Khalid Shehzad ( Department of Anatomy, Ziauddin University )
Asifa Rafique ( Senior Sonologist )
Syed Ejaz Alam ( The Ultrasound Clinic, PECHS, PMRC Research Centre, Jinnah Postgraduate Medical Centre )
Shahida Zaidi ( Department of Obstetrics and Gynecology4, Ziauddin University, Clifton, Karachi. )
Objective:To assess the correlation between ultrasonic manual and automatic ellipsoid measurements of foetalhead and abdominal circumferences.Methods:This comparative cross-sectional study was conducted at the Ultrasound Clinic, Karachi, from January to July 2004. Seventy two normal pregnant women with gestational ages between 14 to 38 weeks were included.Foetal head and abdominal circumferences were measured by manual method using traditional formulae and by automatic ellipsoid mode incorporated into the real-time ultrasound machines (EcoCee and Power Vision,Toshiba, Japan) using convex probes of 3 and 4.2 MHz frequency. Three readings of each parameter i.e. headcircumference (HC) and abdominal circumference (AC) were taken by both methods. Mean values were calculated and compared using paired sample't' test to assess the correlation between ultrasonic manual andautomatic ellipsoid measurements.Results:The mean HC measured through manual method was found to be 20.11 ± 7.04 cm and was significantlydifferent from that of 19.46 ± 6.82 cm measured through ellipsoid automatic technique. There was, however, aperfect positive correlation (r = 0.999) between these measurements. In case of AC, there was no statistical difference between the mean values obtained by the manual technique and those measured automatically (16.79± 6.24 vs. 16.74 ± 6.11; p < 0.44), correlation between these measurements (r = 0.997) was also significantly positive.Conclusion:Correlation was found between the mean measurements of the foetal HC and AC by the manualand the automatic method of calculation; and even stronger for the latter. The automatic mode of measurementis a more rapid method than the manual calculation, and its use may be encouraged (JPMA57:352:2007).
Foetal biometry has been in use for the assessment of gestational age and for monitoring foetal growth since the late 1960s when Campbell's first publication on the subject appeared.1 The parameters in common use, are the biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC) and femur length (FL).2-5Accurate measurements of the head and abdominal circumferences are essential if foetal growth is to be accurately monitored. These can be performed manually by taking two measurements at right angles to each other at anappropriate plane, and the result calculated from the formula for an ellipsoid; it can also be performed on allcommercially available ultrasound machines. Watson and co-workers found a statistically significant difference between the directly measured AC and that of calculated values. The statistical difference wash owever, not found clinically significant.6 Shield JR and co-workers7, while working ellipse calculations versus planimetry, expressed that all equations were found to be equally accurate in calculating HC andAC. Compared to planimetry, the ellipse method is more adequate in calculating circumferences, however, this was not true when calculating their ratios.7Similar author, in another study, working on 122 foetal heads and abdomens,reported that the mean of the cephalic index was 79.1% ±5.21% and that of the abdominal index was 89.1% (± 7.0%),therefore, the simplest equation, (D1 + D2) X 1.57, was recommended for calculating AC and HC except insituations of extreme dolicocephaly.8 The objective of this study was to assess the correlation between ultrasonic manual and automatic ellipsoid measurement of foetal head circumference andabdominal circumference.
Subjects and Methods
This comparative cross sectional study was conducted at the Ultrasound Clinic PECHS, Karachi. The study consisted of a convenient sample of 72 uncomplicated singleton pregnant women scanned at the Ultrasound Clinic, Karachi, between 14th to 38th week of gestation.They were part of a research project titled "Sonographicfoetal biometry in a cohort of Pakistani population." The ultrasound machines used were EcoCee and Power Vision,Toshiba, Japan, with convex probes of 3.0 and 4.2 MHz frequencies. Three readings of each parameter, head circumference (HC) and abdominal circumference (AC)were taken using the manual and the automatic modes. For HC, the mathematical formula used was that foran ellipsoid, viz. Circumference = 0.5 π(D1+ D2) where π= 3.142and D1, D2are the two diameters at right angle to each other,in this case representing the biparietal diameter (BPD) andthe occipito-frontal diameter (OFD) respectively. Thus simplified to (BPD + OFD) X 1.62For the automatic or direct measurements of HC,two calipers were placed on BPD and using trackball to reach the outer margins of skull table from sinciput to theocciput.In case of abdominal circumference, the same formula was modified to cater for the more circular abdomen, viz.Circumference = D1+ D2X 1.57, where 1.57 is the correction factor for a circle.The automatic or direct AC measurements was taken by placing two calipers on the periphery of the abdominal wall using the trackball to adjust calipers to theouter margins. The data was entered and analyzed on SPSS version10. Continuous variables including HC and AC were represented by mean and standard deviation. Paired't' testwas applied to compare the difference in measurements ofthese variables by manual and automatic methods. The correlation between manual and automatic measurements was determined through the correlation coefficient (r) andcoefficient of determinationr2.
