University of ZurichZurich Open Repository and Archive

Winterthurerstr. 190

CH-8057 Zurich

http://www.zora.uzh.ch

Year: 2011

Cerebral measurements made using cranial ultrasound in termUgandan newborns

Hagmann, C F; Robertson, N J; Acolet, D; Nyombi, N; Ondo, S; Nakakeeto, M;Cowan, F M

http://www.ncbi.nlm.nih.gov/pubmed/21353402.Postprint available at:http://www.zora.uzh.ch

Posted at the Zurich Open Repository and Archive, University of Zurich.http://www.zora.uzh.ch

Originally published at:Hagmann, C F; Robertson, N J; Acolet, D; Nyombi, N; Ondo, S; Nakakeeto, M; Cowan, F M (2011). Cerebralmeasurements made using cranial ultrasound in term Ugandan newborns. Early Human Development,87(5):341-347.

http://www.ncbi.nlm.nih.gov/pubmed/21353402.Postprint available at:http://www.zora.uzh.ch

Posted at the Zurich Open Repository and Archive, University of Zurich.http://www.zora.uzh.ch

Originally published at:Hagmann, C F; Robertson, N J; Acolet, D; Nyombi, N; Ondo, S; Nakakeeto, M; Cowan, F M (2011). Cerebralmeasurements made using cranial ultrasound in term Ugandan newborns. Early Human Development,87(5):341-347.

Cerebral measurements made using cranial ultrasound in termUgandan newborns

Abstract

These data provide reliable reference values for linear measurements of many cerebral structures madeusing cUS. The data agree well with those from other populations suggesting that cerebral size is similarin different ethnic groups.

1

ABSTRACT Now 249

Background: Few cUS studies of cerebral measurements are available for normal

term infants. Normative data is important for evaluating cerebral structure size in

symptomatic term infants and assessing preterm brain growth by term age.

Objectives: to (i) make linear measurements using cranial ultrasound (cUS) for

major cerebral structures and intracranial spaces in normal newborn term infants, (ii)

correlate these measurements with gestational age (GA), birthweight (BW), head

circumference (HC), gender and within one infant (iii) examine inter/intra-observer

variation, (iv) compare these data with those currently available.

Design, setting and patients: Linear cUS measurements of major cerebral

structures were made in well term-born Ugandan infants at Mulago University

Hospital, Kampala. Correlations between the measurements and gender, HC, BW

and GA were calculated. Intra- and inter-observer agreement was assessed.

Results: Data from 106 infants (mean GA 39.20 ±1.4SD weeks) were analysed.

Intra/inter-observer agreement was substantial/excellent. Significant correlations

were found between HC and pons anterior-posterior diameter (p<0.01), corpus

callosal (CC) length (p=0.02) and transverse cerebellar diameter (TCD, p<0.01) and

between BW and CC length (p=0.02), vermis height (<0.01) and thalamo-occipital

distance (p=0.03); no significant correlation was found with GA. Within infants CC

length and TCD correlated significantly (p=0.019). Males had larger left ventricular

indices than females (p=0.04). The data was similar to those from other populations.

Conclusions: These data provide reliable reference values for linear measurements

of many cerebral structures made using cUS. The data agree well with those from

other populations suggesting that cerebral size is similar in different ethnic groups.

2

Introduction

Cranial ultrasound (cUS) is usually the first modality used for imaging the neonatal

brain both in symptomatic term infants and most preterm infants. However reference

ranges for linear measurements of cerebral structures from a large cohort of well

term newborn infants have not been published; nor has the relation between these

measurements and gender, birthweight (BW) and head circumference (HC) been

examined. Such normative data is important for evaluating both the size of cerebral

structures in symptomatic term infants and assessing the growth of the preterm brain

by term age.

Few cUS studies of cerebral measurements are available for normal term infants.1-4

There are data from normal fetuses,5-12 and from preterm infants3, 13-17 but these latter

cannot be taken as normative even when the cohorts comprise only low-risk infants

without overt lesions on their scans. Most studies focus on measuring one brain

structure and no study has published reference measurements of many cerebral

structures in normal term infants (Table 1e, appendix).

The aims of this study were (i) to determine linear measurements from cUS for major

cerebral structures, ventricles and extra-axial spaces in a large cohort of well and

neurologically normal newborn term infants, (ii) to evaluate the correlation between

these measurements and gestational age (GA), BW, HC, gender and within one

infant (iii) to assess intra-and interobserver variability for the measurements.

Additionally we wished to know whether there were any differences between

published data and the data from our Ugandan cohort.

3

Methods

The Institutional Review Board of the Ethics Committee, Medical School, Makerere

University, Kampala, Uganda approved the study; permission from the Head of the

Obstetrics Department at Mulago Hospital was also obtained.

