The calculation of intracranial volume using CT scans* David Gault, Francis Brunelle, Dominique Renier, and Daniel Marchac Centre for Craniofacial Anomalies, Neurosurgical Service, Hospital Necker-Enfants Malades, 149, rue de Stvres, F-75730 Paris Ctdex 15, France
Abstract. A m e t h o d o f calculating intracranial volume from horizontal c o m p u t e r i z e d t o m o g r a p h y scan slices is presented. The accuracy o f this technique was confirmed by applying it to 10 dry skulls a n d c o m p a r i n g the values obtained with the true intracranial volumes; as d e t e r m i n e d by filling the skulls with water. Values ranging between 98.14% and 102.6% o f the true values were obtained, the m e a n error being 1.13%. This technique is now being used to study intracranial v o l u m e changes in children with craniostenosis.
Key words: Intracranial v o l u m e - C A T scans - Craniostenosis.
Research on intracranial volume is o f interest to anthropologists, neurosurgeons and paediatricians amongst others. Measurements o f skull height, width, length or circumference have b e e n used to calculate intracranial volumes in adults a n d children, and a wide range o f values has been p u b l i s h e d [6, 7, 9 - 1 1 , 14]. The formulae e m p l o y e d have often been modified to improve their accuracy. However, such techniques do not allow for imp o r t a n t variations in contour a n d are unlikely to be reliable in the d e f o r m e d a n d asymmetric skulls o f children with craniostenosis. In particular, Buda et al. found that the n o r m a l relationship between occipitofrontal circumference and intracranial volume did not pertain to children with craniostenosis . We have found that measurements o f the surface area o f sequential horizontal computerized t o m 0 g r a p h y slices can be used to give an accurate estimate o f intracranial volume. Previous work analysing d a t a collected for 3 D CT scans with a software p r o g r a m to calculate intracranial
* This work was performed during a French exchange fellowship sponsored by the medical Research Council (UK) and the Centre National de la Recherche Scientilique (France)
Offprint requests to: D. Gault, 20 Christians Warehouses, Dockhead, 4 Shad Thames, Southwark, London SE1 2YT, UK
volume has been reported . Neither this p r o g r a m nor the quality scans required were available to us on a routine basis. A n alternative m e t h o d is described in this paper, which is both inexpensive a n d reliable.
Materials and methods The scan is obtained with the orbitomeatal line perpendicular to the scanning table. A CGR CE 10,000 scanner was used with scan time of 2.1 s. Settings of 130 KV and 80 mA were used. The first cut includes the foramen magnum and the skull is then scanned sequentially at 5 mm intervals up to and including the vertex. Scans taken at 10 mm intervals give a volume measurement that is less accurate, the average error being 6%, and scans taken at more frequent intervals would be too time-consuming for routine use. A "bone" window is used centred on 300 Hounsfield units with a range of 1,200 Hounsfield units . Bone is represented as an abnormally thick layer on normal parenchymal scans which, if used, would give inaccurate results . On one of the prints a square or rectangle of known surface area is incorporated to permit calibration of the planimeter. Prints of 9.5 by 9.5 cm were used in this work, the reduction scale being 18 : 50. The surface area of all intracranial sectors appearing on each scan is measured with a planimeter, by manipulating a dot in the centre of the magnifying eyepiece around the inner border of the skull (Fig. 1). Particular attention is required to include the lowermost cut of the middle cranial fossa and to exclude the orbital contents. The planimeter gives a digital readout of surface area, and the total figure obtained for all the cuts is easily converted to square millimeters using the calibration square. This total surface area multiplied by the thickness of the cuts (5 mm) gives the intracranial volume (Fig. 2). To explore the accuracy of this technique 10 dry skulls were scanned, their volumes calculated according to the above protocol and these were compared to the real volumes. To present the technique with a realistic challenge, skulls with scaphocephaly, trigonocephaly, plagiocephaly and oxycephaly were included in this experiment. The determination of the real volume of a dry skull is, however, not an easy task. Traditionally, the foramina are closed with cotton wool and the skull is filled with mustard seeds. Using this technique in a series of 50 skulls we found a mean variation in sequential readings of 26 cm ~ (range 0-65 cm ~) due to variations in packing in both the skull and the measuring cylinder. Davis and Wright  have described a technique wherein a lubricated balloon within a skull is filled with water trader pressure. We were concerned that the balloon would not conform to all the available intracranial crevices. We have, however, discovered that it is possible to occlude all the foramina and
Fig. 1. Use of the planimeter to calculate intracranial surface area
Table 1. Calculated and real intracranial volumes
Fig. 2. Intracranial volume represented as a series of horizontal slices 5 mm thick
potential leakage points by the careful application of Horsley's Bone Wax. It is then possible to measure the real skull volume with water. The wax can be removed subsequently with hot water. This technique was used to establish the real intracranial volume in the ten skulls.
The calculated and real intracranial volumes are presented in Table 1. The mean percentage error was 1.13%, with a range of 0.09-2.6%. The average time taken for the scans was 13.5 rain and the average time taken to calculate the volume from the scans 18 min.
