| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, University of Texas Southwestern Medical Center, 5323
Harry Hines Blvd., Dallas, TX 75390-8896.
2 Department of Obstetrics and Gynecology, University of Texas Southwestern
Medical Center, Dallas, TX 75390-9032.
Received November 3, 2003; accepted after revision February 2, 2004.
Address correspondence to D. M. Twickler.
OBJECTIVE. The purpose of this study was to determine whether differences are seen in calculation of fetal weight using 5-mm sagittal, 3-mm coronal, and 8-mm axial MRI acquisitions compared with term birth weight and contemporaneous sonography.
MATERIALS AND METHODS. Fetal volume measurements were obtained from MRI acquisitions as follows: 5-mm sagittal (2 acquisitions), 3-mm coronal (2 acquisitions), and 8-mm axial (1 acquisition). A 90-sec single-shot fast spin-echo sequence was used. MRI and sonographic studies for fetal weight estimates were performed within 3 hr of term delivery. MRI calculation was based on the equation 0.12 + 1.031 x fetal volume (fetal area x slice thickness) (mL) = MRI fetal weight (kg). The sonographic fetal weight estimate was calculated using the Hadlock formula. MRI and sonographic calculations were compared with birth weight. Concordance coefficient analysis was performed.
RESULTS. Thirty-five retrospective fetal calculations were performed. Concordance coefficients, gram weight means and standard deviations (mean ± SD) between birth weight and MRI acquisitions were as follows: 8-mm axial, 0.91 (3,554 ± 431 g); 3-mm coronal, 0.84 (3,752 ± 578 g); and 5-mm sagittal, 0.83 (3,685 ± 567 g), compared with 0.78 (3,518 ± 332 g) for sonography. The MRI axial concordance coefficient was significantly different from that of the sonographic estimates (p = 0.05). MRI axial concordance coefficient was not statistically different from that of the MRI coronal concordance coefficient (p = 0.22) or the MRI sagittal concordance coefficient (p = 0.19).
CONCLUSION. Calculated weights from a 90-sec single-shot fast spin-echo sequence MR acquisition with 8-mm-thick slices in the axial plane at term are better than sonographic estimates.
Sonography is a widely used imaging technique in obstetrics for numerous fetal and maternal indications. It is currently the screening technique of choice for fetal imaging and estimating antenatal fetal weight because it is commonly available and relatively inexpensive. Fetal MRI is increasingly being used for fetal evaluation as an adjunct to sonography to provide further information regarding fetal anatomy and disorders that may alter the antenatal diagnosis and management of the pregnancy [1-4]. Advances in MRI have made fetal imaging practical by reducing motion artifact with ultrafast acquisitions delivering superior resolution, including single-shot fast spin-echo and HASTE sequences [5, 6].
In addition to evaluation of fetal anatomy and disorders on fetal MRI, estimates of fetal organ volumes and fetal weight have been performed with a high degree of accuracy [6-10]. Accurate estimates of fetal weight before delivery may be helpful in the management of labor, particularly in the settings of suspected macrosomia and intrauterine growth retardation [10]. MRI fetal weight estimates have been reported to have a higher concordance with actual birth weight than sonography does [10-12]. Advantages of MRI compared with other imaging techniques include the ability to obtain multiplanar acquisitions and its theoretic improved resolution. Although it is feasible to calculate MRI fetal weight in sagittal, coronal, or axial planes, to our knowledge, the recommended plane of imaging, slice thickness, and associated number of acquisitions that most accurately determine term fetal weight have not yet been established or published. We therefore sought to determine whether there are differences in fetal weight calculation based on plane of imaging or thickness by comparing sagittal 5-mm, coronal 3-mm, and axial 8-mm MRI acquisitions with term birth weight and sonography fetal weight calculations. Our goal was to help establish an optimal, practical protocol for fetal MRI in the prediction of fetal weight in the term infant.
Materials and Methods
Patients and Imaging
This retrospective study reviewed fetal MRI studies performed between
January 2001 and January 2002 and was part of an ongoing study to evaluate MRI
in term fetuses. Patients were enrolled in this institutional review
board-approved study at the time they were scheduled for their elective repeat
cesarean delivery in the obstetrics clinic or in the preoperative area on the
day of surgery. Patients were eligible for enrollment if they were at term,
required a repeat cesarean delivery, and had an uncomplicated prenatal course.
All patients underwent MRI and sonography to estimate fetal weight within 3 hr
before delivery, and actual fetal weight was recorded at the time of delivery
as part of an ongoing study.