Atotal of 72 subjects were enrolled in the study.Three readings of each parameter, head circumference (HC)and abdominal circumference (AC) were taken using manual and automatic methods.Mean HC calculated by the manual method for the foetuses of different gestational ages (14-38 weeks) was found to be 20.11 (± 7.04) cm, where as that measured by the automatic method was 19.46 (± 6.82) cm; this difference was statistically significant (p < 0.001, 95% C.I of difference = 0.559, 0.731) as shown in table 1. There was apositive correlation (r = 0.999) between these measurements(Figure-1) showing that the differences were constant. Onlytwo observations of manual HC measurements were found higher than the automatic measurements with a difference
of 0.7 and 0.8 cm respectively. Sixteen observations were found equivocal with a difference of 0 to <0.5 cm. While inall remaining 54 observations, automatic HC measurements were found higher than manual HC measurements with a difference of 0.5 to <1.5 cm.The mean AC calculated by manual technique was16.79 (± 6.24) cm, and that measured by automatic technique was 16.74 (± 6.11) cm. This was not statistically significant (p= 0.44, 95% C.I = 0.07, 0.161) as shown in table 1; correlation between these measurements (r = 0.997) was positive (Figure-2). In case of AC, 31 manual observations were found lessthan automatic measurements with a difference ranging from0.1 to 0.6 cm. Thirty three observations were equivocal where the difference was 0 to < 0.5 cm and in the remaining 8 observations, the automatic AC was found higher than themanual AC measurements with a difference of 0.5 to 3.0 cm.
Sonographic measurement of foetal ultrasound parameters form the basis of accurate determination of gestational age, monitoring of foetal growth, and detecting growth abnormalities.9Some selected parameters are usedto estimate foetal weight.10-12 The shape of the foetal head is ovoid, whereas that ofthe upper abdomen (at the plane where AC is measured) ismore circular. Various mathematical formulas are used to measure the circumferences, the one in common use forhead circumference being the one for an ellipsoid, is. Circumference = 0.5 π(D1+ D2)Whereπ= 0 3.142, and D1and D2are the two diameters at right angle to each other and 1.57 is the correction factor for a circle. This equation is commonly used for the calculation of AC. However, the same equationhas been claimed to be accurate for the HC, except insituations of extreme dolicocephaly.11 The formula for ovals or ellipses like the head is HC = BPD + OFD x 1.62 Where, 1.62 is the correction factor for ellipses. Amore rigorous formula13 for ellipses is Circumference = 0.325 D12+ D22 Systematic errors are common with first two groups;the error further increases near term, as the foetal head becomes more ellipsoid (with a difference of 0.8 weeks incalculation of gestational age). The last group is related to mathematical random error because of its complex natureand applicability.14In addition, minor inter-observer biaseshave been reported while calculating circumferences.15The modern ultrasound machines are equipped withthe facility to estimate the circumferences technically using calipers and trackball for both circles and ellipses. In thisstudy an attempt was made to find the correlation between manually measured head and abdominal circumferencesmaking use of traditional formulae compared with the automatic mode of ultrasound machine. There was a statistical difference between the mean values of HC obtained by the two techniques; the differencebeing of the order of 0.6 cm (20.11±7.04 cm Vs 19.46±6.82cm) which forms 3% of the value of 20.1 cm. This may notbe clinically significant; i.e. the value of HC at 22 weeks of gestation ranges from 17.4 cm to 22.3 cm (5th and 95thcentiles).16,17 Therefore, it can be suggested that thedifference obtained by two methods, although statistically significant (p = 0.001), may not be clinically significant. This is supported by studies by Zador and WLu. Zadorin 199118, who described the highly significant correlation between direct measurements of BPD (r= 0.986), OFD(r=0.958) and HC (r= 0.972) and those obtained by theoperators by planimetry. The mean difference (direct minusplanimetry) was 1.87± 1.94 mm for BPD, 2.82± 4.13 for OFDand -0.36± -9.87mm for AC. These differences were independent of operator's identity, instrument used andgestational age. He recommended the use of direct measurement system that can give the foetal headmeasurements which correlate highly with manual determination by a skilled operator and which takes a fractionof the time.18According to Lu Wet al19, the difference between automatic and sonographer's manual measurements were0.12% for BPD and - 0.52% for HC. The 95% CI of the agreements were -3.34%, 3.58% for BPD and -5.50%,4.45% for HC. The results demonstrated that the two measurements were consistent and accurate.19In our study, regarding measurement of AC, there wasno statistical difference between the values obtained by thetwo methods (p=0.44). Similar results were obtained byWatson et al15 who, while working on AC in 235 cases demonstrated that the directly measured AC was found to begreater than the calculated value (p= 0.00014). The magnitude of the difference however, was only 1.3 ±2.2 %,which is smaller than the average inter-observer measurement error. The study further stated that although statistically significant, the difference between these twomeasurement methods is not clinically significant, suggesting that either method is acceptable to determine foetal AC. This discussion suggests that the automatic measurement is a morerapid method than the manual, and is recommended for the measurement of the head and abdominal circumferences.
