- Email: [email protected]
Fetal and Fetal Brain Volume Estimation in the Third Trimester of Human Pregnancy Using Gradient Echo MR Imaging
Magnetic Resonance Imaging, Vol. 16, No. 3, pp. 235–240, 1998 © 1998 Elsevier Science Inc. All rights reserved. Printed in the USA. 0730-725X/98 $19.00 1 .00
PII S0730-725X(97)00281-6
● Original Contribution
FETAL AND FETAL BRAIN VOLUME ESTIMATION IN THE THIRD TRIMESTER OF HUMAN PREGNANCY USING GRADIENT ECHO MR IMAGING Q.Y. GONG,*† N. ROBERTS,† A.S. GARDEN,‡
AND
G.H. WHITEHOUSE*†
*Department of Medical Imaging, †Magnetic Resonance and Image Analysis Research Centre, and the ‡Department of Obstetrics and Gynaecology, University of Liverpool, Liverpool, UK The Cavalieri method has been applied in combination with gradient echo magnetic resonance imaging (MRI) to investigate the increase in the volume of the fetus and fetal brain in the third trimester of pregnancy. Eighteen women with singleton pregnancies were recruited. Birthweights for the fetuses all lay within the 10 –90th centile based on Liverpool data. A regression analysis, weighted using values derived from the coefficient of error predicted for each volume estimate, revealed a linear relationship between total fetal volume and gestational age (R2 5 0.88) and between fetal brain volume and gestational age (R2 5 0.71) during the third trimester. Fetal volume increased by an average of 25.2 ml per day and fetal brain volume increased by an average of 2.3 mL per day. Fetal brain volume is on average a constant proportion (10%, SD 5 2%) of total fetal volume throughout the third trimester. Volume data were also obtained for eight fetuses diagnosed as abnormal. The volume of seven of the eight abnormal fetuses fell outside the 95% confidence interval established from the data obtained for the normal fetuses. However, for only three of the eight abnormal fetuses did brain volume fall outside the 95% confidence interval established for normals, possibly due to brain sparing occurring in asymmetrical growth retardation. The volume of the fetus and fetal brain may be readily estimated directly using the Cavalieri method and magnetic resonance imaging. These parameters represent potentially useful information for assessing fetal growth. © 1998 Elsevier Science Inc. Keywords: Brain; Cavalieri method; Fetus; MRI; Point counting; Stereology; Volume
INTRODUCTION
sional parameters such as biparietal diameter, abdominal circumference, and femor length are obtained.4 Recently, attempts have been made to measure the volume of fetal organs with three-dimensional ultrasonography, but problems are encountered in patients with pronounced oligohydramnios.5 Magnetic resonance imaging (MRI) is not generally limited by large maternal body habitus or lack of amniotic fluid. MRI has been used to investigate fetal anatomy in utero6 – 8 and, with the development of fast MR imaging sequences, it has been possible to obtain measurements of fetal volume and fetal organ volume in vivo.9 –11 However, the echo planar imaging (EPI) sequence used is not yet widely available and generally only suspected abnormal pregnancies have been studied. Many studies have been carried out to assess the
The ability to evaluate fetal growth in utero is important in perinatal management. The identification of growth abnormalities can assist in decisions regarding fetal and maternal management and clinical intervention. Manera1 and Dobson et al.2 report that perinatal mortality is four to ten times greater in pregnancies complicated by intrauterine growth retardation (IUGR) compared with pregnancies in which intrauterine growth is normal. Identification of IUGR after 37 weeks gestation is an indication for delivery to reduce the chance of fetal mortality.3 Imaging plays an important role in assessing and monitoring fetal growth in utero. The most widely used imaging modality is ultrasonography. However, this method is not, in general, used to obtain measures of fetal volume. Instead, arbitrary one- and two-dimenRECEIVED 8/10/97; ACCEPTED 10/19/97. Address correspondence to Dr. Neil Roberts, Magnetic Resonance and Image Analysis Research Centre, University of
Liverpool, P.O. Box 147, Liverpool L69 3BX, UK. E-mail: [email protected] 235
Magnetic Resonance Imaging ● Volume 16, Number 3, 1998
236
Table 1. Characteristics of the pregnancies for the normal (18 cases) and abnormal (8 cases) fetal groups, and p values for comparison between the two groups Normal fetal group
Maternal age (years) Gestation at MR scan (weeks) Gestation at delivery (weeks) Birthweight (kilograms)
Abnormal fetal group
Mean
Range
Mean
Range
p value*
32 34 39 3.29
(22–44) (27–40) (37–42) (2.76–4.24)
24 33 36 2.63
(18–35) (28–39) (29–39) (1.41–3.98)
0.011 0.271 0.001 0.004
*Student’s two-sided t-test.
safety of MRI in general and particularly in obstetrics.12,13 This has led to the establishment of guidelines limiting radio frequency exposure and the switching rate and strength of the magnetic fields employed. For the time being approval for MRI during pregnancy has not been given by the U.S. Food and Drug Administration (FDA).14 The National Radiological Protection Board (NRPB) of the United Kingdom15 advises that, as a precaution, MRI is only performed after the first trimester of pregnancy. Gradient echo sequences allow a series of MR images to be obtained in short imaging times16 and have recently been employed in this laboratory to estimate fetal volume and fetal organ volume in four normal pregnancies.17,18 Our present investigation has two main purposes. Firstly, we have applied the Cavalieri method of modern design stereology17,19,20 in combination with MRI to investigate the relationship between fetal volume and gestational age and between fetal brain volume and gestational age in normal pregnancy. Secondly, we have explored whether measurements of fetal volume and fetal brain volume can provide a noninvasive indication of fetal well being. MATERIALS AND METHODS Approval for this study was obtained from the local Ethics Committee and each woman provided fully informed written consent to participate. Singleton pregnant women were recruited from the antenatal wards and clinics. A group of pregnant women in whom there was a known fetal abnormality were also recruited. However, all pregnancies with maternal complications were excluded from the investigation. The normal fetal group comprised 18 uncomplicated pregnancies in all of which fetal birthweight was between the 10 –90th centile based on Liverpool data. The abnormal fetal group comprised eight pregnancies in which fetal abnormalities have been clinically or pathologically verified. In four cases birthweight was below the 10th centile. The clinical features of the pregnancies studied are described in Table 1. Gestational age was established from a knowledge of the date of the last menstrual period before pregnancy and refers to the time of conception.