Patients

Between 24th July and 31st October 2007 well term infants were recruited from the

postnatal wards at Mulago University Hospital, Kampala, Uganda as controls for a

pilot study of therapeutic hypothermia in a low resource setting18 in order to provide

normative cUS scan data for comparison with Ugandan infants affected by hypoxic-

ischaemic encephalopathy. Maternal informed consent was obtained by the ward

nurses in the mother’s own language and confirmed by the study doctors. Inclusion

criteria were (i) direct admission to the postnatal ward from labour ward, (ii) GA ≥36

weeks (iii) Apgar score ≥8 at 5 minutes, (iv) ≤4 days postnatal age. Antenatal and

perinatal clinical details were obtained from obstetric notes and the mother. Nurses

familiar with the local languages were available throughout; the infants were scanned

in a postnatal ward side-room with the mother present if she wished. They were also

examined neurologically using a standardised protocol.19 HC was measured using a

disposable paper tape measure. BW and HC centiles were calculated using the LMS

growth calculator (www.healthforallchildren.co.uk) from WHO2006 and British 1990

reference data.20

Cranial ultrasounds (cUS)

We used a portable ultrasound machine (z.oneUltra Convertible UltrasoundTM

System, Zonare Medical Systems Inc., USA) with the P10-4 Phased Array

Transducer (bandwidth 4-10 MHz with tissue and compound harmonics of 8MHz).

The probe was cleaned with alcohol wipes between patients; disposable gloves were

worn when scanning. FC, CH, DA, NN and NM performed the examinations. We

obtained minimally five coronal, one mid-sagittal, two left and two right parasagittal

4

views. cUS data were stored digitally on the cUS system and later downloaded and

analysed later by FC and CH, from DICOM formatted images using OsiriX 3.5.1

software (Geneva, Switzerland).

Both neurological and cUS examinations were done by FC, DA, CH, NN and NM.

FC, DA and CH worked together before and are experts in these methodologies. NN

and NM were trained by FC and DA and later supervised by CH.

Linear measurements20 were made on the following structures and the planes 21

chosen as previously described. Detailed description of the planes and

measurements are given in the appendix (Table 2e, appendix).

1. mid-sagittal plane: length of the corpus callosum22 (CC), height of cerebellar

vermis22 and cisterna magna22 (CM) beneath the cerebellar vermis, anterior-

posterior (AP) diameter of the pons22 and depth of 4th ventricle22 (Figure 1a)

2. parasagittal plane: thalamo-occipital distance22 (TOD) and cortical depth of the

cingulate sulcus (Figure 1b)

3. mid-coronal plane (through the foramina of Monro): ventricular index14 (VI) width

and height of the cavum septum pellucidum22 (CSP), thickness (height) of the

CC, width of 3rd ventricle (Figure 2a), width of the basal ganglia22, diagonal width

of the caudate head22, extracerebral space22 (ECS) and interhemispheric fissure

(IHF) 22 (Figure 2b)

4. posterior coronal plane through the thalami and maximal diameter of the

cerebellum: maximum transverse cerebellar diameter22, 23(TCD) (Figure 2c)

Data analyses

Statistical analyses were performed with SPSS 14 using parametric or non-

parametric tests depending on data normality. Paired t tests were used to evaluate

differences between right and left measurements. Linear regression analyses were

5

performed to explore associations between the measurements and GA, BW, HC,

gender and laterality. The level of significance was taken at p0.05. Inter-and

intraobserver agreement was assessed using the intraclass correlation coefficient

(ICC) from 15 infants in whom FC and CH independently and FC twice made the

measurements.Agreement is substantial when the ICC is between 0.6 and 0.8 and

excellent when >0.8.

6

Results

Of the 111 infants eligible for the study five were excluded either due to evidence of

infarction or haemorrhage21 or due to poor image quality. No significant differences in

measurements were found between infants with or without minor scan abnormalities

(choroid plexus or subependymal cysts, mildly increased white matter or basal

ganglia echogenicity, lenticulostriate vasculopathy24), therefore data from these

remaining 106 infants was pooled and analysed.

The demographic details of the 106 infants are shown in Table 3 and were similar in

males and females. The mean HC centile was significantly higher than the mean BW

centile (p<0.01). Using the WHO growth reference data the HC of 5% of the infants

was <10th centile (3% <3rd centile), 95% of infants were between the 10th-97th centile

and 5% were >97th centile. The distribution of the BW centile was 13% <10th centile

(5% <3rd centile), 82% were between the 10th-97th centile and 5% were >97th

percentile. Male infants had a significantly larger mean HC (35.21cm1.17) than

females (34.6cm1.27, p=0.023). The comparison between the centiles calculated

with British or WHO references are given in Table 3. BW correlated with GA

(B=0.10, p=<0.01, 95%CI 0.036 -0.17) and HC (B=0.15, p=<0.01, 95% CI 0.07-0.23).

No correlation was found between HC and GA.

Linear measurements

The values for all cUS measurements are given in Table 4. There were no significant

differences between left and right hemispheric measurements using a paired t test,

so where appropriate, mean data from both hemispheres was used.

7

The CSP could be measured in 66 (64%) infants, the median width being 3.66mm

(0.61-8.37). When the 40 infants in whom the CSP was not patent were included the

median width was 3.05mm (0-8.37). The 90th centile was 6.7mm.

The ECS and IHF were seen and measurable in only 34 (33%, median 1.30mm

(0.36-2.2)) and 55 (53%, median 1.68mm (0.46-3.76)) infants respectively. If the

infants in whom ECS and IHF was not visible, were included as 0mm, then the

median values were 0.00mm (range 0.0-2.2) and 0.88mm (range 0.0-3.76)

respectively.