Before the images were acquired, each maternal patient was counseled about fetal safety issues and written informed consent was obtained. MRI was performed using a 1.5-T Signa magnet (GE Healthcare). A torso surface coil was placed around the mother's pelvis and centered over the gravid uterus. No maternal sedation was administered because the short acquisition times of the sequences limited fetal motion as a potential artifact. A 15-sec localizer three-plane gradient-echo T2*-weighted sequence was obtained to plan the orthogonal planes relative to the lie of the fetus. Then a 90-sec single-shot fast spin-echo sequence was used to obtain the fetal images. Two acquisitions each were required to obtain 5-mm sagittal images and 3-mm coronal images, for a total of approximately 75-80 images per patient. One acquisition was required to obtain 8-mm axial images, for a total of approximately 40 images per patient. There were no gaps in the acquisitions. The orthogonal MRI planes were in relation to the term uterus. The parameters used to obtain the images were the following: TR range/TEeff range, 30,000-98,000/50-100; field of view, 30-48 cm (average, 40 cm); matrix, 256 x 128 or 512 x 256; bandwidth, 31.2 or 62.5 kHz; number of excitations, 0.5; interecho spacing, 4.5 msec; and slice thickness, 3, 5, or 8 mm.
Sonography was performed by a single experienced sonographer (who was blinded to MRI and actual birth weight information) using an XP/10 or Acuson Sequoia scanner (Siemens) or an Elegra scanner with software version 6 (Siemens) with 3- or 5-MHz curved linear transducers. Biparietal diameter, head circumference, abdominal circumference, and femur length were measured.
Image Analysis
Region-of-interest (ROI) tracings were drawn around fetal parts both
freehand and by using the ellipse function on an Advantage Windows
postprocessing workstation (GE Healthcare) on fetal MRI studies including
axial, sagittal, and coronal images (Figs.
1A,
1B,
1C,
1D,
2A,
2B,
2C,
2D,
3A,
3B,
3C,
3D). The single observer was
not aware of actual birth weight or fetal weight estimated on sonography. MRI
volume for each plane of imaging for each fetus was determined by the sum of
ROI area measurements. Fetal weight was then calculated with the following
formula [10]:
 |
|
|
|
|
|
|
|
|
|
|
|
|
Fetal lie on MRI was also documented. Estimated fetal weight on sonography was calculated using the formula of Hadlock et al. [13, 14]. Estimates of fetal weight on MRI and sonography were compared with actual recorded birth weights. The data were analyzed using concordance coefficients. Confidence intervals for the correlations were based on the Fisher's z-transformation, and comparisons between correlations were performed using the Student's t test on the Fisher's z-transformations. A p value less than or equal to 0.05 was considered statistically significant.
Results
Thirty-five fetal MRI and sonographic calculations were compared with actual fetal birth weights over a 1-year period. The patient demographics were 85% Hispanic, 9% African American, 4% Asian, and 2% white American. The average time for completion of these five acquisitions in each woman was approximately 9-12 min, with a total time to complete the study of 20-30 min, including getting the women on and off the MRI gantry. The average amount of postprocessing time for MRI weight assignment was approximately 30 min for the single axial acquisition and 1 hr each for the two sagittal and two coronal acquisitions. The average sonography time for the biometry and weight assessment was 45 min, including setting up the women in the sonography room.
Gram weight means and SDs and weight ranges for the 35 patients are presented in Table 1. All concordance correlations compared with term birth weight were significant (p < 0.01). The relationship and concordance coefficients between 8-mm axial MRI, 3-mm coronal MRI, 5-mm sagittal MRI, and sonographic weight estimates are presented in Figures 4, 5, 6, 7. The MRI 8-mm axial concordance coefficient was significantly different from that of the sonographic estimates (p = 0.05). However, the MRI 8-mm axial concordance coefficient was not statistically different from that of the MRI 3-mm coronal concordance coefficient (p = 0.22) or MRI 5-mm sagittal concordance coefficient (p = 0.19).
|
|
|
|
|
Thirty-two of 35 fetuses were breech or cephalic, with the transverse lie being much less frequent. Because of the small number of fetuses in the transverse lie, it was not possible to analyze this data set.
Discussion
To our knowledge, ours is the first study comparing MRI fetal weight estimates using 8-mm axial, 3-mm coronal, and 5-mm sagittal imaging planes within 3 hr before delivery with birth weight and contemporaneous sonography. In our retrospective study of 35 patients, we found an excellent correlation between actual term fetal birth weight and fetal weight estimates by all MRI acquisitions, with the single 90-sec 8-mm axial second single-shot fast spin-echo sequence having the best correlation coefficient and significantly better weight estimates than sonography. Although the 8-mm single axial MRI fetal weight estimates were statistically superior to those of sonography, there was no statistical difference compared with the other orthogonal MRI planes.
There were a larger than expected number of fetuses whose weight exceeded 4,000 g. As noted in the Materials and Methods and Results sections, the patient population consisted of women scheduled for elective repeat cesarean delivery, who by definition were multiparous and likely to have larger infants. Larger sizes and the ethnicity of the mothers also may have impacted the birth weights of the infants. The indications for the initial cesarean delivery were not available at the time the data were retrospectively reviewed. However, it is likely that some, if not many, patients were delivered by cesarean initially for dystocia that may have been related to large fetal size.