In the normal fetal group, two pregnancies were imaged on two occasions and one pregnancy was imaged on four occasions. In the abnormal fetal group, two women had fetuses with gastroschisis (gs); one fetus had a diaphragmatic hernia (dh); one had an abnormality confirmed at delivery to be microhydrancephaly (mh); one fetus was hydropic with cerebral cystic lesions (hcc); one fetus had mild hydrops and cystic scalp lesions of undetermined nature (hcs); one fetus had exompholocele (ep) and one had clinically definite evidence of faltered growth (fg). MR images were acquired on a 1.5 T SIGNA whole body imaging system (GE Medical Systems, Milwaukee, Wisconsin, USA) with a standard body coil. A gradient recalled acquisition in the steady state (GRASS) sequence was used with repetition time (TR) of 25 ms, echo time (TE) of 12 ms and a flip angle of 10°, as previously described.16 Section thickness was 10 mm with a gap of 15 mm between consecutive sections. A field of view (FOV) of 48 cm was used with an acquisition matrix comprising 256 readings of 128 phase encoding steps. The display matrix was 256 3 256 pixels. One image slice was produced in 3 s, so that a series of 10 to 20 images took between 30 s and 1 min to acquire. Each woman assumed the left lateral position during imaging to prevent aortocaval compression. Foam padding was used to support the pregnant uterus and to enable the pregnant woman to maintain her position comfortably. MR images were obtained for maternal axial, sagittal and coronal orientations that corresponded approximately to axial, sagittal, and coronal orientations in the fetus. The best slice orientation for depicting fetal anatomy was selected for volume estimation. The MRI sequences were monitored to ensure that exposure to radio frequency energy and magnetic field gradients remained within the guidelines laid down by the NRPB.15 Total examination time was approximately 20 min. MR images were transferred to a SPARC 10 workstation (SUN Microsystems, California, USA) and input to ANALYZE image analysis software (Mayo Foundation, Minnesota, USA)21 for fetal volume and fetal brain volume estimation by the Cavalieri method. Point count-
Estimation of fetal and fetal brain volume ● Q.Y. GONG
ET AL.
237
Fig. 1. To estimate total fetal volume or fetal brain volume, a stereological test system for point counting, available in ANALYZE software (MAYO Foundation, Minnesota, USA),21 is overlaid with uniform random position on the relevant MR images. Each cross (1) signifies one test point. The left-most and middle panels refer to the same, approximately sagittal, section through a normal fetus. The left-most panel shows the superimposed test system before point counting and in the middle panel the points overlying the fetal transects have been removed by clicking the computer mouse. The right-most panel refers to a different MR image section and illustrates the use of point counting to estimate fetal brain volume. The points overlying the fetal brain transect have been removed by clicking the computer mouse. In both cases, transect area is estimated as the sum of the relevant point counts multiplied by the area associated with each test point. Volume is subsequently obtained by multiplying the sum of the transect areas on each of the relevant systematic MR image sections (obtained with random starting position) by the distance between sections.
ing was performed by a single observer (QYG) who is an experienced radiologist. The volume of the fetal brain included the cerebral hemispheres, cerebral ventricles and cerebellum. Stereology may be defined as the statistical inference of geometric parameters from sampled information. The volume of an object can be estimated efficiently with no systematic error or sampling bias using the Cavalieri method. The method has recently been applied in several volumetric studies with MRI.17,19,20 Structure volume is estimated with known precision as the sum of the transect areas of the sections through the structure, multiplied by the constant gap between the MR sections acquired with a random starting position. The point counting method of determining section area consists of overlying the images entirely with a systematic array of test points and counting the number of occasions where a point lies within the transects through the structure of interest (Fig. 1). This approach is far more efficient than manual planimetry.22 The procedure used for investigating the optimal number of points to be counted per section and the optimal number of sections to be analysed per volume estimate, as well as the method of predicting the coefficient of error (CE) on the volume estimates due to the combined effects of sectioning and point counting, has been described in previous studies carried out in this laboratory.17,19 We have shown that fetal volume, or fetal brain volume, can be estimated with a CE of less than 5% by counting of the order 200 points on about seven systematic sections which exhaustively sample the fetus, or fetal brain, respectively.
Statistical analyses were performed using S-plus software (StatSci, Washington, USA). Pearson’s product moment correlation coefficient was calculated for correlation analysis and a weighted least squares regression analysis was used to examine the relationship between fetal volume and gestational age and between fetal brain volume and gestational age. Weighted regression analysis is used when the assumption of variance homogeneity does not hold. The general approach is to weight each value in inverse proportion to its variance.23 In the present study, the reciprocals of the square of the standard error derived from the predicted CE were used as weights for the analysis of the relationship between fetal volume, or fetal brain volume, and gestational age. RESULTS In the normal fetal group, estimated volume ranged from 1273.0 to 3781.0 mL with a standard deviation of 664.3 mL. The mean predicted CE on the volume estimates was 7.3% with a standard deviation of 2.4%. Estimated total fetal volumes are plotted against gestational age in Fig. 2. The correlation of total fetal volume with gestational age was highly significant (r 5 0.94, p , 0.01). Weighted regression analysis provided the following linear relation between fetal volume and gestational age fetal volume (mL) 5 176.0 3 gestational age (weeks) 2 3621.0
Magnetic Resonance Imaging ● Volume 16, Number 3, 1998
238
In the abnormal fetal group, the total fetal volume estimates ranged from 392.8 –2046.0 mL with a mean CE of 6.8% (SD 5 1.8%). Corresponding fetal brain volumes ranged from 63.3 to 250.5 mL with a mean CE of 6.4% (SD 5 2.0%). For seven out of the eight fetuses total fetal volume was found to be small for gestational age, with measurements falling outside the 95% confidence interval established for the normal fetal group (Fig. 2). In contrast, for fetal brain only three out of the eight measurements fell outside the 95% confidence interval established from the normal fetal group (Fig. 3). The ratio of brain volume to total fetal volume in the normal fetal group ranged from 8% to 15% with a mean of 10% (SD 5 2%). There was a tendency for the ratio to decrease with increasing gestational age (Fig. 4) but this was not significant (p 5 0.38). DISCUSSION
Fig. 2. Estimates of total fetal volume are plotted against gestational age. The weighted-least-squares regression line (solid line) is fitted with 95% confidence limits (dotted lines) refers to the normal fetal group. F 5 normal fetal group; ƒ 5 abnormal fetal group. The meaning of the abbreviations used to denote the abnormal fetuses is given in the text.