Differences in relation to gender

The only small but significant difference found was for the left VI, which was larger in

male infants (mean 11.75mm vs 11mm, p=0.04).

Correlation with GA, BW and HC and within infants

No significant correlation was found between GA and any of the measurements. The

TCD, pons AP diameter and CC length correlated significantly with HC also after

correction for BW, GA and age at scanning (Table 5e, appendix). CC length, vermis

height and TOD correlated significantly with BW. Within infants, there was a

significant correlation between CC length and vermis height, pons width, TCD, and

VI on univariate analysis. When corrected for HC, only the correlation between the

largest structures, the CC and TCD remained statistically significant (p=0.019).

Inter-and intraobserver agreement

The intra-and interobserver agreements were substantial to almost perfect (Table 6e,

appendix).

8

Discussion

Our study provides normative linear measurements for many cerebral structures

made using cUS soon after birth in normal term Ugandan African infants. We had

substantial to almost perfect ICC for most measurements. Agreement was worst for

the VI likely because the anterior horns of many lateral ventricles were barely visible

making their borders difficult to distinguish.25, 26

Surprisingly, none of the linear measurements correlated with GA. This may be

because most infants were in a narrow GA range but the reliability of GA assessment

may not be optimal.27, 28 CC length, TCD, AP pons width correlated with HC whilst CC

length, vermis height, TCD correlated with BW. Though BW correlated with GA and

HC, HC did not correlate with GA. The median HC centile was significantly larger

than the median BW centile, although both were within the 10th to 97th range and

similar to previously published Ugandan29 and African data.30, 31 There may be

relative growth restriction in this population accounting for the larger HC centile. We

expected a better correlation between the measurements and HC than BW but this

was not the case. We did not find significant differences in BW and HC centiles using

the WHO2006 or British reference data. Unfortunately the WHO2006 reference does

not allow adjustment for GA at birth.

The CC is a major cerebral structure that is easily measured on cUS. Poorer growth

and microstructure of the CC in preterms has been related to poor

neurodevelopmental outcome. 32, 33, 34Fetal reference ranges for normal CC length

and thickness, established with transvaginal ultrasonography in a mid-sagittal plane5,

8 near to term age are similar to our measurements and those of Leijser et al.22 The

strength of our data is that it is large cohort of normal term infants.

9

Cerebellar injury may be responsible to some extent for the high incidence of

dysfunction found in infants born preterm.35 It is only recently that the cerebellum has

been better imaged using cUS and cognisance taken of abnormal appearances,

though rarely are measurements made to document normal growth. Fetal ultrasound

studies established measurements of normal vermis size in the mid-sagittal7, 36, 37 and

TCD in the axial plane.37 Reference growth curves for cerebellar vermis height were

made in a postnatal study of newborn infants born from 25-42 weeks GA finding a

strong association with GA, possibly independent of gender.38 Our vermis

measurements are similar to those from the term infants in that study. Makhoul et al

showed that TCD is linearly correlated with GA and HC in appropriate and

symmetrically small infants and seemed to be preserved in the majority of

asymmetrically grown small infants23. In our population, no correlation was found

between GA or gender and TCD but there was a strong correlation with HC. This

maybe explained by the narrow range of GA in our cohort or some inaccuracy in the

estimate of GA. We measured the TCD in the coronal plane through the anterior

fontanelle23 and not through the mastoid fontanelle in an axial plane as it is done in

fetal studies and recommended if cerebellar injury is suspected. However, for

assessing size our experience is that the measurements are similar in both planes

and in a recent study of TCD measurements made through both fontanelles in 60

preterm infants no significant differences were found.39

The CM should be seen beneath the vermis and if not this suggests the cerebellum

may be too low and/or the posterior fossa is small. CM enlargement is associated

with cerebral and spinal abnormalities and aneuploidy and is a marker for cerebellar

hypoplasia. The mean CM size in our study was 3.27±1.1mm a little less than

reported in two other studies where the mean was nearer 4.5mm.4,40 The 90th centile

from our data was 4.85mm concordant with the proposal that a CM exceeding 7.2mm

(preterms) and 6.5mm (term infants) maybe associated with pathology4. An isolated

10

enlarged CM detected on fetal screening may be associated with later subtle

developmental problems.40

The CSP was clearly patent in 65% of our infants. Faruggia et al found a patent CSP

in 50% of term infants41 and Mott et al could see the CSP in all infants <36 weeks GA

but in only 36% by 40 weeks GA.15 The mean CSP width (5.7mm1.9) at 40 weeks in

their study was larger than in ours (3.961.65), but the mean heights were similar. In

an autopsy study of infants thought not to have CNS abnormality a patent CSP was

found in 97% of term infants <1 week old.42 This high incidence may be because

even a 1mm gap was included, likely below the level discernable using cUS. A CSP

seen neonatally probably represents a normal variant but when large (or absent

prenatally) is thought to be a marker of abnormal neurodevelopment.43-46 The largest

CSP width we found was 8.37mm (a well 38 week GA infant) well over the 90th

centile of 5.77mm when infants in whom the CSP was not patent were included. Our

data suggest that cognisance should be taken of a CSP over 6-7 mm in symptomatic

term infants.

