Francis J. Hahn1 Wei-Kom Chu 1

John Y. Cheung2

This article appears in the March/April 1984 issue of AJNR and the June 1984 issue of AJR.

Received August 13, 1982; accepted after revi­sion August 22, 1983.

, Department of Radiology, University of Ne­braska Medical Center, 42d and Dewey Ave., Omaha, NE 68105. Address reprint requests to F. J. Hahn.

2 School of Electrical Engineering and Computer Sciences, University of Oklahoma, Oklahoma City, OK 73104.

AJNR 5:155-157, March/April 1984 0195-6108/84:0502-0155 $00.00 © American Roentgen Ray Society

155

CT Measurements of Cranial Growth: Normal Subjects

Growth patterns of the cranium measured directly as head circumference have been well documented. With the availability of computed tomography (CT), cranial dimensions can be obtained easily. The objective of this project was to establish the mean values and their normal variance of CT cranial area of subjects at different ages. Cranial areas and its long and short axes were measured on CT scans for 215 neurologic patients of a wide age range who presented no evidence of abnormal growth of head size. Growth patterns of the cranial area as well as the numeric product of its linear dimensions were determined via a curve fitting process. The patterns resemble that of the head circum­ference growth chart, with the most rapid growth observed in the first 12 months of age and reaching full size during adolescence. It is believed that the availability of such reference, in addition to the head circumference measurement, will be valuable to the CT reviewer in determining the growth status of head size.

The growth pattern of the cranium of children, as measured by their head circumferences, has been extensively studied and well documented [1-5] . Before the availability of computed tomographic (CT) scanners, the measurement of head circumference was the only feasible means of estimating cranial size. With CT scanners, area or linear dimensions of the cranium can be easily obtained. However, mean values for the cranial area as determined on CT scans are not widely available. The primary objective of our study was to establish a set of mean values of cranial areas and their normal variances for subjects at different ages. By comparing individual patient measurements with these mean values, abnormal growth of the cranium may be detected by CT.

Subjects and Methods

The subjects consisted of 215 pediatric neurologic patients who demonstrated no evidence of any disease that might affect their cranial size. Patients with indications as "rule out hydrocephalus" or "rule out microcephalus" were automatically eliminated. There were 125 boys and 90 girls with ages ranging from birth to 18 years of age. The CT scanner used was aGE 8800 with JC software package. The scan time was 9.6 sec, and the images were displayed on a 320 x 320 matrix. All the patients were scanned in the standard supine position with an approximately 5 0 _100 tilt from the canthomeatal line. The midventricular slice of the CT scan, which shows the most prominent frontal horns of the lateral ventricle , was selected for the estimation of the cranial area since it reflects the maximum size of the cranium. The same window setting (center at 35 with width 100) was used for screen viewing and hard-copy filming. Changing the window level and width to obtain an optimal display did not influence the measurement of the cranial area or its linear dimensions. A representative slice of the CT head scan is shown in figure 1 A. Using a built-in cursor, the outer edge of the cranial vault was traced and the enclosed cranial area was automatically calculated by the computer. In addition , the maximum anteroposterior (AP) dimension as well as the maximum width of the cranium were also estimated (fig . 1 B). Outer diameters of the skulls were assessed as AP and lateral dimensions. Both the head areas and cranial linear dimensions were recorded for each subject. More than 90% of the measurements were performed by

156 HAHN ET AL. AJNR:5, Mar/Apr 1984

Fig. 1.-A, Midventricular slice of CT head scan showing prominent frontal horns of lateral ventricles selected for measurement of cranial area and linear dimensions. B, Illustration of measurements of maximum AP and lateral (LAT) dimensions of outer margin of cranium.

220

2 10

200

190

180

N 170 E 160 3 /'

<1: 150 /,'

W 140

/ a: :/. : <1: 130 /"

-' . / , ~ 120 I Z <1: 110 , , ./ a: 1' / u 100

90 I I x=MUl TlPLE DATA SAMPLES

80 / 70

.J.