(R2 5 0.88, p , 0.01), indicating that the daily increase in total fetal volume during the third trimester in normal fetuses is 25.2 mL. Based on the weighted linear fit, the predicted fetal volume at birth for the normal fetuses lay between 2892.0 and 3772.0 mL, with a standard deviation of 236.2 mL. Assuming a fetal density of 1.0 g/mL, the median difference between actual and estimated fetal birthweight, expressed as a percentage of the actual birthweight, was 1.7% (range from 213.7% to 30.4%). Application of Student’s two-sided t-test revealed no significant difference between the predicted fetal volume and actual birth weight (p . 0.05). Brain volume in the normal fetal group ranged from 121.5 to 371.2 mL (mean 5 255.4 6 65.0 mL). The mean predicted CE on the volume estimates was 6.9% with a standard deviation of 2.3%. Estimated brain volume is plotted against gestational age in Fig. 3. Weighted regression analysis provided the following linear relation between fetal brain volume and gestational age
This study has demonstrated a good correlation between fetal volume and gestational age (r 5 0.94, p , 0.01) in the normal fetus with a linear relationship (R2 5 0.88, p , 0.01) in the third trimester. The measured rate of increase of fetal volume of 25.2 mL per day is almost identical to the average of 25.1 mL per day obtained for a smaller group of subjects in a previous study18 and, with the fetal density assumed to be unity, is consistent with the average weight gain of 20 –30 g per day typically measured for the postnatal period.24 The corre-
fetal brain volume (mL) 5 15.8 3 gestational age (week) 2288.2 (R2 5 0.71, p , 0.01), indicating that the daily increase in fetal brain volume during the third trimester in normal fetuses is 2.3 mL.
Fig. 3. Estimates of fetal brain volume are plotted against gestational age. The weighted-least-square regression line (solid line) is fitted with 95% confidence limits (dotted lines) refers to the normal fetal group. F 5 normal fetal group; ƒ 5 abnormal fetal group. The meaning of the abbreviations used to denote the abnormal fetuses is given in the text.
Estimation of fetal and fetal brain volume ● Q.Y. GONG
Fig. 4. The ratio of fetal brain volume to total fetal volume is plotted against gestational age for the normal fetal group. No significant relationship is demonstrated (p . 0.05).
sponding rate of increase of brain volume observed for the normal fetus during the third trimester was 2.3 mL per day. On average, fetal brain volume remained a constant proportion (10%, SD 5 2%) of total fetal volume throughout the third trimester. A previous study reported a good correlation (R2 5 0.97) between birthweight and fetal volume estimated within one week of delivery.10 In the present study MRI was performed as many as 10 weeks before delivery. However, we were able to predict fetal volume at delivery using the results of the weighted linear regression analysis of fetal volume and gestational age for the normal fetal group. The correlation between predicted volume and birthweight is highly significant (r 5 0.57, p , 0.01). With the fetal density assumed to be unity, a two sided t-test shows no statistical difference between predicted and actual birthweight (p . 0.05). An important finding of this study is that volume estimation, performed several weeks before delivery, may be able to identify fetuses with growth retardation. Small-for-gestation fetuses25–27 are not necessarily associated with growth retardation, but it is advisable that they are identified and closely monitored.28 Although growth retardation can frequently be identified using one- or two-dimensional parameters derived from ultrasound images, the three-dimensional volume measurement provided by MRI is likely to provide a more sensitive index. Fetal brain volume is, like total fetal volume, closely related to gestational age (R2 5 0.71). Baker et al.11 applied planimetric techniques to estimate fetal brain volume from
ET AL.
239
echo planar MR images obtained in pregnancies with complications and also found a linear relation with gestational age (R2 5 0.79). Measurement of fetal brain volume was reported to be unhelpful in identifying fetuses subsequently delivered with individualised birthweight ratio above and below the 10th centile.11 This is in agreement with the present study in which fetal brain volume could not always be used to identify abnormal fetuses. Five of the eight abnormal fetuses had brain volumes inside the 95% CIV established from normal fetuses. The fact that total fetal volume is more sensitive than brain volume in distinguishing between normal and abnormal fetuses may be due to relative sparing of the brain in the abnormal fetus. Two types of abnormal growth patterns are recognised. In symmetrical growth, all of the body organs show similar proportional reductions in volume. In asymmetrical growth the volumes of all organs are reduced except the brain. Symmetrical and asymmetrical growth retardation are thought to arise from unfavourable effects on growth that respectively occur early and late in pregnancy.26,28 The organ depicted in black in the centre of the left-most panel of Fig. 1 is the fetal liver. Measurement of the length of the fetal liver on MR images has been reported to be helpful in identifying fetuses subsequently delivered with abnormal individualised birthweight ratio.11 The incorporation of the measurement of fetal liver volume18 alongside fetal brain volume and total fetal volume represents an obvious refinement in the use of MRI for distinguishing between normal and abnormal fetal growth, and also for distinguishing between symmetrical and asymmetrical growth retardation in utero. However, further studies of normal pregnancies will be required so that reliable data are obtained regarding the increase in fetal brain and liver volume and total fetal volume during the third trimester. Fetal motion artefacts were occasionally present on the GRASS images obtained in the present study. Maternal sedation with diazepam has been successfully used to suppress fetal motion but is not recommended in routine practice.7 Fetal motion artefact can be effectively removed by using EPI sequences. Other motion artefacts arise from maternal respiratory and cardiac movement and flow, which can be reduced to a minimum with MR gating techniques to compensate for individual breathing, cardiac and pulse-rate patterns.29 In conclusion, our results indicate that MRI in conjunction with stereology provides a useful tool for studying fetal growth in utero. A linear relationship (p , 0.01) exists between fetal volume and gestational age. Total volume estimates obtained for fetuses with abnormalities were, unlike brain volume, frequently significantly reduced relative to normal fetuses. Estimation of fetal volume and fetal organ volume by MRI and stereology
240
Magnetic Resonance Imaging ● Volume 16, Number 3, 1998
provides useful information for differentiating between asymmetrical and symmetrical growth retardation. Although it is unlikely that MRI will replace ultrasonography as a primary obstetric imaging tool, because of its high running cost and current lack of real-time capability, MRI is complementary to ultrasonography and provides a more convenient opportunity for obtaining volume measurements. Acknowledgment–We are grateful to Ms. Julie M. Walton for assistance in obtaining fetal birth weights and related information.
