Ultrasound

Cerebral Ventricular Size in Children

A Two-Dimensional Ultrasonic Study1

William J. Garrett, M.D., D.PhiI., George Kossoff, M.E., B.Sc., and Peter S. Warren, MB., M.R.A.C.R.

The width of the bodies of both lateral ventricles was studied in 833 children using B-mode
echography. It could be accurately measured in 99.9 % of those under the age of 2, 94 % of
those 2 to 5, and 65 % of those 6 to 12. The normal lateral ventricular ratio or LVR (width of
body of LV/O.5 X internal skull diameter) was 0.24-0.36, but there was a borderline area
and the upper part of this range (0.34-0.36) included some cases of early hydrocephalus.
In 72 % of the children with a head circumference greater than the 98th percentile, the LVR
was 0.24-0.36. The greatest transverse diameter of the atrium of the normal lateral ventricle
in children appears to be 10 mm, that of the normal third ventricle 5 mm. A water-delay sys-
tern is described as being superior to a contact system for B-mode echoencephalography.

INDEX TERMS: Brain, ultrasound studies, 1 [01.1298 Children,

#{149} central nervous system Infants,
#{149} central nervous
system (Skull
#{149} and contents, congen. hydrocephalus, i[0].145)

Radiology 136:711-715, September 1980

T HE attenuating and refracting properties of the skull

seriously interfere with the transmission of ultrasound,
making static B-mode examination unsatisfactory in adults
for all practical purposes. Real-time examination is more
successful, since the circle of Willis and other moving
landmarks are identified more easily (1 1). However, this
limitation does not apply to infants and young children,
in whom the ventricular system has been outlined and
measured (1 2). Considerable
, cerebral detail may be
displayed, and such structures as the thalamus, the head
of the caudate nucleus, and the aqueduct can be identified
(7). We wanted to correlate echographic quality with age
and measure the cerebral ventricles in younger pa-
tients.
Fig. 1 . Measurement of the lateral ventricles. The lateral walls are
seen as lines parallel to the midline. LVR = A/B; C = atrium.
MATERIALS AND METHODS

A total of 833 children (1,063 examinations) were are inadequate. By contrast, with
the child liesthe Octoson
studied using the Commonwealth Acoustic Laboratories comfortably on a polythene the water
membrane covering
contact echoscope with a hand-held focused 3-MHz tank which contains the transducers. The water-delay
transducer (1 .5 cm in diameter, 9-cm radius of curvature) system, with the transducer standing back from the patient,
in the first half of the series and the Ultrasonics Institute offers three major advantages: (a) the focal zone of the
Octoson, a semi-automatic water-delay system with eight transducers can be placed in the region of interest to op-
focused 3-MHz transducers (7 cm in diameter, 32.5-cm timize resolution; (b) near-field problems are avoided, so
radius of curvature) (6) in the second half. that the child need only be examined from one side; and
Compound scans were routinely taken parallel to Reid’s (C) the child is not disturbed when the transducer arm deep
base line and in the coronal plane. Recently, a third series in the tank changes position, thereby generally avoiding
angled 15#{176}
to Reid’s base line was also taken to display anesthesia. The water is a comfortable 37#{176}C,and there
the posterior fossa in a transverse plane. Owing to difficulty is no heat loss even with the smallest babies. The posterior
with the contact system in the near field, the examination fossa is clearly displayed, as the ear is covered with the
has to be conducted from both sides of the skull, causing contact medium and allows sound transmission.
patient irritation and requiring anesthesia with ketamine The widths of the bodies and atria of the lateral ventri-
or a similar drug in most cases. Skin contact is lost over des were measured on transverse sections (Fig. 1). The
the ear, and as a result scans of the posterior cranial fossa body of the lateral ventricle was measured near the mid-

1 From the Department of Diagnostic Ultrasound, Royal Hospital for Women, Paddington, N.S.W., Australia (W.J.G., P.S.W.), and the Ultrasonics
Institute, Millers Point, Sydney, N.S.W., Australia (G.K.). Received Oct. 30, 1979 and accepted Jan. 9, 1980. sjh

