| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Obstetric Imaging |
1 From the Department of Radiology, Beth Israel Deaconess Medical Center, 330 Brookline Ave, Boston MA 02215 (D.L., T.S.M.); Department of Radiology, Saint-Luc, Hospitalier de l’Universite de Montreal St-Denis, Montreal, Quebec, Canada (I.T.); and Department of Radiology, Lucille Salter Packard Children’s Hospital at Stanford, Palo Alto, Calif (P.D.B.). Received August 6, 2001; revision requested September 28; revision received October 15; accepted December 10. Supported by NIH grant NS37942. Address correspondence to D.L. (e-mail: dlevine@caregroup.harvard.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
MATERIALS AND METHODS: A retrospective review of MR and US images was performed for 110 normal fetuses and 94 fetuses with central nervous system abnormalities to assess lateral ventricular morphology as having (a) a normal appearance, (b) mild, disproportionate dilatation of the occipital horns with overall preservation of ventricular morphology, (c) colpocephaly with or without normal orientation of the frontal horns, (d) abnormal orientation of the frontal horns without colpocephaly, (e) an angular appearance, (f) fused frontal horns, (g) global dilation, or (h) a distorted appearance. Ventricular morphology on US and MR images was compared and correlated with reference standard diagnoses.
RESULTS: US and MR imaging classifications were concordant in 145 of 188 (77%) examinations. Mild disproportion of occipital horns with respect to frontal horns was seen only on MR images. This ventricular configuration was present in eight of 110 normal fetuses and in 10 of 16 fetuses with isolated mild ventriculomegaly (P < .001). An angular configuration of the lateral ventricles, which is seen in fetuses with neural tube defects (NTDs), was present on review of MR images in 11 fetuses and on US images in one fetus. The ventricles of fetuses with NTDs and angular ventricles (3–12 mm) were significantly smaller than those of fetuses with NTDs and global dilatation of the ventricles (13–25 mm; P < .05).
CONCLUSION: Ventricular contours differ with differing diagnoses of central nervous system abnormalities.
© RSNA, 2002
Index terms: Fetus, abnormalities, 856.862 • Fetus, central nervous system, 856.874
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
There have been multiple publications (4–11) concerning diagnosis of fetal CNS anomalies with the use of MR imaging. The purpose of our study was to elucidate further the MR imaging appearance of fetal cerebral ventricles by comparing US and MR images.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
One hundred seventy-four patients underwent MR imaging as part of a research protocol: 106 for CNS abnormalities seen at US, 19 for CNS anomalies suspected on the basis of patient history (eg, screening for tuberous sclerosis), 19 for placenta abnormalities such as placenta accreta or abruption, and 30 for abnormalities seen at US not involving the CNS (eg, thoracic and/or abdominal masses) but having the potential for better clarification on MR images. The remaining 13 examinations were ordered by clinicians for maternal uterine or adnexal abnormalities. The 174 research examinations were performed according to protocols approved by the Committee on Clinical Investigations at Beth Israel Deaconess Medical Center. Written informed consent was obtained from all patients prior to imaging. The committee allowed review of records for the other 13 patients without the need for written informed consent.
US Imaging
US and MR imaging were performed on the same day in all cases, with the exception of one patient who refused to undergo US. US images conformed to AIUM guidelines (1), and views of the fetal head included the biparietal diameter, head circumference, posterior fossa, and lateral ventricles. Acuson (models 128 or 128XP; Acuson, Mountain View, Calif) or ATL (HDI UM9 or 3000; ATL, Bothell, Wash) machines were used. Transabdominal transducers included 2.5–4-MHz sector transducers on Acuson machines, 2–4-MHz or 4–7-MHz curvilinear transducers on ATL machines, and 5–7-MHz sector transducers used transvaginally on both types of machines. In one case, images of the head were missing at the time of retrospective review. In 16 examinations (performed as part of transabdominal US to investigate possible placenta accreta or as part of pelvic US to identify an adnexal mass), images of the fetal head were not obtained, and one patient (whose fetus had agenesis of the corpus callosum) refused to undergo US. Thus, only 208 of the 226 fetal MR examinations yielded US images of the fetal head for comparison, 20 of which were repeat examinations.