The technique described is sufficiently accurate to rec o m m e n d its use to analyse the results o f craniofacial operations specifically designed to augment intracranial volume, e.g. frontal advancement for brachycephaly. Manipulation of the flexible arm o f the planimeter around the inner table of the cranium is most easily done seated at a
lighted table and, despite the reliance on a h u m a n skill, use of "the bone window" enables accurate recognition of the object boundary. The obliquity o f the bone in the final cephalad sections often produces some blurring o f the inner limits o f the skull and here the centre of the area o f attenuation was taken as the true boundary . Only a small part o f the total volume is addressed in these final scans and this m a y explain how the accuracy o f the results were conserved. Although the scans in this series were all obtained in dry skulls with bone air contrast, when the same technique is applied to children before and after surgery the inner table of the skull is still sufficiently demarcated to enable volume calculation. Previous attempts at volume determination o f visceral organs from CT scans showed mean percentage errors between 3.59 and 4.95%, using 10 m m scan spacing . We have shown that the 5-mm scan spacing in this study is an important factor in obtaining improved results. One important application of this technique will be to study how volume abnormalities relate to intracranial pressure in children with craniostenosis. T h e age o f the child and the type o f craniostenosis are likely to be important factors influencing this relationship. Raised
273 intracranial pressure in some children with craniostenosis has already been noted [12, 13]. Should there be a relationship between this and reduced intracranial volume, it may be possible to use intracranial volume measurement as a diagnostic tool to indicate whether early surgery is required for functional reasons. At present, the most reliable way to determine if urgent release of prematurely fused skull sutures is required is to record intracranial pressure though a burr hole . In some cases volume measurements alone m a y suggest whether urgent surgery is required, thus avoiding an additional, albeit minor, operation in the pre-operative evaluation of craniostenosis. Initial studies have shown that when simple bone flaps are fashioned in young children without any rigid cranial remodelling, there is considerable augmentation in intracranial volume in the first week following surgery. This suggests that a significant "brain push" exists in infants and tends to support the assumption that reduced volume and raised intracranial pressure are related.
Acknowledgements. This work was performed during a French exchange fellowship sponsored by the Medical Research Council (UK) and the Centre National de la Recherche Scientifiqne (France). We gratefully acknowledge the assistance of the Mus6e de l'Homme of Paris in preparing this report.
References 1. Breiman RS, Beck JW, Korobkin M, Glenny R, Akwari OE, Heaston DK, Moore AV, Ram PC (1982) Volume determinations using computed tomography. Am J Roentgenol 138: 329-333 2. Buda FB, Reed JC, Rabe EF (1975) Skull volume in infants. Methodology, normal values and application. Am J Dis Child 129:1171-1174
3. Checkley DR, Zhu XP, Antoun N, Chen SZ, Isherwood I (1984) An investigation into the problems of attenuation and area measurements made from CT images of pulmonary nodules. J Comput Assist Tomogr 8:237-243 4. Davis PJM, Wright EA (1977) A new method for measuring cranial cavity volume and its application to the assessment of cerebral atrophy at autopsy. Neuropathol Appl Neurobiol 3: 341-358 5. Dufresne CR, McCarthy JG, Cutting CB, Epstein F J, Hoffman WY (1987) Volumetric quantification of intracranial and ventricular volume following cranial vault remodeling: a preliminary report. Plast Reconstr Surg 79:24-32 6. Fernandez F, Moule N, Singhi P, Singhi S (1984) Roentgenographic skull volume in Jamaican children between the ages of one month and five years. West Indian Med J 33:227-230 7. Gordon IRS (1966) Measurement of cranial capacity in children. Br J Radio139:377-381 8. Koehler PR, Anderson RE, Baxter B (1979) The effect of computed tomography viewer controls on anatomical measurements. Radiology 130:189-194 9. Lichtenberg R (1960) Radiographic du crfme de 226 enfants normaux de la naissance ~t 8 ans. Impressions digitiformes, capacit6, angles et indices. Thesis, University of Paris 10. McKinnon IL, Kennedy JA, Davies TV (1956) The estimation of skull capacity from roentgenologic measurements. Am J Roentgeno176 : 303-310 11. Olivier G, Aaron C (1975) D6termination de la capacit6 cr~nienne par les dimensions endocrhniennes. Bull Assoc Anat (Nancy) 59: 707-717 12. Renier D, Sainte-Rose C, Marchac D, Hirsch JF (1982) Intracranial pressure in craniostenosis. J Neurosnrg 57:370377 13. Renier D, Sainte-Rose C, Marchac D (1987) Intracranial pressure in craniostenosis: 302 recordings. In: Marchac D (ed) Craniofacial surgery. Springer, Berlin Heidelberg New York, pp 110-113 14. Singhi S, Wafia BNS, Singhi P, Walia HK (1985) A simple non-roentgenographic alternative for skull volume measurement in children. Indian J Med Res 82:150-156 Received December 23, 1987