The mean AP pons width in our study was 16.861.75mm compared to

14.941.96mm in the study of Leijser et al22 probably because their population

included some preterm infants. Measurement of the pons is rarely done22 but it is not

difficult 47 and useful in discerning developmental hypoplasias and swelling in acute

injurious situations.

Levene et al established the VI by GA in 273 term and preterm infants.14 The mean

VI (11.3±1.64mm) of our cohort is consistent with Levene’s term infant data. The left

VI was significantly larger in males than females in our cohort consistent with data

reporting larger left ventricles26 and male infants having larger ventricles than

11

females.13 Davies et al13 reported on anterior horn, 3rd and 4th ventricular width and

TOD measurements in preterm infants (23-33 weeks GA) and their mean TOD

measurements were larger in males than females as we found at term though the

mean TOD in our study was less (12.56±2.36mm,range 7.38-20.01) than in their

preterm cohort (16.7±4mm,range 8.7-24.7). However, preterm infants are known to

have larger posterior horns of the ventricles compared to healthy term controls. 48 We

did not find any correlation between the TOD and GA which is in agreement with

Davies et al.13 Significant but very small differences (1-2mm) are found between cUS

and MRI TOD measurements.22 This difference is most likely explained by the

difficultly in obtaining the correct plane for TOD measurement on standard sagittal

MRI images.22, 49

The ECS and IHF were very small and often not measureable. The mean mid-frontal

sub-arachnoid space (mean 1.320.48mm) did not exceed 4mm, the 95th centile

found by Libicher et al.3 In newborn preterms Armstrong et al17 found the

subarachnoid space to be <3.5mm in 95%, very similar to the data from Frankel et

al50 but larger than in our term infants. There was no significant gender ECS or IHF

difference in our cohort to explain the larger HC in male infants.

Knowledge of the normal range of cortical thickness on cUS is limited. Although only

small regions of cortex are amenable to measurement via the anterior fontanelle it is

possible to visualise the central and cingulate sulci. Leijser et al found that the

cortical depth of the cingulate sulcus was significantly and consistently greater

(2.00mm0.28; mean cUS-MR difference 0.69mm) than on MRI22 but our

measurements (mean 1.22mm0.28) were similar to their MR measurements. There

are few existing data for the size of the caudate head or the basal ganglia22 but the

12

central grey matter is the most important tissue likely to be injured in hypoxic-

ischaemic encephalopathy.51, 52

One limitation of the study is the lack of neurodevelopmental follow-up, which would

be desirable but logistically difficult in this population. All our infants had normal 5

minute Apgar scores and a normal neonatal neurological exam.19 We might have

improved visualisation of the ECS and IHF by using a higher frequency probe.3 Slight

uncertainty about GA may explain the lack of correlation with measurements.

In conclusion cUS is easy to use and gives highly reproducible measurements for

many cerebral structures. The values from this African population agree well with

data from smaller studies of individual structures suggesting that cerebral size is

similar in different ethnic groups. Many intracranial spaces are often not visible or

patent yet published normal ranges do not usually include these zero values thus

overestimating the normal range. We suggest that in infants with suspected cerebral

abnormalities the size of cerebral structures can and should be assessed objectively

using cUS; such measurements can also be made in preterm infants at term age for

assessing brain growth.

13

References

1. Govaert P, Pauwels W, Vanhaesebrouck P, et al. Ultrasound measurement of

the subarachnoid space in infants. Eur J Pediatr. 1989 Feb;148(5):412-3.

2. Lam WW, Ai VH, Wong V, et al. Ultrasonographic measurement of

subarachnoid space in normal infants and children. Pediatr Neurol. 2001

Nov;25(5):380-4.

3. Libicher M, Troger J. US measurement of the subarachnoid space in infants:

normal values. Radiology. 1992 Sep;184(3):749-51.

4. Castriota Scanderbeg A, Mulas M, Bisceglia M, et al. [Echographic

measurements in the posterior cranial fossa: the normal values of the term and

premature newborn infant]. Radiol Med. 1997 Nov;94(5):428-32.

5. Achiron R, Achiron A. Development of the human fetal corpus callosum: a

high-resolution, cross-sectional sonographic study. Ultrasound Obstet Gynecol. 2001

Oct;18(4):343-7.

6. Jou HJ, Shyu MK, Wu SC, et al. Ultrasound measurement of the fetal cavum

septi pellucidi. Ultrasound Obstet Gynecol. 1998 Dec;12(6):419-21.

7. Malinger G, Ginath S, Lerman-Sagie T, et al. The fetal cerebellar vermis:

normal development as shown by transvaginal ultrasound. Prenat Diagn. 2001

Aug;21(8):687-92.

8. Malinger G, Zakut H. The corpus callosum: normal fetal development as

shown by transvaginal sonography. AJR Am J Roentgenol. 1993 Nov;161(5):1041-3.

9. Serhatlioglu S, Kocakoc E, Kiris A, et al. Sonographic measurement of the

fetal cerebellum, cisterna magna, and cavum septum pellucidum in normal fetuses in

the second and third trimesters of pregnancy. J Clin Ultrasound. 2003

May;31(4):194-200.