REFERENCES 1. Manara, L.R. Intrapartum fetal morbidity and mortality in intrauterine growth retarded infants. J. Am. Osteopath. Assoc. 80:101–104; 1980. 2. Dobson, P.C.; Abell, D.A.; Beischer, N.A. Mortality and morbidity of fetal growth retardation. Aust. NZ J. Obstet. Gynaecol. 21:69 –72; 1981. 3. Romero, R.; Jeanty, P. The detection of fetal growth disorders. Semin. Ultrasound CT MR 5:130 –143; 1984. 4. Nzeh, D.A.; Rimmer, S.; Moore, W.M.O.; Hunt, L. Prediction of birthweight by fetal ultrasound biometry. Br. J. Radiol. 66:987–989; 1992. 5. Lee, A.; Kratochwil, A.; Stumpflen, I. Fetal lung volume determination by three-dimensional ultrasonography. Am. J. Obstet. Gynecol. 175:588 –592; 1996. 6. McCarthy, S.M.; Filly, R.A.; Stark, D.D.; Hricak, H.; Brant-Zawadzki, M.N.; Callen, P.W.; Higgings, C.B. Obstetrical magnetic resonance imaging: Fetal anatomy. Radiology. 154:427– 432; 1985. 7. Winreb, J.C.; Lowe, T.; Cohen, J.M.; Kutler, M. Human fetal anatomy: MR imaging. Radiology. 157:715–720; 1985. 8. Powell, M.C.; Worthington, B.S.; Buckley, J.M.; Symonds, E.M. Magnetic resonance imaging (MRI). II Fetal anatomy. Br. J. Obstet. Gynaecol. 95:38 – 46; 1988. 9. Baker, P.N.; Johnson, I.R.; Gowland, P.A.; Freeman, A.; Adams, V.; Mansfield, P. Estimation of fetal lung volume using echo-planar magnetic resonance imaging. Obstet. Gynecol. 83:951–954; 1994. 10. Baker, P.N.; Johnson, I.R.; Gowland, P.A.; Hykin, J.; Harvey, P.R.; Freeman, A.; Adams, V.; Worthington, B.S.; Mansfield, P. Fetal weight estimation by echo-planar magnetic resonance imaging. Lancet. 343:644 – 645; 1994. 11. Baker, P.N.; Johnson, I.R.; Gowland, P.A.; Hykin, J.; Adams, V.; Mansfield, P.; Worthington, B.S. Measurement of fetal liver, brain and placental volumes with echo-planar magnetic resonance imaging. Br. J. Obstet. Gynaecol. 102: 35–39; 1995. 12. Johnson, I.R.; Symonds, E.M.; Kean, D.M.; Worthington, B.S.; Pipkin, F.B.; Hawkes, R.C.; Gyngell, M. Imaging the pregnant human uterus with magnetic resonance. Am. J. Obstet. Gynecol. 148:1136 –1139; 1984. 13. McRobbie, D.; Foster, M.A. Pulsed magnetic field exposure during pregnancy and implications for NMR fetal imaging: A study with mice. Magn. Reson. Imaging. 3:231–234; 1985.
14. U.S. Food and Drug Administration. Magnetic resonance diagnostic device: Panel recommendation and report on petitions for MR reclassification. Fed. Reg. 53:7575–7579; 1988. 15. National Radiological Protection Board ad hoc advisory group on nuclear magnetic resonance imaging. Revised guidelines on acceptable limits of exposure during nuclear magnetic resonance clinical imaging. Br. J. Radiol. 56: 974 –977; 1983. 16. Garden, A.S.; Griffiths, R.D.; Weindling, A.M.; Martin, P.A. Fast scan magnetic resonance imaging in fetal visualization. Am. J. Obstet. Gynecol. 164:1190 –1196; 1991. 17. Roberts, N.; Garden, A.S.; Cruz-Orive, L.M.; Whitehouse, G.H.; Edwards, R.H.T. Estimation of fetal volume by magnetic resonance imaging and stereology. Br. J. Radiol. 67:1067–1077; 1994. 18. Garden, A.S.; Roberts, N. Fetal and fetal organ volume estimations with magnetic resonance imaging. Am. J. Obstet. Gynecol. 175:442– 448; 1996. 19. Roberts, N.; Cruz-Orive, L.M.; Reid, N.M.K.; Brodie, D.A.; Bourne, M.; Edwards, R.H.T. Unbiased estimation of human body composition by the Cavalieri method using magnetic resonance imaging. J. Micros. 171:239 –253; 1993. 20. Light, K.; Roberts, N.; Whitehouse, G.H.; Edwards, R.H.T. Unbiased and efficient estimation of bladder volume by magnetic resonance imaging (MRI). J. Mag. Reson. Imaging. 5:33– 42; 1995. 21. Robb, R.A. A software system for interactive and quantitative analysis of biomedical images. In: K.H. Hone, H. Fuchs, S.M. Pizer (Eds). 3D Imaging in Medicine. NATO ASI Series, f(60), 1990: pp. 333–361. 22. Mathieu, O.; Cruz-Orive, L.M.; Hoppeler, H.; Weibel, E.R. Measuring error and sampling variation in stereology: Comparison of the efficiency of various methods for planar image analysis. J. Micros. 121:301–13; 1981. 23. Carroll, R.J.; Ruppert, D. Transformation and weighting in regression-monographs on statistics and applied probability. New York: Chapman and Hall; 1988. 24. Usher, R.; McLean, F. Normal fetal growth and the significance of fetal growth retardation. In: J. Davis, J. Dobbin (Eds). Scientific Foundations of Pediatrics. London: William Heinemann Medical Books Ltd., 1974: pp. 69 – 80. 25. Gablraith, R.S.; Karchmar, E.J.; Piercy, W.N.; Low, J.A. The clinical prediction of intrauterine growth retardation. Am. J. Obstet. Gynecol. 133:281–286; 1979. 26. Doubilet, P.M.; Benson, C.B. Fetal growth disturbances. Semin. Roentgenol. 15:309 –316; 1990. 27. Chang, T.C.; Robson, S.C.; Boys, R.J.; Spencer, J.A.D. Prediction of the small for gestational age infant: Which ultrasonic measurement is best? Obstet. Gynecol. 80:1030 –1038; 1992. 28. Stewart, A. Fetal growth: mortality and morbidity. In: F. Sharp, R.D.G. Milner, R.B. Fraser (Eds). Fetal Growth. London, England: Peacock Press, 1989: pp. 403– 412. 29. Wendt, R.E. III; Rokey, R.; Vick, G.W.; Johnston, R.L. Electrocadiographic gating and monitoring in NMR imaging. Magn. Reson. Imaging. 6:89 –95; 1988.
