact zone).
Enlarged cisterns and fourth ventricle due to atroph y.
6
7
neurosurgeon or radioth erapi st ca n be provid ed with all necessa ry
inform ati on.
As does sagittal CT, coronal scanning of th e sell ar region as
described in thi s paper produ ces negligibl e artifac ts, so an optimal
bi planar evalu ati on of all anatomic and path ologic co nditions of th e
sell ar area is possible . Th e di ag nosti c c riteri a ruled out by means
of d ifferent investi gati onal techni q ues [1 , 10-1 3] ca n be easil y
app lied . In our opini on, va ri ants of an empty se ll a do not necessarily
req uire an additional metriza mid e or air stud y . Situ atio ns and complicati ons before and after transsph enoid al surgery ca n be easily
determin ed on sag ittal sec ti ons [4 , 6 , 7] as far as th e operati ve
route, th e fillin g material after tum or remova l, and leak age of CSF
or air are co ncern ed.
Excep t for NMR , sagittal CT is th e only technique by whi ch
interpetrosa l artifacts d o not occur. To date, different approaches
to thi s prob lem have fail ed to offer a d efinite solution [1 2 , 14]. Thi s
underscores its valu e for th e assessment of lesion s related to the
brainstem (e.g ., pontomed ullary and extrin sic brainstem tum ors
ari sin g from th e cl ivus). As it facilitates th e definiti on of tumors of
th e entire posteri or fossa in combin ation with se lected ax ial and
coro nal sections, it may be an altern ative to c istern al metrizamid eenh anced studi es [1 5].
Th e detection of cerebellar and brainstem atrophy is notabl y
improved by midsagittal slices (fig. 7) [16 , 17], but, to explore th e
entire posterior fossa sagittal reconstru c tion , at least 24 and often
34 overl apping 5-mm-thi c k axial scans have to be obtain ed [1 8].
The sylvi an aqueduct and the third ventricle are normally not seen
on these computer-rearranged vi ews [18]. To apply 1 .5 mm overlapping axial slices does not seem routin ely practi ca ble. Lin ear
interpolation [2] (greater im age size than display matri x pi xel area)
accounts for th e coarse and blurred structure co ntours (fig . 58) .
For this reason and th e stripelik e vi ew, topography is less easily
delineated th an on a direct secti on with co mplete info rm ati on.
Th erefore, wheneve r possible , a lesion should be examin ed directly.
Tum ors and malformat ion s in the midlin e predominate in infantile
intrac ranial path ology [11]. Sagittal scans are easily obtain ed in
c hild ren, since th e motility of their cervic al spine is normal. Th e
results are excellent with regard to tumor stagings and definition of
prim ary and second ary lesions in th e regi on of the aqueduct. Th e
understandin g of co mpl ex malfo rmations and their ru dimentary
variants is better [7]. Al so , th e radi atio n dose received by th e lenses
of th e eye during sagittal scanning is significantly lower in th e
midlin e region [19]. This has been confirmed by our own th erm oluminescent dosimetric data.
AJNR:4 , May / June 1983
CT OF THE HEAD
Like the coronal section, sagittal CT may be especially helpful if
set lateral to metallic c lips, so th at streaks deg rading the image may
be avoided. Thi s can be very important for early detection of tum or
recurrence.
But sagittal CT has its limitation s. Diffi cu lty is experi enced w ith
large or obe se patients and those w ho have inflexible ce rvical
spines because of preexisting degenerative pathology. Also, amalgam fillings and other metallic tooth implantations may cause streak
artifacts that detract from the image q uality. Fortunately , the inc isions usuall y have fewer implants th an the molars.
Editor 's Note: Stru ctures demon strated here by d irect sag ittal
section are also often visualized by multiplanar reconstructed images on high-resolution in strum ents.
REFERENCES
1. Bluemm RG . Direkte sagittale Computertomographi e des Hirnschaedels. Methodik und erste Ergebnisse . Dtsch Med Wochenschr 1981 ; 106 : 1416-141 7
2. Bluemm RG. Direct sagittal (position al) comp uted tomography
of the head. Neuroradiology 1982;22 : 199-201
3. Bluemm RG . Direct sagittal comp uted tomography of the head .
Presented at th e 6th Carvat international meeting on body CT
and new imaging, Rome , February 1982
4. Bluemm RG . Erg ebnisse der sagittalen Computertomographie
des Schaedels. Sci entific exhibit. Presented at the annual
meeting of the German Radiological Society, Berlin, May 1982
5. Bluemm RG , Akuamoa-Boateng E, Dieckmann J. Direkte sagittale Computertomographie des Gesichtssch aedels. Methodik
und erst klinische Ergebnisse . Presented at the annual meeting
of the German Society of Oral and Maxillo-Facial Surgery,
Lu ebeck-Travemuende, May 1982
6. Bluemm RG , van Waes PFGM . Direct sagittal computed tomography of the base of the skull. Presented at the 9th International Congress of Radiology in Otolaryngology, Fontevraud,
487
France, June 1982
7. Bluemm RG . Direct sagi ttal co mputed tomography in ch il dren.
Presented at th e 8th meeting of the European Society of
Pediatri c Surgery, Rennes, France, June 1982
8. Mass S, Norman D, Newton TH . Coronal compu ted tomog raphy: indicati ons and accu racy. AJR 1978;13 1 : 875-879
9 . Nakag awa H, Wolf BS. Delineation of lesions of the base of th e
sku ll by co mputed tomography. Radiology 1977;1 24: 75-80
10. Aubin ML, Bentson J, Vign aud J. Tomodensitom etrie de la tig e
pituitaire. J Neuroradiol 1978;5: 1 53-160
11 . Clar HE, Bock JW, Weic hert HC . Comparison of encepha lotomograms and co mputerized tomograms in mid line tumors in
infants. Acta Neurochir (Wien) 1979;50 : 91-1 01
12. Earn est NF, McCullough EC , Frank DA. Fact or artifact: an
analys is of artifact in high-resolution co mpu ted tomographi c
scanning of th e sella. Radiology 1981 ; 140: 1 09-1 13
13. Syvertsen A, Haugh ton VM, Williams AL , Cusic k JF . The computed tom og raphic appearance of the normal pituitary gland
and pituitary microadenomas. Radiology 1979;133 : 385-39 1
14. Glover GH , Pelc NJ . Nonlinear partial-volume artifacts in x-ray
co mputed tomography . Med Ph ys 1980;7: 238-248
15. Glanz S, Geehr RB , Dunca n CC, Pi epmeier JM . Metrizamideenhanced CT for evaluati on of brainstem tumors . AJNR
1980;1 :3 1-34;AJR 1980;134:821 - 824
16. Baker HL. The c linica l usefu lness of routine corona l and sagittal reconstru ction s in c ran ial co mputed tomography . Radiology 1981 ;140 : 1-9
17 . Bluemm RG , Gehlen W. Aspects of direct (positional) sag ittal
CT-scanning of th e brain with respect to c li nical findin gs. Acta
Neurochi (Wien) 1982;66: 2 13-220
18. Delavelle J, Megret M . CT sag ittal reconstruction of posteri or
fossa tum ors . Neuroradiology 1980; 19 : 8 1-88
19. Kronholz HL, Castrup W , Glasmeier KH . Zur Strahlenbelastun g
von Kindem bei Schaedeluntersuchungen an Computertomographen der neuen Generation. Presented at the annual meeting of the German Radiological Society, Munich, May 1981