10. Zalel Y, Seidman DS, Brand N, et al. The development of the fetal vermis: an

in-utero sonographic evaluation. Ultrasound Obstet Gynecol. 2002 Feb;19(2):136-9.

14

11. Malinger G, Lerman-Sagie T, Achiron R, et al. The subarachnoid space:

normal fetal development as demonstrated by transvaginal ultrasound. Prenat Diagn.

2000 Nov;20(11):890-3.

12. Goldstein I, Makhoul IR, Tamir A, et al. Ultrasonographic nomograms of the

fetal fourth ventricle: additional tool for detecting abnormalities of the posterior fossa.

J Ultrasound Med. 2002 Aug;21(8):849-56.

13. Davies MW, Swaminathan M, Chuang SL, et al. Reference ranges for the

linear dimensions of the intracranial ventricles in preterm neonates. Arch Dis Child

Fetal Neonatal Ed. 2000 May;82(3):F218-23.

14. Levene MI. Measurement of the growth of the lateral ventricles in preterm

infants with real-time ultrasound. Arch Dis Child. 1981 Dec;56(12):900-4.

15. Mott SH, Bodensteiner JB, Allan WC. The cavum septi pellucidi in term and

preterm newborn infants. J Child Neurol. 1992 Jan;7(1):35-8.

16. Anderson N, Wells E, Hay R, et al. Cerebellar vermis measurement at cranial

sonography for assessing gestational age in the newborn weighing less than 2000

grams. Early Hum Dev. 1996 Jan 5;44(1):59-70.

17. Armstrong DL, Bagnall C, Harding JE, et al. Measurement of the

subarachnoid space by ultrasound in preterm infants. Arch Dis Child Fetal Neonatal

Ed. 2002 Mar;86(2):F124-6.

18. Robertson NJ, Nakakeeto M, Hagmann C, et al. Therapeutic hypothermia for

birth asphyxia in low-resource settings: a pilot randomised controlled trial. Lancet.

2008 Sep 6;372(9641):801-3.

19. Dubowitz L, Mercuri E, Dubowitz V. An optimality score for the neurologic

examination of the term newborn. J Pediatr. 1998 Sep;133(3):406-16.

20. Cole TJ, Green PJ. Smoothing reference centile curves: the LMS method and

penalized likelihood. Stat Med. 1992 Jul;11(10):1305-19.

21. Rennie J, Hagmann C, Robertson N. Neonatal Cerebral Investigation:

Cambridge University Press; 2008.

15

22. Leijser LM, Srinivasan L, Rutherford MA, et al. Structural linear

measurements in the newborn brain: accuracy of cranial ultrasound compared to

MRI. Pediatr Radiol. 2007 Jul;37(7):640-8.

23. Makhoul IR, Goldstein I, Epelman M, et al. Neonatal transverse cerebellar

diameter in normal and growth-restricted infants. J Matern Fetal Med. 2000 May-

Jun;9(3):155-60.

24. Hagmann C, Robertson N, Acolet D, et al. Cranial ultrasound findings in well

newborn Ugandan infants. . Arch Dis Child Fetal Neonatal Ed in press 2010.

25. Nelson MD, Jr., Tavare CJ, Petrus L, et al. Changes in the size of the lateral

ventricles in the normal-term newborn following vaginal delivery. Pediatr Radiol. 2003

Dec;33(12):831-5.

26. Winchester P, Brill PW, Cooper R, et al. Prevalence of "compressed" and

asymmetric lateral ventricles in healthy full-term neonates: sonographic study. AJR

Am J Roentgenol. 1986 Mar;146(3):471-5.

27. Harris RD, Marks WM. Compact ultrasound for improving maternal and

perinatal care in low-resource settings: review of the potential benefits,

implementation challenges, and public health issues. J Ultrasound Med. 2009

Aug;28(8):1067-76.

28. Okunlola MA, Owonikoko KM, Fawole AO, et al. Gestational age at antenatal

booking and delivery outcome. Afr J Med Med Sci. 2008 Jun;37(2):165-9.

29. Simpkiss MJ. Birth weight, maternal age, and parity among the african

population of uganda. Br J Prev Soc Med. 1968 Oct;22(4):234-7.

30. Ebomoyi E, Adetoro OO, Wickremasinghe AR. Birthweight and

sociobiological factors in Ilorin, Nigeria. J Biosoc Sci. 1991 Oct;23(4):417-23.

31. Feleke Y, Enquoselassie F. Maternal age, parity and gestational age on the

size of the newborn in Addis Ababa. East Afr Med J. 1999 Aug;76(8):468-71.

16

32. Rademaker KJ, Lam JN, Van Haastert IC, et al. Larger corpus callosum size

with better motor performance in prematurely born children. Semin Perinatol. 2004

Aug;28(4):279-87.

33. Counsell SJ, Edwards AD, Chew AT, et al. Specific relations between

neurodevelopmental abilities and white matter microstructure in children born

preterm. Brain. 2008 Dec;131(Pt 12):3201-8.

34. Anderson NG, Laurent I, Cook N, et al. Growth rate of corpus callosum in

very premature infants. AJNR Am J Neuroradiol. 2005 Nov-Dec;26(10):2685-90.