60

2 4 6 8 10 12 14 16 18 202224 4 6 8 10 12 14 16 18 _---- MONTH .. .--YEAA ----+-

AGE

Fig. 2.-Growth pattern of head sizes as measured by CT cranial area from birth to 18 years of age. One-month-old group includes newborn to 45 days: 24-month-old group includes 24-30 months. Each dot represents one subject.

the primary author, with less than 10% being measured by CT technicians. The interobserver error was determined by requesting different observers (two radiologists and two CT technicians) to measure the same CT head scan. The error was less than 3% in both the area and dimensional measurements. The accuracy of the area and distance measurements were better than 1 % when meas­urements were performed on phantoms with comparable dimensions.

Results

The data recorded for each subject included age, cranial area, and AP and lateral cranial dimensions. Under 2 years of age, the ages were recorded in monthly units , and over 2 years in yearly units. The 2 year dividing point was arbitrarily chosen because rapid head growth was observed during the first 12 months, gradually slowing in the second 12 months; and the patient ages recorded on the requisition forms were monthly below 2 years and yearly beyond that age. A scatter plot showing all the cranial areas that constitute the data

TABLE 1: Mean Cranial Areas and Their Normal Ranges

Age NO. of Subjects

Subjects s 2 years old (months): 1 2 . 3 . 4 . 5 . 6 . 7 . 8 .. 9 .

10 . 11 12 . 13 .... . . . . . 14 . . 15 . 16 . 17 . 18 . 19 20 21 . 22 ........ . 23 . 24 .

17 6 4 6 8 6 3 7 7 1 3 9 1 1 1 3 3 3 1 1 2 1 2

14 Subjects ~ 3 years old (years):

3 . 13 4 . . 14 5 5 6 ....... . . 8 7 . . . . . . . . . 7 8 . . . . . . . . . 8 9 . 6

10 . 8 11. . . . . . . . . 7 12 . . . . . . . . . 1 13 . . . . . . . . . 11 14 . 5 15 . 6 16 .... . .. .. 6 17 . 7 18 . 3

Mean Area' in em' (Ranget)

75.5 (63.0-88.0) 93.9 (81.4-106.4)

104.7 (92.2-117.2) 112.3 (99.8-124.8) 118.2 (105.7- 130.7) 123.0 (110.5-135.5) 127.1 (114.6- 139.6) 130.7 (118.2-143.2) 133.8 (121.3-146.3) 136.6 (124.1- 149.1) 139.1 (126.6- 151.6) 141.4 (128.9-153.9) 143.5 (131 .0-156.0) 145.5 (133.0-158.0) 147.3 (134.8- 159.8) 149.0 (136.5-161.5) 150.6 (138.1-163.1) 152.1 (139.6-164.6) 153.6 (141.1-166.1) 154.9 (142.4-167.4) 156.2 (143.7-168.7) 157.5 (145.0-170.0) 158.7 (146.2-171.2) 159.8 (147.3-172.3)

163.0 (146.4- 179.4) 169.4 (1 52.9-185.9) 174.4 (157.9-190.9) 178.6 (162.1-195.1) 182.0 (165.5-1 98.5) 185.0 (168.5-201.6) 187.7 (171 .2-204.2) 190.1 (173.6- 206.6) 192.2 (195.7-208.7) 194.2 (177.7-210.7) 196.0 (179.6-212.6) 197.7 (180.7-213.7) 199.2 (1 82.7- 215.7) 200.7 (184.2- 217.2) 202.1 (185.6- 218.6) 204.4 (187.9- 220.9)

. Mean area is derived from the regression formula provided in the text. t Range reflects the 90% of the population closest to the mean.

base together with the mean and variance values is shown in fig . 2. Analyses were performed separately for data below 2 years old and above 2 years old . The mean values as well as their normal limits are also shown in table 1. The regression formula for cranial area as a function of age for ages less than or equal to 24 months is: area (cm2) = 75.55 + 26.5 x logn (age in months); for this equation, r = 0.96. The regres­sion formula for cranial area as a function of age for ages more than 2 years is: area (cm2) = 140.11 + 21.44 x logn (age in years); for this equation, r = 0.86.