711
712 WILLIAM J. GARRETT AND OTHERS September 1980

Fig. 2. Ultrasonic echograms of the neonatal brain. a. Transverse section through the anterior horns ( 1), bodies (2),
and atria (3) of the lateral ventricles, showing prominent high-level echoes from the choroid plexuses (4).
b. Transverse section through the foramina of Monro (5), showing the third ventricle (6), posterior horns of the lateral
ventricles ( 7), thalamus (8), sulcus over the insula (9), and head of the caudate nucleus ( 10).
C. Transverse section at the level of the colliculi, showing the cerebral peduncles ( 1 1), aqueduct( 12), opening in the tentorium
( 13), and the posterior ( 7) and inferior horns ( 14) of the lateral ventricles.
d. Coronal section through the anterior horns of the lateral ventricles, showing the sulcus over the insula (9), head of the
caudate nucleus ( 10), and other structures of the basal ganglia. The central low-level echoes are emanating from the pons
( 15).
e. Coronal section through the foramina of Monro, showing high-level echoes from the choroid plexuses (4). The septum
pellucidum ( 16), third ventricle (6), and thalamus (8) are displayed.
f. Coronal section through the thalami, showing the bodies of the lateral ventricles (2), massa intermedia ( 17), and lower
portion of the third ventricle (6).
Vol. 136 CEREBRAL VENTRICULAR SIZE IN CHILDREN 713 Ultrasound

3a-c

Fig. 3. 3-year-old girl with shunted communicating hydrocephalus. a. Transverse section through the dilated atria of
the lateral ventricles, showing a porencephalic defect (P) resulting from penetration of the ventricular contents into the brain,
a gap in the septum pellucidum (SF’), and the position of the shunt tube (5).
b. Coronal section through the third ventricle (TV), thalamus (1), and brainstem (BS), showing a high-level echo from
the shunt tube (5) in the left lateral ventricle.
C. Coronal section showing the greatly dilated atria (A) of the lateral ventricles, tentorium cerebelli (TC), and contents
of the posterior fossa (P19.

point of its anteroposterior diameter, which is easier to TABLE I: EFFECT OF SKULL THICKNESS ON VISUALIZATION OF
THE LATERAL VENTRICLES
determine in transverse sections than in the coronal plane.
We also calculated the lateral ventricular ratio (LVR), which Age(yr.) 0-1 2-5 6-12
is the width of the body of the lateral ventricle divided by No.ofchildren 780 216 51
half the greatest internal transverse diameter of the cal- Both lateral ventricles not seen 1 13 18
varia. The transverse diameter of the third ventricle and Failurerate 0.1% 6% 35%
the position of the anatomical midline in relation to the skull
were studied in both transverse and coronal planes. TABLE II: LATERAL VENTRICULAR RATIO
In the present series, the children were all referred
- LVR No. of Patients
because of a clinical problem, most commonly relative
macrocephaly. The fact that 70.6 % of the patients were <0.24 10
male is unexplained but may reflect the selection of per- 0.24-0.25 43
0.26-0.29 171
centile charts used by the referring pediatricians. Very few 0.30-0.33 203
subjects were premature. 0.34-0.36 91
0.37-0.39 45
0.40-0.42 41
RESULTS >0.45 89
693
Excluding 14 cases of gross brain disorganization and
. Excluding those children with a history of surgery or meningitis
2 cases in which movement was excessive, both lateral
ventricles were recorded well enough for satisfactory
measurement of the transverse diameters in 1 ,015 out of increased skull thickness but also aggravated by air
1 ,047 examinations. The transverse diameter of the third trapped in the thicker hair of older children. Despite these
ventricle was measured satisfactorily in 1,001 examina- difficulties, the images were adequate for measurement
tions; in 26 of the 46 failures, it was probably 1 mm or less of both lateral ventricles in two-thirds of the older group
and therefore difficult to assess. Skull thickness did not (Fig. 4; TABLE I).
interfere significantly with ventricular measurement in In a clinical series, it is difficult to give a categorical
children under the age of 6 (TABLE I). statement as to the normal range of any measurement. In
In addition to the ventricular outlines, good anatomical 693 patients (excluding children with a history of surgery
detail of brain tissue was generally obtained in the first two or meningitis), the distribution of results for estimation of
years of life (Fig. 2). In the 2- to 5-year-old group, the the LVR suggests that the normal range is 0.24-0.36, while
ventricles were displayed well but there was some loss of the upper part of this range (0.34-0.36) is a borderline area
soft-tissue detail. Where hydrocephalus or a lesion was containing both normal subjects and cases of mild hy-
the clinical problem, the slightly diminished detail did not drocephalus (TABLE II). The earliest signs of hydrocephalus
invalidate the method (Fig. 3). In the 6- to 12-year-old appeared to be fullness and rounding of the atria of the
group, there was considerable variation, due partly to the lateral ventricles, which sometimes occurred before there
 714 WILLIAM J. GARRETT AND OTHERS September 1980