35. Parker J, Mitchell A, Kalpakidou A, et al. Cerebellar growth and behavioural &

neuropsychological outcome in preterm adolescents. Brain. 2008 May;131(Pt

5):1344-51.

36. Achiron R, Kivilevitch Z, Lipitz S, et al. Development of the human fetal pons:

in utero ultrasonographic study. Ultrasound Obstet Gynecol. 2004 Oct;24(5):506-10.

37. Sherer DM, Sokolovski M, Dalloul M, et al. Nomograms of the axial fetal

cerebellar hemisphere circumference and area throughout gestation. Ultrasound

Obstet Gynecol. 2007 Jan;29(1):32-7.

38. Pogliani L, Radaelli G, Manfredini V, et al. Height of the cerebellar vermis and

gestational age at birth. Ultrasound Obstet Gynecol. 2008 Apr;31(4):401-5.

39. Graca A, Cardoso K, Cowan F. Assessment of gestational age in very

preterm neonates using cerebellar measurements at cranial ultrasound-what is the

best approach. EAPS; Copenhagen, Denmark2010.

40. Goodwin L, Quisling RG. The neonatal cisterna magna: ultrasonic evaluation.

Radiology. 1983 Dec;149(3):691-5.

41. Farruggia S, Babcock DS. The cavum septi pellucidi: its appearance and

incidence with cranial ultrasonography in infancy. Radiology. 1981 Apr;139(1):147-

50.

42. Shaw CM, Alvord EC, Jr. Cava septi pellucidi et vergae: their normal and

pathogical states. Brain. 1969 Mar;92(1):213-23.

17

43. Bodensteiner JB, Schaefer GB. Wide cavum septum pellucidum: a marker of

disturbed brain development. Pediatr Neurol. 1990 Nov-Dec;6(6):391-4.

44. Bodensteiner JB, Schaefer GB, Craft JM. Cavum septi pellucidi and cavum

vergae in normal and developmentally delayed populations. J Child Neurol. 1998

Mar;13(3):120-1.

45. Schaefer GB, Bodensteiner JB, Thompson JN, Jr. Subtle anomalies of the

septum pellucidum and neurodevelopmental deficits. Dev Med Child Neurol. 1994

Jun;36(6):554-9.

46. Barkovich AJ, Norman D. Absence of the septum pellucidum: a useful sign in

the diagnosis of congenital brain malformations. AJR Am J Roentgenol. 1989

Feb;152(2):353-60.

47. Helmke K, Winkler P, Kock C. Sonographic examination of the brain stem

area in infants. An echographic and anatomic analysis. Pediatr Radiol. 1987;17(1):1-

6.

48. Inder TE, Wells SJ, Mogridge NB, et al. Defining the nature of the cerebral

abnormalities in the premature infant: a qualitative magnetic resonance imaging

study. J Pediatr. 2003 Aug;143(2):171-9.

49. Horsch S, Bengtsson J, Nordell A, et al. Lateral ventricular size in extremely

premature infants: 3D MRI confirms 2D ultrasound measurements. Ultrasound Med

Biol. 2009 Mar;35(3):360-6.

50. Frankel DA, Fessell DP, Wolfson WP. High resolution sonographic

determination of the normal dimensions of the intracranial extraaxial compartment in

the newborn infant. J Ultrasound Med. 1998 Jul;17(7):411-5; quiz 7-8.

51. Voit T, Lemburg P, Neuen E, et al. Damage of thalamus and basal ganglia in

asphyxiated full-term neonates. Neuropediatrics. 1987 Aug;18(3):176-81.

52. Eken P, Jansen GH, Groenendaal F, et al. Intracranial lesions in the fullterm

infant with hypoxic ischaemic encephalopathy: ultrasound and autopsy correlation.

Neuropediatrics. 1994 Dec;25(6):301-7.

18

Acknowledgments

We thank the mothers of the babies and the nurses on the postnatal wards at Mulago

hospital for all their help with this study

Competing interests: None

Funding

This project was initiated as part of the Uganda Women’s Health Initiative (UWHI)

involving University College London, Institute for Women’s Health, Mulago Hospital,

Makerere University and Hospice Africa Uganda. We are grateful to Professor Ian

Jacobs and Dr Anthony Gakwaya Co Directors UWHI and Mr Graham Evans Project

Manager UWHI for their support and guidance. The UWHI and this project were

supported by generous donations from Lee and Roger Myers and Ann-Margaret and

John Walton.

19

Figure Legends

Figure1a, b. Sagittal cranial ultrasound images showing (a) mid-sagittal view of

measurements of the length of the corpus callosum, the height of the cerebellar

vermis and cisterna magna and AP width of the pons and the 4th ventricle and (b) a

parasagittal image showing the measurement of the thalamo-occpital distance and

cortical depth. A small cyst is seen in the choroid plexus.