PII S0730-725X(97)00281-6
● Original Contribution
FETAL AND FETAL BRAIN VOLUME ESTIMATION IN THE THIRD TRIMESTER OF HUMAN PREGNANCY USING GRADIENT ECHO MR IMAGING Q.Y. GONG,*† N. ROBERTS,† A.S. GARDEN,‡
AND
G.H. WHITEHOUSE*†
*Department of Medical Imaging, †Magnetic Resonance and Image Analysis Research Centre, and the ‡Department of Obstetrics and Gynaecology, University of Liverpool, Liverpool, UK The Cavalieri method has been applied in combination with gradient echo magnetic resonance imaging (MRI) to investigate the increase in the volume of the fetus and fetal brain in the third trimester of pregnancy. Eighteen women with singleton pregnancies were recruited. Birthweights for the fetuses all lay within the 10 –90th centile based on Liverpool data. A regression analysis, weighted using values derived from the coefficient of error predicted for each volume estimate, revealed a linear relationship between total fetal volume and gestational age (R2 5 0.88) and between fetal brain volume and gestational age (R2 5 0.71) during the third trimester. Fetal volume increased by an average of 25.2 ml per day and fetal brain volume increased by an average of 2.3 mL per day. Fetal brain volume is on average a constant proportion (10%, SD 5 2%) of total fetal volume throughout the third trimester. Volume data were also obtained for eight fetuses diagnosed as abnormal. The volume of seven of the eight abnormal fetuses fell outside the 95% confidence interval established from the data obtained for the normal fetuses. However, for only three of the eight abnormal fetuses did brain volume fall outside the 95% confidence interval established for normals, possibly due to brain sparing occurring in asymmetrical growth retardation. The volume of the fetus and fetal brain may be readily estimated directly using the Cavalieri method and magnetic resonance imaging. These parameters represent potentially useful information for assessing fetal growth. © 1998 Elsevier Science Inc. Keywords: Brain; Cavalieri method; Fetus; MRI; Point counting; Stereology; Volume
INTRODUCTION
sional parameters such as biparietal diameter, abdominal circumference, and femor length are obtained.4 Recently, attempts have been made to measure the volume of fetal organs with three-dimensional ultrasonography, but problems are encountered in patients with pronounced oligohydramnios.5 Magnetic resonance imaging (MRI) is not generally limited by large maternal body habitus or lack of amniotic fluid. MRI has been used to investigate fetal anatomy in utero6 – 8 and, with the development of fast MR imaging sequences, it has been possible to obtain measurements of fetal volume and fetal organ volume in vivo.9 –11 However, the echo planar imaging (EPI) sequence used is not yet widely available and generally only suspected abnormal pregnancies have been studied. Many studies have been carried out to assess the
The ability to evaluate fetal growth in utero is important in perinatal management. The identification of growth abnormalities can assist in decisions regarding fetal and maternal management and clinical intervention. Manera1 and Dobson et al.2 report that perinatal mortality is four to ten times greater in pregnancies complicated by intrauterine growth retardation (IUGR) compared with pregnancies in which intrauterine growth is normal. Identification of IUGR after 37 weeks gestation is an indication for delivery to reduce the chance of fetal mortality.3 Imaging plays an important role in assessing and monitoring fetal growth in utero. The most widely used imaging modality is ultrasonography. However, this method is not, in general, used to obtain measures of fetal volume. Instead, arbitrary one- and two-dimenRECEIVED 8/10/97; ACCEPTED 10/19/97. Address correspondence to Dr. Neil Roberts, Magnetic Resonance and Image Analysis Research Centre, University of
Liverpool, P.O. Box 147, Liverpool L69 3BX, UK. E-mail: [email protected] 235
Magnetic Resonance Imaging ● Volume 16, Number 3, 1998
236
Table 1. Characteristics of the pregnancies for the normal (18 cases) and abnormal (8 cases) fetal groups, and p values for comparison between the two groups Normal fetal group
Maternal age (years) Gestation at MR scan (weeks) Gestation at delivery (weeks) Birthweight (kilograms)
Abnormal fetal group
Mean
Range
Mean
Range
p value*
32 34 39 3.29
(22–44) (27–40) (37–42) (2.76–4.24)
24 33 36 2.63
(18–35) (28–39) (29–39) (1.41–3.98)
0.011 0.271 0.001 0.004
*Student’s two-sided t-test.
safety of MRI in general and particularly in obstetrics.12,13 This has led to the establishment of guidelines limiting radio frequency exposure and the switching rate and strength of the magnetic fields employed. For the time being approval for MRI during pregnancy has not been given by the U.S. Food and Drug Administration (FDA).14 The National Radiological Protection Board (NRPB) of the United Kingdom15 advises that, as a precaution, MRI is only performed after the first trimester of pregnancy. Gradient echo sequences allow a series of MR images to be obtained in short imaging times16 and have recently been employed in this laboratory to estimate fetal volume and fetal organ volume in four normal pregnancies.17,18 Our present investigation has two main purposes. Firstly, we have applied the Cavalieri method of modern design stereology17,19,20 in combination with MRI to investigate the relationship between fetal volume and gestational age and between fetal brain volume and gestational age in normal pregnancy. Secondly, we have explored whether measurements of fetal volume and fetal brain volume can provide a noninvasive indication of fetal well being. MATERIALS AND METHODS Approval for this study was obtained from the local Ethics Committee and each woman provided fully informed written consent to participate. Singleton pregnant women were recruited from the antenatal wards and clinics. A group of pregnant women in whom there was a known fetal abnormality were also recruited. However, all pregnancies with maternal complications were excluded from the investigation. The normal fetal group comprised 18 uncomplicated pregnancies in all of which fetal birthweight was between the 10 –90th centile based on Liverpool data. The abnormal fetal group comprised eight pregnancies in which fetal abnormalities have been clinically or pathologically verified. In four cases birthweight was below the 10th centile. The clinical features of the pregnancies studied are described in Table 1. Gestational age was established from a knowledge of the date of the last menstrual period before pregnancy and refers to the time of conception.