Fig. 4. 62-year-old girl with shunted hydrocephalus. a. Transverse section angled 15#{176}
to Reid’s base line, showing
the bodies of the lateral ventricles (LV) and position of the shunt tube (S) passing through a gap in the septum pellucidum.
Note the blurred inner table of the near parietal bone at this age.
b. Section parallel to and 4 cm below Figure 4, a, showing the anterior horns (AH), foramen of Monro (FM), and third ventricle
(TV).
c. Coronal section showing the bodies of the lateral ventricles and septum pellucidum (SF’) but no detail of the soft tissue
at this age.

TABLE Ill: SIZE OF THE ATRIA OF THE LATERAL VENTRICLES TABLE IV: HEAD CIRCUM FERENCE > 98TH PERCENTILE
IN 378 PATIENTS
Boys Girls
Width of Atrium (mm) LVR Left Right Left Right
LVR 8 9-10 11-12 13 Total
0.24-0.36 71 71 38 36
0.24-0.33 343 65 48 36 492 0.37-0.45 13 14 1 3
0.34-0.36 45 19 9 27 100 >0.45 11 10 15 15
0.37-0.39 10 4 6 37 57 Total A:
0.40 5 10 8 84 107
756
of boys in this study, there were more girls than boys with
TABLE V: TRANSVE RSE DIAMETER OF THE THIRD VENTRICLE an LVR of >0.45 (TABLE IV).
Third Ventricle (mm) The size of the third ventricle is highly variable. Out of
LVR 3 4-5 6-7 8-10 >10 693 patients with no history of surgery or meningitis, the
<0.24 10 maximum transverse diameter of the third ventricle was
0.24-0.33 289 97 16 13 2 5 mm in 89.7% of 508 children whose LVR was in the
0.34-0.36 48 22 10 11 suggested normal range of 0.24-0.36 and 92.5 % of 417
0.37-0.39 14 10 9 8 4
children whose LVR was 0.24-0.33 (TABLE V). A large
0.40-0.45 7 8 7 12 7
>0.45 3 7 16 20 43 suprapineal recess was observed in 23 patients, in 3 of
whom the transverse diameter of the third ventricle was
less than 5 mm.
was any significant increase in the LVR. Out of 378 chil-
dren under the age of 2 with no history of surgery or DISCUSSION
meningitis, 82.9 % of those whose LVR was less than 0.34
had atria 10 mm or less in transverse diameter (TABLE Ill). In assessing the size of the lateral ventricles in children,
From these observations, it appears reasonable to regard absolute figures expressed in millimeters are less useful
10 mm as the upper limit of normal. than ratios related to the overall size of the brain. The
Since hydrocephalus is commonly associated with ventricle index proposed by Schiersmann for use with
macrocephaly, patients with macrocephaly may be sus- pneumoencephalography divided the bony biparietal di-
pected of having hydrocephalus. In this series, there were ameter by the maximum width of the combined bodies of
149 children whose
head circumference was greater than the lateral ventricles as seen on the radiograph (9). Kazner
the 98th percentile. Of these, 75 % of the boys and 69% and Hopman (5) modified Schiersmann’s index for use with
of the girls had an LVR in the suggested normal range of A-mode ultrasound by dividing the width of the combined
0.24-0.36. Severe hydrocephalus was noted in 1 1 % of lateral ventricles by the external biparietal diameter of the
the boys and 28 % of the girls. Despite the preponderance head as obtained with calipers and named the result the
 Vol. 136 CEREBRAL VENTRICULAR SIZE IN CHILDREN 7i5 Ultrasound