Figures 2 a,b and c. Coronal cranial ultrasound images showing (a) measurement of

the width and height of the cavum septum pellucidum, height of the corpus callosum,

the ventricular index and the width of the 3rd ventricle (b) diagonal width of the head

of the caudate nucleus, the maximal width of the basal ganglia, the interhemispheric

fissure and the extracerebral space and (c) a more posterior coronal view through

the thalami showing the maximal width of the cerebellum

20

Tables

Table 1e

Relevant references giving normative term data for comparison with data from this

study

Table 2e

Detailed description of planes and measurements

Table 3

Demographic data of the 106 infants including a comparison of birth weight and head

circumference centiles using British1990 and WHO 2006 reference data. Mean birth

weight centile was significantly lower than mean head circumference centile

(p<0.01).

Table 4

Mean and standard deviation cUS measurements. +these data are given as median

values and include as 0 the infants whose interhemispheric fissure and extracerebral

space were too small to be measureable and whose cavum septum pellucidum was

closed.

Table 5e

Linear regression showing correlation between cUS measurements and head

circumference and birth weight.

Table 6e

Intercorrelation coefficients (ICC) of 8 different structural measurements showing

substantial to almost perfect intra-and interobserver agreement

Table 1e. Relevant references giving normative term data for comparison with data from this study

Reference (no) Age at measurement

Cerebral structure Mean (SD) in mm or otherwise stated

Term infants Govaert et al (1) 40 weeks to 13

months Subarachnoid space <0.2cm

Lam et al (2) Term to one year At term age: Subarachnoid space Interhemispheric fissure Craniocortical width

2.2mm 2.8mm 2.4mm

Libicher et al (3) 1 to 12 months Subarachnoid space Median 1.0, 5th centile 0.6, 95th centile 3.5mm

Castriota Scanderbeg et al (4) Term Cisterna magna Term infants: 4.5(1.0) Fetal studies up to term age Achiron et al (5) 16 to 37 weeks Length of Corpus callosum 27.2(1.2) Jou et al (6) 19 to 42 weeks Width of CSP Malinger et al (8) 18 to 42 weeks At 40-42 weeks:

Length of corpus callosum 44.0(3.8)

Serhatlioglu et al (9) 16 to 38 weeks Third trimester. Width of CSP

5.0(1.4)

Malinger et al (7) 21 to 39 weeks At 39-42 weeks: Cerebellar cranio caudal diameter

25.7(2.3)

Zalel et al (10) 18 to 38 weeks At 37 to 38 weeks: Height of vermis

15.3(0.87)

Malinger et al (11) 16 to 40 weeks Subarachnoid space Craniocortical width

3.0(0.9) 3.3(1.0)

Goldstein et al (12) 13 to 40 weeks Anterioposterior diameter of fourth ventricle

y=-0.84+0.23x gestational age

22

Table 2e Description of planes and measurements Mid-coronal view

Defined as the coronal plane passing through the foramina of Monro

Ventricular Index (VI)

The length of a horizontal line drawn between the midline and the outer border of the lateral ventricle, which is seen as an abrupt end of an echogenic line14

Thickness of the corpus callosum (CC)

The distance between the upper and lower borders of the CC seen as a linear hypoechogenic structure above the CSP22

Width and height of the cavum septum pellucidum (CSP)

Width: the maximum distance between the lateral borders of the CSP. Height: the distance between the upper and lower border of the CSP22

Diagonal width of the head of the caudate nucleus (CN)

Measured at right angles to a line that runs from its upper outer corner to the inner lower corner and the maximum distance between the ventricular wall and the lower margin of the caudate head that runs adjacent to the anterior limb of the internal capsule22

Width of basal ganglia The maximum distance between the most lateral border of the basal ganglia and the midline. The lateral border of the basal ganglia is distinct being more echogenic than the adjacent white matter22

Width of 3rd ventricle Measured from the inner edge to inner edge of the 3rd ventricle at its maximum width

Interhemispheric fissure (IHF)

The maximum width between the hemispheres, measured from the surface of opposite gyri

Extracerebral space (ECS) The perpendicular distance between the skull and the surface of the cerebral cortex near the midline2,3, 22

Parasagittal plane Thalamo-occipital distance (TOD)

The TOD was measured in a parasagittal plane that was angled laterally more posteriorly than anteriorly so that the entire length of the lateral ventricle was seen. It was measured from the most posterior part of the thalamus to the most posterior angle of the occipital ventricular horn13

Cortical depth of the cingulate sulcus

The cingulate cortex was visualised in a parasagittal plane just above the lateral ventricle and the cortical depth was measured from its echogenic upper margin to the lower margin defined as increased signal from the sub-cortical white matter 22

Posterior coronal plane Defined as a plane through the thalami and the maximum width of the cerebellum

Maximum transverse cerebellar diameter (TCD)

The maximum horizontal distance between the lateral borders of the cerebellar hemispheres22, 23

Mid-sagittal plane Length of CC The distance between the genu and splenium of the CC22 Cerebellar vermis height The distance between the upper and lower borders of the

vermis22 Height of cisterna magna The distance (as a vertical line) between the lower border

of the vermis and the skull just posterior to the foramen magnum22

Anterior-posterior diameter of pons

The distance between the anterior border of the pons at its maximal forward prominence and the posterior border at

23

the level of the upper border of the 4th ventricle22 Depth of the 4th ventricle The distance between the posterior border of the pons and

the anterior border of the vermis at the maximum depth of the 4th ventricle22

24

Table 3. Demographic data of the 106 infants including a comparison of birth weight

and head circumference centiles using British1990 and WHO 2006 reference data.