In the normal fetal group, two pregnancies were imaged on two occasions and one pregnancy was imaged on four occasions. In the abnormal fetal group, two women had fetuses with gastroschisis (gs); one fetus had a diaphragmatic hernia (dh); one had an abnormality confirmed at delivery to be microhydrancephaly (mh); one fetus was hydropic with cerebral cystic lesions (hcc); one fetus had mild hydrops and cystic scalp lesions of undetermined nature (hcs); one fetus had exompholocele (ep) and one had clinically definite evidence of faltered growth (fg). MR images were acquired on a 1.5 T SIGNA whole body imaging system (GE Medical Systems, Milwaukee, Wisconsin, USA) with a standard body coil. A gradient recalled acquisition in the steady state (GRASS) sequence was used with repetition time (TR) of 25 ms, echo time (TE) of 12 ms and a flip angle of 10°, as previously described.16 Section thickness was 10 mm with a gap of 15 mm between consecutive sections. A field of view (FOV) of 48 cm was used with an acquisition matrix comprising 256 readings of 128 phase encoding steps. The display matrix was 256 3 256 pixels. One image slice was produced in 3 s, so that a series of 10 to 20 images took between 30 s and 1 min to acquire. Each woman assumed the left lateral position during imaging to prevent aortocaval compression. Foam padding was used to support the pregnant uterus and to enable the pregnant woman to maintain her position comfortably. MR images were obtained for maternal axial, sagittal and coronal orientations that corresponded approximately to axial, sagittal, and coronal orientations in the fetus. The best slice orientation for depicting fetal anatomy was selected for volume estimation. The MRI sequences were monitored to ensure that exposure to radio frequency energy and magnetic field gradients remained within the guidelines laid down by the NRPB.15 Total examination time was approximately 20 min. MR images were transferred to a SPARC 10 workstation (SUN Microsystems, California, USA) and input to ANALYZE image analysis software (Mayo Foundation, Minnesota, USA)21 for fetal volume and fetal brain volume estimation by the Cavalieri method. Point count-
Estimation of fetal and fetal brain volume ● Q.Y. GONG
ET AL.
237
Fig. 1. To estimate total fetal volume or fetal brain volume, a stereological test system for point counting, available in ANALYZE software (MAYO Foundation, Minnesota, USA),21 is overlaid with uniform random position on the relevant MR images. Each cross (1) signifies one test point. The left-most and middle panels refer to the same, approximately sagittal, section through a normal fetus. The left-most panel shows the superimposed test system before point counting and in the middle panel the points overlying the fetal transects have been removed by clicking the computer mouse. The right-most panel refers to a different MR image section and illustrates the use of point counting to estimate fetal brain volume. The points overlying the fetal brain transect have been removed by clicking the computer mouse. In both cases, transect area is estimated as the sum of the relevant point counts multiplied by the area associated with each test point. Volume is subsequently obtained by multiplying the sum of the transect areas on each of the relevant systematic MR image sections (obtained with random starting position) by the distance between sections.
ing was performed by a single observer (QYG) who is an experienced radiologist. The volume of the fetal brain included the cerebral hemispheres, cerebral ventricles and cerebellum. Stereology may be defined as the statistical inference of geometric parameters from sampled information. The volume of an object can be estimated efficiently with no systematic error or sampling bias using the Cavalieri method. The method has recently been applied in several volumetric studies with MRI.17,19,20 Structure volume is estimated with known precision as the sum of the transect areas of the sections through the structure, multiplied by the constant gap between the MR sections acquired with a random starting position. The point counting method of determining section area consists of overlying the images entirely with a systematic array of test points and counting the number of occasions where a point lies within the transects through the structure of interest (Fig. 1). This approach is far more efficient than manual planimetry.22 The procedure used for investigating the optimal number of points to be counted per section and the optimal number of sections to be analysed per volume estimate, as well as the method of predicting the coefficient of error (CE) on the volume estimates due to the combined effects of sectioning and point counting, has been described in previous studies carried out in this laboratory.17,19 We have shown that fetal volume, or fetal brain volume, can be estimated with a CE of less than 5% by counting of the order 200 points on about seven systematic sections which exhaustively sample the fetus, or fetal brain, respectively.
Statistical analyses were performed using S-plus software (StatSci, Washington, USA). Pearson’s product moment correlation coefficient was calculated for correlation analysis and a weighted least squares regression analysis was used to examine the relationship between fetal volume and gestational age and between fetal brain volume and gestational age. Weighted regression analysis is used when the assumption of variance homogeneity does not hold. The general approach is to weight each value in inverse proportion to its variance.23 In the present study, the reciprocals of the square of the standard error derived from the predicted CE were used as weights for the analysis of the relationship between fetal volume, or fetal brain volume, and gestational age. RESULTS In the normal fetal group, estimated volume ranged from 1273.0 to 3781.0 mL with a standard deviation of 664.3 mL. The mean predicted CE on the volume estimates was 7.3% with a standard deviation of 2.4%. Estimated total fetal volumes are plotted against gestational age in Fig. 2. The correlation of total fetal volume with gestational age was highly significant (r 5 0.94, p , 0.01). Weighted regression analysis provided the following linear relation between fetal volume and gestational age fetal volume (mL) 5 176.0 3 gestational age (weeks) 2 3621.0
Magnetic Resonance Imaging ● Volume 16, Number 3, 1998
238
In the abnormal fetal group, the total fetal volume estimates ranged from 392.8 –2046.0 mL with a mean CE of 6.8% (SD 5 1.8%). Corresponding fetal brain volumes ranged from 63.3 to 250.5 mL with a mean CE of 6.4% (SD 5 2.0%). For seven out of the eight fetuses total fetal volume was found to be small for gestational age, with measurements falling outside the 95% confidence interval established for the normal fetal group (Fig. 2). In contrast, for fetal brain only three out of the eight measurements fell outside the 95% confidence interval established from the normal fetal group (Fig. 3). The ratio of brain volume to total fetal volume in the normal fetal group ranged from 8% to 15% with a mean of 10% (SD 5 2%). There was a tendency for the ratio to decrease with increasing gestational age (Fig. 4) but this was not significant (p 5 0.38). DISCUSSION
Fig. 2. Estimates of total fetal volume are plotted against gestational age. The weighted-least-squares regression line (solid line) is fitted with 95% confidence limits (dotted lines) refers to the normal fetal group. F 5 normal fetal group; ƒ 5 abnormal fetal group. The meaning of the abbreviations used to denote the abnormal fetuses is given in the text.