When US was performed to determine CNS anomalies and when the fetus was in a cephalic position, images from both transabdominal and transvaginal evaluation of the fetal CNS were available for review. In the remainder of patients, only transabdominal US images of the fetal head were obtained.
MR Imaging
MR imaging was performed with a 1.5-T superconducting system (Vision or Symphony; Siemens, Erlangen, Germany) with a four-element phased-array surface coil. A half-Fourier rapid acquisition with relaxation enhancement (RARE) technique was used. Images were obtained in the transverse, coronal, and sagittal planes with respect to maternal anatomy when a placental or adnexal mass was indicated. Imaging parameters changed over the course of the study and were modified according to patient body habitus and indication for examination. A typical protocol for fetal imaging was a half-Fourier single-shot RARE technique with an echo spacing of 4.2 msec, echo time (TE) of 60 msec, echo train length of 72, section thickness of 3–5 mm, field of view of 26 x 35 cm, acquisition matrix of 128 x 256, and refocusing flip angle of 130°, for a total acquisition time of 19 seconds, with 420 msec per section. When a fetal anomaly was indicated, image planes were selected with respect to the fetus in the fetal sagittal, coronal, and transverse planes, with each sequence serving as the scout for subsequent imaging. Thus, for the indication of fetal CNS anomalies, images tended to be orthogonal to the fetal brain. However, in cases in which US was performed to investigate an extra-CNS anomaly or in cases in which fetal motion was consistently present, images of the CNS could be oblique to fetal CNS anatomy.
Reference Standard
Reference standard diagnoses for the 125 fetuses referred for evaluation of CNS anomalies were based on results of postpartum imaging (n = 45), surgical reports (n = 17), and autopsy (n = 13). When none of these results were available (n = 50), final diagnosis was based on physical examination findings in combination with imaging findings, or, if those were not available, on prenatal imaging findings alone. There were 23 diagnoses based on physical examination findings in combination with imaging findings (two neonates had encephaloceles, one had alobar holoprosencephaly, and 20 had a normal CNS at prenatal examination and at physical examination after birth). Nineteen of these 20 healthy infants have had normal development at follow-up to 3 years (mean follow-up, 1.2 years). In one child who had developmental delay and was diagnosed with carnitine deficiency, the developmental delay resolved after treatment. In 27 cases, the final diagnosis was based on prenatal imaging findings (MR imaging and US) alone and was determined by means of consensus of two of the authors (P.D.B., D.L.) at the time of examination. These 27 patients (a) had insufficient specimens for diagnosis after fetal termination, (b) declined autopsy, or (c) were lost to follow-up. Clinical follow-up was performed for all surviving patients.
Image Review
Together, three of the authors (D.L., T.S.M., I.T.) reviewed the fetal US and MR images. Only static images were used, and they were reviewed in random order. The three reviewers individually determined the ventricular pattern classification and voiced their decisions. When differences of opinion occurred, the final classification was decided by means of consensus. This occurred in five cases. When one of the reviewers recognized a particular fetus, that reviewer recused herself and did not state her opinion to avoid influencing the other two, unless a consensus opinion was needed. Original reports of findings were not used as part of the review process.
For categorization of ventriculomegaly, atrial dimensions were measured at US, and residual cortex was measured by one author (I.T.) at MR imaging. Ventriculomegaly was defined as an atrial diameter larger than 10 mm and was further categorized as mild (atrial diameter between 10 and 15 mm), moderate (atrial diameter larger than 15 mm; residual cortex larger than 2 mm in diameter) or severe (residual cortex smaller than 2 mm in diameter). These definitions of ventriculomegaly were established for this project and were modified from the criteria of Hudgins et al (12).