Mean birth weight centile was significantly lower than mean head circumference

centile (p<0.01).

Cohort demography (n=106) Median (range)

Gestational age (weeks) 39.28 (36.1-42.4)

Distribution of gestational age 77% between 38-41 weeks

Birth weight (grams) 3300 (1900-4400)

Mean birth weight centile (WHO 2006) 44.7 (0.1-99.0)

Mean birth weight centile (British 1990) 43.2

p value

0.18

Mean head circumference (cm)

males

females

35.0 (31.0-37.5)

35.21±1.17

34.6±1.27 (p=0.023)

Head circumference centile (WHO 2006) 66.85(0.1-99.5)

Head circumference centile (British 1990) 59.72

p value

0.90

Median age at scanning (days) 1 (0-4)

Table 4. Mean and standard deviation cUS measurements. +these data given median values and include as 0 the infants whose

interhemispheric fissure, and extracerebral space were too small to be measureable and whose cavum septum pellicidum was closed

Cerebral structure Imaging plane Parameter Measurement (mm, mean SD

p value right/left

10th / 90th percentile

Corpus callosum Mid-sagittal Length 45.953.24 41.63/ 49.955 Coronal Thickness 1.360.24 1.06/1.78 Cerebellar vermis Mid-sagittal Height 22.362.14 19.56/25.02 Cerebellar hemispheres Coronal Maximal

width 50.683.98 45.77/56.08

Fourth ventricle (n=84) Mid-sagittal Depth 3.510.93 2.25/4.80 Pons Mid-sagittal AP depth 16.861.75 14.35/18.86 Cisterna magna Mid-sagittal Height 3.271.1 1.75/4.89 Interhemispheric fissure (n=55)

Anterior coronal Width 1.68 (0.46-3.76) (median, range)

0.89/2.63

+Interhemispheric fissure Anterior coronal Width 0.88 (0.0-3.76) (median,range) 0/2.2 Extracerebral space (n=33) Anterior coronal Height 1.30 (0.36-2.2)(median, range) 0.75/ 2.04 +Extracerebral space Anterior coronal Height 0.0 (0.0-2.2) (median, range) 0/1.56 Cortical depth Parasagittal from

cingulate gyrus Depth 1.220.28 0.85/1.61

Right 11.241.71 0.20 9.09/13.50 Ventricular Index Coronal at foramen of Monro Left 11.421.79 9.35/13.60

Right 12.362.45 0.09 9.59/15.52 Thalamo-occipital distance Parasagittal Left 12.782.73 9.64/16.03

Third ventricle (n=29) Coronal Width 1.660.35mm 1.23/4.89 Cavum septum pellucidum (n = 66)

Maximal width on ant. coronal

Width 3.66 (0.61-8.37)(median, range) 2.07/6.77

+Cavum septum pellucidum Maximal width on ant. coronal

Width 3.05 (0.00-8.37) (median, range) 0.00/5

Cavum septum pellucidum Coronal Height 4.191.54 2.04/6.24 Basal ganglia width Coronal Right 17.801.87 0.63 15.35/20.09 Left 17.881.90 15.30/20.31 Caudate head width Mid-coronal Right 4.990.69 0.82 4.17/5.92 Left 5.010.90 3.89/6.34

26

Table 5e. Linear regression showing correlation between cUS measurements and

head circumference and birth weights

Head circumference Regression coefficient

P value 95% CI

Transverse cerebellar width (mm) 0.05 <0.01 0.02 to 0.08 Corpus callosum length (mm) 0.03 0.02 0.05 to 0.054 Pons AP width (mm) 0.02 <0.01 0.008 to 0.037 Birth weight Corpus callosum length (mm) 0.032 0.02 0.146 to 3.23 Cerebellar vermis height (mm) 1.50 <0.01 0.49 to 2.52 Mean thalamo-occipital distance (mm) 1.40 0.03 0.10 to 2.70

Table 6. Intercorrelation coefficients of 8 different structural measurements showing

substantial to almost perfect intra-and interobserver agreement.

Structure Interobserver ICC (95%CI)

Intraobserver ICC (95%CI)

Vermis height 0.88 (0.65 to 0.96) 0.86 (0.44 to 0.96) Cerebellar diameter 0.9 (0.71 to 0.97) 0.68 (0.26 to 0.92) Length of corpus callosum 0.86 (0.55 to 0.95), 0.93 (0.75 to 0.98) AP width pons 0.83 (0.48 to 0.94) 0.84 (0.45 to 0.95) Left ventricular index 0.76 (0.34 to 0.92) 0.66(-0.002 to 0.88) Right ventricular index 0.82 (0.3 to 0.94) 0.86(0.59 to 0.95) Left thalamo-occipital distance 0.92 (0.75 to 0.97) 0.90 (68 to 0.97) Right thalamo-occipital distance 0.88 (0.61 to 0.96) 0.83 (0.45 to 0.94)

27

Figure 1a

Figure 1b

28

Figure 2a

Figure 2b

29

Figure 2c