(R2 5 0.88, p , 0.01), indicating that the daily increase in total fetal volume during the third trimester in normal fetuses is 25.2 mL. Based on the weighted linear fit, the predicted fetal volume at birth for the normal fetuses lay between 2892.0 and 3772.0 mL, with a standard deviation of 236.2 mL. Assuming a fetal density of 1.0 g/mL, the median difference between actual and estimated fetal birthweight, expressed as a percentage of the actual birthweight, was 1.7% (range from 213.7% to 30.4%). Application of Student’s two-sided t-test revealed no significant difference between the predicted fetal volume and actual birth weight (p . 0.05). Brain volume in the normal fetal group ranged from 121.5 to 371.2 mL (mean 5 255.4 6 65.0 mL). The mean predicted CE on the volume estimates was 6.9% with a standard deviation of 2.3%. Estimated brain volume is plotted against gestational age in Fig. 3. Weighted regression analysis provided the following linear relation between fetal brain volume and gestational age
This study has demonstrated a good correlation between fetal volume and gestational age (r 5 0.94, p , 0.01) in the normal fetus with a linear relationship (R2 5 0.88, p , 0.01) in the third trimester. The measured rate of increase of fetal volume of 25.2 mL per day is almost identical to the average of 25.1 mL per day obtained for a smaller group of subjects in a previous study18 and, with the fetal density assumed to be unity, is consistent with the average weight gain of 20 –30 g per day typically measured for the postnatal period.24 The corre-
fetal brain volume (mL) 5 15.8 3 gestational age (week) 2288.2 (R2 5 0.71, p , 0.01), indicating that the daily increase in fetal brain volume during the third trimester in normal fetuses is 2.3 mL.
Fig. 3. Estimates of fetal brain volume are plotted against gestational age. The weighted-least-square regression line (solid line) is fitted with 95% confidence limits (dotted lines) refers to the normal fetal group. F 5 normal fetal group; ƒ 5 abnormal fetal group. The meaning of the abbreviations used to denote the abnormal fetuses is given in the text.
Estimation of fetal and fetal brain volume ● Q.Y. GONG
Fig. 4. The ratio of fetal brain volume to total fetal volume is plotted against gestational age for the normal fetal group. No significant relationship is demonstrated (p . 0.05).
sponding rate of increase of brain volume observed for the normal fetus during the third trimester was 2.3 mL per day. On average, fetal brain volume remained a constant proportion (10%, SD 5 2%) of total fetal volume throughout the third trimester. A previous study reported a good correlation (R2 5 0.97) between birthweight and fetal volume estimated within one week of delivery.10 In the present study MRI was performed as many as 10 weeks before delivery. However, we were able to predict fetal volume at delivery using the results of the weighted linear regression analysis of fetal volume and gestational age for the normal fetal group. The correlation between predicted volume and birthweight is highly significant (r 5 0.57, p , 0.01). With the fetal density assumed to be unity, a two sided t-test shows no statistical difference between predicted and actual birthweight (p . 0.05). An important finding of this study is that volume estimation, performed several weeks before delivery, may be able to identify fetuses with growth retardation. Small-for-gestation fetuses25–27 are not necessarily associated with growth retardation, but it is advisable that they are identified and closely monitored.28 Although growth retardation can frequently be identified using one- or two-dimensional parameters derived from ultrasound images, the three-dimensional volume measurement provided by MRI is likely to provide a more sensitive index. Fetal brain volume is, like total fetal volume, closely related to gestational age (R2 5 0.71). Baker et al.11 applied planimetric techniques to estimate fetal brain volume from
ET AL.
239
echo planar MR images obtained in pregnancies with complications and also found a linear relation with gestational age (R2 5 0.79). Measurement of fetal brain volume was reported to be unhelpful in identifying fetuses subsequently delivered with individualised birthweight ratio above and below the 10th centile.11 This is in agreement with the present study in which fetal brain volume could not always be used to identify abnormal fetuses. Five of the eight abnormal fetuses had brain volumes inside the 95% CIV established from normal fetuses. The fact that total fetal volume is more sensitive than brain volume in distinguishing between normal and abnormal fetuses may be due to relative sparing of the brain in the abnormal fetus. Two types of abnormal growth patterns are recognised. In symmetrical growth, all of the body organs show similar proportional reductions in volume. In asymmetrical growth the volumes of all organs are reduced except the brain. Symmetrical and asymmetrical growth retardation are thought to arise from unfavourable effects on growth that respectively occur early and late in pregnancy.26,28 The organ depicted in black in the centre of the left-most panel of Fig. 1 is the fetal liver. Measurement of the length of the fetal liver on MR images has been reported to be helpful in identifying fetuses subsequently delivered with abnormal individualised birthweight ratio.11 The incorporation of the measurement of fetal liver volume18 alongside fetal brain volume and total fetal volume represents an obvious refinement in the use of MRI for distinguishing between normal and abnormal fetal growth, and also for distinguishing between symmetrical and asymmetrical growth retardation in utero. However, further studies of normal pregnancies will be required so that reliable data are obtained regarding the increase in fetal brain and liver volume and total fetal volume during the third trimester. Fetal motion artefacts were occasionally present on the GRASS images obtained in the present study. Maternal sedation with diazepam has been successfully used to suppress fetal motion but is not recommended in routine practice.7 Fetal motion artefact can be effectively removed by using EPI sequences. Other motion artefacts arise from maternal respiratory and cardiac movement and flow, which can be reduced to a minimum with MR gating techniques to compensate for individual breathing, cardiac and pulse-rate patterns.29 In conclusion, our results indicate that MRI in conjunction with stereology provides a useful tool for studying fetal growth in utero. A linear relationship (p , 0.01) exists between fetal volume and gestational age. Total volume estimates obtained for fetuses with abnormalities were, unlike brain volume, frequently significantly reduced relative to normal fetuses. Estimation of fetal volume and fetal organ volume by MRI and stereology
240
Magnetic Resonance Imaging ● Volume 16, Number 3, 1998
provides useful information for differentiating between asymmetrical and symmetrical growth retardation. Although it is unlikely that MRI will replace ultrasonography as a primary obstetric imaging tool, because of its high running cost and current lack of real-time capability, MRI is complementary to ultrasonography and provides a more convenient opportunity for obtaining volume measurements. Acknowledgment–We are grateful to Ms. Julie M. Walton for assistance in obtaining fetal birth weights and related information.