Ventricular configuration on US and MR images was categorized according to one of the following classes: (a) normal configuration (Fig 1); (b) primitive fetal ventricular configuration (Fig 2), defined as mild, disproportionate dilatation of the occipital horns when compared with the frontal horns, with overall preservation of ventricular morphology; (c) colpocephaly (Fig 3), defined as disproportionate dilatation of the occipital horns when compared with the frontal horns, with lack of preservation of ventricular morphology (this pattern could be seen with or without normal orientation of the frontal horns) (Fig 4); (d) abnormal orientation of the frontal horns without colpocephaly; (e) angular configuration (Fig 5), defined as a sharp angle at the frontal and occipital margins of the lateral ventricles; (f) fused frontal horns (Fig 6), defined as no or little septal tissue between the frontal horns, which have a continuous smooth appearance with an intact corpus callosum (this diagnosis was provided only when the finding was isolated to the frontal horns [ie, global dilatation was not present]); (g) global dilatation of the ventricular system (Fig 7); or (h) intracranial anatomy too distorted to characterize ventricular morphology (Fig 8).
|
|
|
|
|
|
|
|
|
|
The ventricular atrium and frontal horns were measured on all transverse MR images of the fetal head. Ventricular measurements, the ratio of atrial to frontal horn measurements, and the difference between atrial and frontal horn measurements as a percentage of atrial measurement were compared on MR images in fetuses with subjectively normal ventricular configuration, primitive fetal ventricular configuration, colpocephaly, and global dilatation.
Measurements of the ventricles obtained on MR images were compared with those obtained on US images. Since the ventricles were only routinely and prospectively measured on US images of fetuses with ventriculomegaly, this comparison was limited to fetuses with a ventricular size of 10 mm or larger on US images.
If a fetus underwent more than one examination during the study period, then images from all examinations were evaluated, but only findings from the earliest examination were used for comparison (for a total of 204 MR imaging and 188 US examinations).
Statistical Analysis
The 
2 test (Excel software; Microsoft, Redmond, Wash) was used to compare proportions of fetuses with primitive fetal ventricular configuration versus proportions of fetuses with normal ventricular size or mild ventriculomegaly. A paired t test was used to compare ventricular measurements obtained from US and MR images (Minitab software; Minitab, State College, Pa). The Student t test (Excel software, Microsoft) was used to compare (a) ventricular size in fetuses with neural tube defects (NTDs) and angular ventricles with that in fetuses with NTDs and global dilatation, (b) mean age of fetuses with normal ventricular size with that of fetuses with primitive fetal ventricular configuration, and (c) mean age of fetuses with NTDs and angular ventricles with that of fetuses with global dilatation. Analysis of variance (Minitab software, Minitab) was used to compare atrial measurements, frontal horn measurements, and the ratio of these measurements between groups of fetuses with normal ventricular configuration, global dilatation, primitive fetal ventricular configuration, and colpocephaly.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
|
|
|
An angular configuration of the lateral ventricles was present on review of MR images in 11 fetuses, all of which had NTDs. This configuration was seen on US images in only one fetus. The one fetus with an NTD and a normal ventricular configuration on US and MR images had no associated Chiari malformation. Three fetuses with spinal NTDs had global dilatation of the ventricles. The ventricular sizes of fetuses with NTDs and angular ventricles (3–12 mm) were significantly smaller than those of fetuses with NTDs and global dilatation of the ventricles (13–25 mm; P < .05). In fetuses with spinal NTDs, the mean gestational age of fetuses with angular ventricles (mean, 21.20 weeks ± 3.57 [SD]; range, 16–29 weeks) was significantly less than that of fetuses with global dilatation (mean, 28.50 weeks ± 4.33; range, 26–33 weeks; P < .05).
Ventricular Measurements
In cases of ventriculomegaly, measurements obtained on US and MR images were compared. There were no significant differences, with mean measurements of 17.1 mm ± 9.2 on US images and mean measurements of 17.3 mm ± 8.5 on MR images in 45 fetuses with ventriculomegaly on US images. A comparison of the absolute measurements with the two techniques was also performed. In 45 fetuses with ventriculomegaly that underwent 46 examinations, the measurements were the same in 17 (38%) fetuses, within 1 mm in 17 (38%) fetuses, within 2 mm in 10 (22%) fetuses, and within 4 mm in one (2%; atrial measurement of 45 mm at US) fetus.