celIa media index or CMI. In i967, SjOgren (iO) defined a REFERENCES

lateral ventricular index (LVI) which is essentially the re-
ciprocal of Kazner and Hopman’s CMI, and this was used 1. Garrett WJ, Kossoff G: Grey scale examination of the brain in
children. [In] White 0, Brown RE, ed: Ultrasound in Medicine.
by Mostafawy (8). Both the LVI and CMI, which were de- New York, Plenum, 1977, Vol 3A, pp 821-828
signed for A-mode studies, include the thickness of the 2. Garrett WJ, Kossoff G, Jones RFC: Ultrasonic cross-sectional
skull, thereby lessening their accuracy. In the present visualization of hydrocephalus in infants. Neuroradiology 8:
279-288, 10 Mar 1975
series, the widest internal diameter of the calvaria was
3. Grumme T, Kazner E: Echoencephalography: the normal
used in calculating the lateral ventricular ratio, corre- echoencephalogram. [In] de Vlieger M, Holmes JH, Kazner E,
sponding closely to the biparietal diameter of the brain et al: Handbook of Clinical UltraSOUnd. New York, Wiley, 1978,
before it has become compressed by a subdural hema- Chapt 61, pp 649-656
4. Haber K, Wachter RD, Christenson PC, et al: Ultrasonic evalu-
toma or other lesion. Use of a contact echoscope may ation of intracranial pathology in infants: a new technique.
result in an error on the order of 2 % , since the measure- Radiology 134:173-178, Jan 1980
ment has to be made in two parts from the midline to avoid 5. Kazner E, Hopman l+ Possibilities and reliability of echoventric-
the near-field artifact, and gravity may deflect the midline ulography. Ultrasound Med Biol 1:17-32, Sep 1973
6. Kossoff G, Carpenter DA, Radovanovich G, et al: Octoson: a
slightly. Likewise, with the Octoson an error on the order new rapid multi-transducer general water-coupling
purpose
of 2 % may result in older children, who have thicker skulls, echoscope. [In] Kazner E, de Vlieger M, Muller HR. et al, ed:
so that the echo from the near parietal bone may diffuse Proceedings of the Second European Congress on Ultrasonics
on the film, blurring the edge of the inner table. In plagio- in Medicine, Munich, 12-16 May 1975. International Congress
Series No. 363. Amsterdam, Excerpts Medica, 1975, pp 90-
cephaly, half of the internal biparietal diameter is taken 95
bilaterally at right angles to the midline and the mean is 7. Kossoff G, Garrett WJ, Radovanovich G: Ultrasonic atlas of
calculated. Despite these difficulties, the system is very normal brainof infant. Ultrasound Med Biol 1:259-266, Aug
1974
satisfactory and the LVR is expressed as relating tQ one
8. Mostafawy A: Pediatric SOnOencephalOgraphy. Bc in, Springer,
side at a time rather than taking the two lateral ventricles 1971, pp 22-25
together as with the LVI or CMI, which is particularly 9. Schlersmann 0: ElnfUhrung In die Enzephalographie. Leipzig,
worthwhile when serial postoperative examinations are Thieme, 1942 [quoted by Kazner and Hopman (5)]
10. SjOgren I: Echoencephalography In paediatric practice with
undertaken. special regard to measurement of the ventricular size. Acta
Our figure of 5 mm as the upper limit of normal for the Paediatr Scand (Suppl 178): 1-67, 1967
third ventricle agrees with published accounts (3). Care 11 . Smith SW, Phillips DJ, von Ramm OT, et al: Real time B-mode
should be taken to distinguish the ventricle from the cavum echoencephalography. [In] White DN, Barnes R, ed: Ultrasound
In Medicine. New York, Plenum, 1976, Vol 2, pp 373-382
septi pellucidi or cavum vergae in premature infants 12. Warren PS, Garrett WJ, Kossoff G: High resolution imaging of
(i2). the neonatal brain. [In] Proceedings of the 24th Annual Meeting
Interference with sound transmission by the skull is not of the American Institute of Ultrasound In Medicine and the 8th
Annual of the American Society of Ultrasound Technical
Meeting
a great problem in children under the age of 6, permitting
Specialists, August 27-31, 1979, Montreal, Canada. Oklahoma
B-mode ultrasound to make a contribution in this age group City, OkIa., AlUM, 1979, Vol 1, p 104
not only in providing anatomical information but also in
avoiding excessive ionizing radiations. In the present se-
. ries, the Octoson was far preferable to the contact system Department of Diagnostic Ultrasound
using a hand-held transducer applied directly to the scalp. Royal Hospital for Women
With its water-delay system, the Octoson gives excellent 188 Oxford St.
Paddington, 2021
tissue detail and is particularly suited to pediatric practice N.S.W.
(4). Australia