Although the head area can be measured easily when scans are displayed on a CT display monitor, their values cannot be obtained easily when the scans are presented in hard-copy format. To overcome this difficulty, linear dimensions of the cranium, which can be easily measured on films, are thought to be useful in determining the growth status of the cranium . Since the cranium is often elliptical , both its long (AP) and

AJNR '~ , Mar/Apr 1984 CT MEASUREMENTS IN NORMAL CRANIAL GROWTH 157

300

280

260

240 N 220 E S 200 I- 180 U => 0 160 / a / 0: 140 /

I (l. 120 I /

/

100 /

80 I

80

.,-.--/'

.--"

95%_

95'" ~ / ....... 5% .--"

i I I I I , I I , I i I I Iii 10 12 14 16 18 20 22 24 4 6 B 10 12 14 16 18

----- MONTH ----... ....--- YEAR ------+-

AGE

Fig. 3.-Growth pattern of head sizes as product of CT cranial linear dimensions from birth to 18 years of age.

short lateral axes have to be measured. To better estimate its size from the measured linear dimensions, numeric prod­ucts of these two dimensions were obtained and used as the cranial size estimator. Figure 3 shows the predicted mean values and their normal ranges as estimated by regression analyses of all data collected . The regression formula for cranial product as a function of age when age is less than or equal to 24 months is: product (cm2) = 96.18 + 34.51 x logn (age in months); for this equation, r = 0.96. The regression formula for cranial product as a function of age for ages greater than 2 years is: product (cm2) = 181.64 + 28.06 x logn (age in years); for this equation, r = 0.83 .

Discussion

Quantitative analyses of ventricular sizes on CT scans have been reported using the cerebroventricular ratios [6-8] . The cranial size as determined by the number of pixels has also been investigated [9] . Sabattin [10] has presented a critical review of various methods used to measure the size of the intracranial cerebrospinal fluid spaces.

The most rapid head growth is noted during the first 12 months of age with the cranial area increasing from about 75 cm2 to about 140 cm2, a roughly twofold increase. The growth rate slows down after 12 months and is 160 cm2 at 24 months. The cranium reaches full size at adolescence with a value of

200 cm2. The same growth pattern is seen on the standard Nellhaus head circumference charts used by the bedside physician [1]. An identical growth pattern is observed for the product of the linear dimensions measured from CT scans, which show the mean values to increase from 95 cm2 to 180 cm2 during the first 12 months of age, and then increase to 200 cm2 at 2 years. When fully grown the mean value of the product is about 260 cm2. Although the profiles of both area and product curves are similar, the area curves should be used to obtain the head size. The product should only be used when area values are not available. We also believe that the normal ranges are of more clinical value than the mean values to determined abnormal growth.

ACKNOWLEDGMENTS

We thank W. Boyd, M. Ferguson , and D. Plettner for measure­ments; R. Torkelson and A. Fisher for clinical information; F. Messner, Sacred Heart Hospital, Yankton, SO, for sending CT measurements of his patients; and C. A. Dobry, Department of Radiology, for interest and encouragement in this project.

REFERENCES

1. Nellhaus G. Head ci rcumference from birth to eighteen years. Pediatrics 1968;41 : 1 06-114

2. Pryor HB. Charts of normal body measurements and revised width-weight tables in graphic form. J Pediatr 1966;68: 615-631

3. McCammon RW. Human growth and development. Springfield, IL: Thomas, 1970

4. Marks HG, Borns P, Steg NL, Stine SB, Stroud HH, Zates TS. Catch-up brain growth-demonstrations by CAT scans. J Pediatr 1978;93: 254-256

5. Sher PK, Brown SB. A longitudinal study of head growth in preterm infants. II : Differentiation infantile hydrocephalus. Dev Med Child Neurol 1975;17:711-718

6. Hahn FJY, Rim K. Frontal ventricular dimensions on normal computed tomography . AJR 1976;126:593-596

7. Hahn FJY , Rim L. A quantitative analysis of ventricular size on computed tomographic scans. CT 1977;1 :121-1 25

8. Hahn FJY, Schapiro RL. The excessively small ventricle on computed axial tomography of the brain. Neuroradiology 1976;12 :137-139

9. Zatz LM , Jernigan TL, Ahumada AJ Jr. Changes on computed cranial tomography with aging: intracranial fluid volume. AJNR 1982;3: 1-11

10. Sabattin L. Evaluation and measurement of the normal ventricular and subarachnoid spaces by CT. Neuroradiology 1982;23: 1-5