The measurements of the head circumference by the manual and the automatic methods were statistically significant; however, the difference may not be clinically significant, as it would fall within the standard deviation of approximately 4 days allowed for the HC. There was strong positive correlation (r = 0.999).As regards the measurement of AC by the two methods, there was no statistical difference between the two, however, there was strong positive correlation betweenthe two methods (r = 0.997)Automatic measurements are performed morerapidly than the manual ones, and may be routinely used for the measurement of head and abdominal circumferences.
The authors thank Dr. Amir Omair, Department of Community Health Sciences, Ziauddin University for hisstatistical assistance.
1.Campbell S. An improved method of fetal cephalometery by ultrasound. J.Obstet Gynaecol Brit Cwelth 1968; 75: 568-76.
2.Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal biparietal diameter: acritical re-evaluation of the relation to menstrual age by means of real-timeultrasound. J ultrasound Med 1982; 1: 97-104.
3.Chitty LS, Altman DG, Henderson A, Campbell S. Charts of fetal size: 2. Headmeasurements. Br J Obstet Gynaecol 1994; 101: 35-43.
4.Warda AH, Deter RL, Rossavik IK, Carpenter RJ, Hadlock FP. Fetal femurlength: a critical reevaluation of the relationship to menstrual age. ObstetGynaecol 1985; 66: 69-75.
5.Exacoustos C, Rosati P, Rizzo G, Arduini D. Ultrasound measurements of fetallimb bones. Ultrasound Obstet Gynecol 1991; 1: 325-30.
6.Watson WJ, Chescheir NC, Seeds JW. Ultrasound determination of fetalabdominal circumference: a comparison of measurement methods. Am JPerinatol 1991; 7: 182-3.7.Shields JR, Mediaris AL, Bear MB. Fetal head and abdominalcircumferences: ellipse calculations versus planimetry. J Clin Ultrasound1987; 15 237-9.
8.Shields JR, Medeanis AL, Bear MB. Fetal head and abdominalcircumferences: effect of profile shape on the accuracy of ellipse equations. JClin Ultrasound 1987; 15 241-4.
9.Degani S. Fetal biometry: clinical, pathological, and technical considerations.Obstet Gynecol Surv 2001; 56:159-67.
10.Hadlock FP, Deter RL, Carpenter RJ, Park SK. Estimating fetal age: effect ofhead shape on biparietal diameter. Am J Roentgenology 1981; 137: 83-5.
11.Campbell S, Wilkin D. Ultrasonic measurement of fetal abdomen circumferencein the estimation of fetal weight. Br J Obstet Gynaecol 1975; 82: 689-97.
12.Hadlock FP, Harist RB, Carpenter RJ, Deter RL, Park SK. Sonographicestimation of fetal weight. The value of femur length in addition to head andabdomen measurements. Radiology 1984; 150: 535-40.
13.Durr-e-sabih. Measurements in ultrasound. In: Aprimer of practicalultrasound.1st ed. Islamabad: PINSTECH, 1995: pp 263-4.
14.Kurtz AB, Kurtz RJ. The ideal fetal head circumference calculation.Ultrasound Med 1989; 8: 25-9.
15.Hadlock FP, Kent WR, Loyd JL, Harrist RB, Deter RL, Park SK et al. Anevaluation of two methods for measuring fetal head and body circumferences.J Ultrasound Med 1982; 1: 359-60.
16.Jeanty P, Romero R. Obtaining the head perimeter. In: Jeanty P, Romero R,eds. Obstetrical ultrasound. 1st ed. New York: Mc Graw- Hill Book Company,1984, pp 87-90.
17.Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal head circumference:Relation to menstrual age. Am J Roentgnology 1982; 138: 649-53.
18.Zador IE, Salari V, Chik L, Sokol RJ. Ultrasound measurement of the fetalhead: computer versus operator. Ultrasound Obstet Gynecol 1991; 1: 208-11.
19.WLu, Tan J, Floyd R. Automated fetal head detection and measurement inultrasound images by iterative randomized hough transform. Ultrasound MedBiol 2005; 31: 929-36.