In spite of our finding of no statistically significant difference between the MRI orthogonal planes in estimating fetal weight, the axial MRI acquisition, in our anecdotal experience, was the better plane for estimating fetal birth weight at term for several reasons. Because of the widespread use of predominantly axial imaging in both CT and routine MRI examinations, radiologists and other interpreters of images are more familiar with the axial plane than with other planes. Because 32 of 35 fetuses were cephalic or breech at term, axial images of the fetus were obtained by acquiring axial images of the maternal abdomen. Not only was the axial plane more convenient and familiar for the observer, but it also allowed the observer to use the ellipse ROI function on the postprocessing workstation because many fetal parts (head, abdomen, thighs, arms) are elliptic in the axial imaging plane. The slice thickness of 8 mm required only a single acquisition to cover the entire gravid uterus and a postprocessing time of 30 min; two acquisitions were required to include the entire term fetus with slice thicknesses of 5 and 3 mm, and the postprocessing time was doubled.
A major limitation of our study is that both slice thickness and plane of imaging were varied in each acquisition sequence. Although there is hypothetic improved volume ascertainment with thinner slices, this did not occur in our series when performed in the coronal and sagittal planes. Axial-plane 5-mm and 3-mm zero-gap slices would have been ideal for comparison of thickness but were not performed in our series. Given the problems noted with acquiring two acquisitions in the other planes, it is doubtful that fetal weight estimates would improve with thinner axial slices.
The strengths of our study design and methodology are the use of a single experienced observer for each technique and comparison of fetal weight estimates using MRI and sonography with actual fetal birth weights obtained within 3 hr of those studies.
The equation used to calculate MRI fetal weights may not apply to earlier gestational ages or smaller fetuses in which less fat is present. Thus, future studies are needed to verify our findings in smaller fetuses at various gestational ages and fetuses with growth restriction. Future investigation of fetal MRI for weight estimates should also include designing and evaluating methods of obtaining thinner slices in shorter acquisition times as faster MRI scanners are developed. Other studies could evaluate different types of MRI sequences and semiautomatic or automatic postprocessing to greatly decrease postprocessing times and ultimately reduce cost. Less-labor-intensive postprocessing may result in more widespread use of fetal MRI volume and weight assessment for certain clinical fetal and maternal indications.
Our conclusion adds to the growing data showing the usefulness of fetal MRI, with applications in areas of fetal central nervous system and other abnormalities as well as organ volumes, including brain and lungs [8, 9, 15]. The significance of our study is that a single 90-sec axial 8-mm single-shot fast spin-echo MRI acquisition at term can be used to calculate exceptionally accurate fetal weight for those clinical situations in which more precise fetal weight than sonography offers is needed. Potential clinical scenarios for MRI calculated weight estimates include infants at risk for dystocia, such as occurs with maternal diabetes or postterm gestations, and growth-restricted infants, when timing and mode of delivery are affected by fetal weight.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. R. Hatab, M. V. Zaretsky, J. M. Alexander, and D. M. Twickler Comparison of Fetal Biometric Values with Sonographic and 3D Reconstruction MRI in Term Gestations Am. J. Roentgenol., August 1, 2008; 191(2): 340 - 345. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Cannie, J. C. Jani, F. Van Kerkhove, J. Meerschaert, F. De Keyzer, L. Lewi, J. A. Deprest, and S. Dymarkowski Fetal Body Volume at MR Imaging to Quantify Total Fetal Lung Volume: Normal Ranges Radiology, April 1, 2008; 247(1): 197 - 203. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Gorincour, D. Eurin, F. E. Avni, For the GRRIF Group (Groupe Radiopediatrique de Re, M. Cannie, J. C. Jani, F. De Keyzer, S. Dymarkowski, and J. A. Deprest Prenatal Prediction of Pulmonary Hypoplasia: US and MR Imaging Working Together Radiology, November 1, 2007; 245(2): 608 - 609. [Full Text] [PDF]
|
|
|
|
 |
|
 D. L. Conway Obstetric Management in Gestational Diabetes Diabetes Care, July 1, 2007; 30(Supplement_2): S175 - S179. [Full Text] [PDF]
|
|
|
|
|
|
M. Cannie, J. C. Jani, F. D. Keyzer, R. Devlieger, D. V. Schoubroeck, I. Witters, G. Marchal, S. Dymarkowski, and J. A. Deprest Fetal Body Volume: Use at MR Imaging to Quantify Relative Lung Volume in Fetuses Suspected of Having Pulmonary Hypoplasia Radiology, December 1, 2006; 241(3): 847 - 853. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