REFERENCES 1. Manara, L.R. Intrapartum fetal morbidity and mortality in intrauterine growth retarded infants. J. Am. Osteopath. Assoc. 80:101–104; 1980. 2. Dobson, P.C.; Abell, D.A.; Beischer, N.A. Mortality and morbidity of fetal growth retardation. Aust. NZ J. Obstet. Gynaecol. 21:69 –72; 1981. 3. Romero, R.; Jeanty, P. The detection of fetal growth disorders. Semin. Ultrasound CT MR 5:130 –143; 1984. 4. Nzeh, D.A.; Rimmer, S.; Moore, W.M.O.; Hunt, L. Prediction of birthweight by fetal ultrasound biometry. Br. J. Radiol. 66:987–989; 1992. 5. Lee, A.; Kratochwil, A.; Stumpflen, I. Fetal lung volume determination by three-dimensional ultrasonography. Am. J. Obstet. Gynecol. 175:588 –592; 1996. 6. McCarthy, S.M.; Filly, R.A.; Stark, D.D.; Hricak, H.; Brant-Zawadzki, M.N.; Callen, P.W.; Higgings, C.B. Obstetrical magnetic resonance imaging: Fetal anatomy. Radiology. 154:427– 432; 1985. 7. Winreb, J.C.; Lowe, T.; Cohen, J.M.; Kutler, M. Human fetal anatomy: MR imaging. Radiology. 157:715–720; 1985. 8. Powell, M.C.; Worthington, B.S.; Buckley, J.M.; Symonds, E.M. Magnetic resonance imaging (MRI). II Fetal anatomy. Br. J. Obstet. Gynaecol. 95:38 – 46; 1988. 9. Baker, P.N.; Johnson, I.R.; Gowland, P.A.; Freeman, A.; Adams, V.; Mansfield, P. Estimation of fetal lung volume using echo-planar magnetic resonance imaging. Obstet. Gynecol. 83:951–954; 1994. 10. Baker, P.N.; Johnson, I.R.; Gowland, P.A.; Hykin, J.; Harvey, P.R.; Freeman, A.; Adams, V.; Worthington, B.S.; Mansfield, P. Fetal weight estimation by echo-planar magnetic resonance imaging. Lancet. 343:644 – 645; 1994. 11. Baker, P.N.; Johnson, I.R.; Gowland, P.A.; Hykin, J.; Adams, V.; Mansfield, P.; Worthington, B.S. Measurement of fetal liver, brain and placental volumes with echo-planar magnetic resonance imaging. Br. J. Obstet. Gynaecol. 102: 35–39; 1995. 12. Johnson, I.R.; Symonds, E.M.; Kean, D.M.; Worthington, B.S.; Pipkin, F.B.; Hawkes, R.C.; Gyngell, M. Imaging the pregnant human uterus with magnetic resonance. Am. J. Obstet. Gynecol. 148:1136 –1139; 1984. 13. McRobbie, D.; Foster, M.A. Pulsed magnetic field exposure during pregnancy and implications for NMR fetal imaging: A study with mice. Magn. Reson. Imaging. 3:231–234; 1985.
14. U.S. Food and Drug Administration. Magnetic resonance diagnostic device: Panel recommendation and report on petitions for MR reclassification. Fed. Reg. 53:7575–7579; 1988. 15. National Radiological Protection Board ad hoc advisory group on nuclear magnetic resonance imaging. Revised guidelines on acceptable limits of exposure during nuclear magnetic resonance clinical imaging. Br. J. Radiol. 56: 974 –977; 1983. 16. Garden, A.S.; Griffiths, R.D.; Weindling, A.M.; Martin, P.A. Fast scan magnetic resonance imaging in fetal visualization. Am. J. Obstet. Gynecol. 164:1190 –1196; 1991. 17. Roberts, N.; Garden, A.S.; Cruz-Orive, L.M.; Whitehouse, G.H.; Edwards, R.H.T. Estimation of fetal volume by magnetic resonance imaging and stereology. Br. J. Radiol. 67:1067–1077; 1994. 18. Garden, A.S.; Roberts, N. Fetal and fetal organ volume estimations with magnetic resonance imaging. Am. J. Obstet. Gynecol. 175:442– 448; 1996. 19. Roberts, N.; Cruz-Orive, L.M.; Reid, N.M.K.; Brodie, D.A.; Bourne, M.; Edwards, R.H.T. Unbiased estimation of human body composition by the Cavalieri method using magnetic resonance imaging. J. Micros. 171:239 –253; 1993. 20. Light, K.; Roberts, N.; Whitehouse, G.H.; Edwards, R.H.T. Unbiased and efficient estimation of bladder volume by magnetic resonance imaging (MRI). J. Mag. Reson. Imaging. 5:33– 42; 1995. 21. Robb, R.A. A software system for interactive and quantitative analysis of biomedical images. In: K.H. Hone, H. Fuchs, S.M. Pizer (Eds). 3D Imaging in Medicine. NATO ASI Series, f(60), 1990: pp. 333–361. 22. Mathieu, O.; Cruz-Orive, L.M.; Hoppeler, H.; Weibel, E.R. Measuring error and sampling variation in stereology: Comparison of the efficiency of various methods for planar image analysis. J. Micros. 121:301–13; 1981. 23. Carroll, R.J.; Ruppert, D. Transformation and weighting in regression-monographs on statistics and applied probability. New York: Chapman and Hall; 1988. 24. Usher, R.; McLean, F. Normal fetal growth and the significance of fetal growth retardation. In: J. Davis, J. Dobbin (Eds). Scientific Foundations of Pediatrics. London: William Heinemann Medical Books Ltd., 1974: pp. 69 – 80. 25. Gablraith, R.S.; Karchmar, E.J.; Piercy, W.N.; Low, J.A. The clinical prediction of intrauterine growth retardation. Am. J. Obstet. Gynecol. 133:281–286; 1979. 26. Doubilet, P.M.; Benson, C.B. Fetal growth disturbances. Semin. Roentgenol. 15:309 –316; 1990. 27. Chang, T.C.; Robson, S.C.; Boys, R.J.; Spencer, J.A.D. Prediction of the small for gestational age infant: Which ultrasonic measurement is best? Obstet. Gynecol. 80:1030 –1038; 1992. 28. Stewart, A. Fetal growth: mortality and morbidity. In: F. Sharp, R.D.G. Milner, R.B. Fraser (Eds). Fetal Growth. London, England: Peacock Press, 1989: pp. 403– 412. 29. Wendt, R.E. III; Rokey, R.; Vick, G.W.; Johnston, R.L. Electrocadiographic gating and monitoring in NMR imaging. Magn. Reson. Imaging. 6:89 –95; 1988.