Primitive fetal ventricular configuration was seen only with an atrial diameter of 6–15 mm (Table 4). The atrial diameter of fetuses with primitive fetal ventricular configuration (10.2 mm ± 2.0) was significantly greater than that of fetuses with normal ventricular configuration (6.5 mm ± 2.1); P < .05.
|
Repeat Examination
Of the 19 fetuses that underwent more than one examination, 15 had no change in ventricular appearance at MR imaging: 12 fetuses were normal, one fetus had fused frontal horns (this fetus had agenesis of the septi pellucidi), one fetus had primitive fetal ventricular configuration (the fetus was examined at 16 and 22 weeks), and one fetus had ventricles that were too distorted to characterize. Four fetuses had a changing ventricular appearance. One fetus with encephalocele had angular ventricles at examination at 20 weeks, but the ventricles appeared normal at 33 weeks. Three fetuses with primitive fetal ventricular configuration at 21–23 weeks had a normal ventricular appearance at 32–34 weeks.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Our distinction between normal ventricular configuration, primitive fetal ventricular configuration, and colpocephaly was a subjective one. This subjective impression was proven to be valid by finding statistically significant differences in ventricular measurements between the groups of fetuses with these classifications. One of the goals of this study was to determine if a subjective, mild disproportion of the ventricular system was abnormal. Since this finding was of unknown clinical importance, when there was subjectively mild disproportion of the occipital horns with overall preservation of ventricular contour we termed it primitive fetal ventricular configuration. This was a relatively common MR imaging finding in eight of 110 (7%) patients in our normal population. When the disproportion was marked, or when the frontal horns appeared abnormal, we then used the term colpocephaly. Our results demonstrate the utility of this distinction, since only when fetuses were determined to have colpocephaly or when the frontal horns were parallel was complete or partial agenesis of the corpus callosum present. However, it should be realized that this subjective impression was influenced by our 5 years of experience in performing fetal MR imaging. When we described this finding in the fetus in 1997 (5), we erroneously termed this colpocephaly in a fetus with mild ventriculomegaly. At the time, the patient was counseled as if her fetus had isolated, mild ventriculomegaly, since we believed that the finding of colpocephaly in the fetus had unknown clinical importance. The results of our current study help clarify further that mild disproportion of the occipital horns to the frontal horns is commonly seen in the fetus, especially in fetuses with mild ventriculomegaly.
Our findings also emphasize the importance of visualizing the entirety of the lateral ventricles with US. Unless an image of the frontal horns of the lateral ventricles is obtained, agenesis of the corpus callosum with nondilated ventricles will continue to be missed. This problem was noted by Bennett et al (2) in 1996, in a study in which 15 fetuses with normal intracranial findings at 16–22 weeks were found to have agenesis of the corpus callosum on images obtained in the third trimester. This also helps explain why primitive fetal ventricular configuration is not routinely noted at US. Static US images of the fetal head are obtained for biometric measurement at the level of the thalamus, measurement of the ventricular atrium, and visualization of the choroids plexus and posterior fossa. These images do not necessarily include the frontal horns; therefore, the finding of primitive fetal ventricular configuration on US images is not expected to be seen in a retrospective study of this nature.
One finding of unclear importance in the current study is the relationship between primitive fetal ventricular configuration and mild ventriculomegaly. Follow-up of patients is ongoing to allow comparison of outcomes of fetuses with mild ventriculomegaly associated with a normal configuration, primitive fetal ventricular configuration, or global dilatation of the ventricles.
The association between primitive fetal ventricular configuration and gestational age is also unclear in the current study. There were three fetuses with two examinations each in which the ventricles had primitive fetal ventricular configuration at 20–22 weeks and then appeared normal at repeat examination at 30–32 weeks. However, although there was a trend for primitive fetal ventricular configuration to be seen at less than 24 weeks, there was no statistically significant difference in gestational age when comparing the entire group of fetuses with primitive fetal ventricular configuration and those with normal ventricular configuration. The fact that primitive fetal ventricular configuration was seen in a relatively large number of our normal fetuses (7%) and that it tends to have resolved at follow-up suggests that primitive fetal ventricular configuration can be a normal finding, especially for fetuses in the early second trimester. The finding of primitive fetal ventricular configuration should nevertheless trigger a search for associated abnormalities, with particular attention paid to the frontal horns. On the basis of our findings, we believe that the term colpocephaly, when applied to the fetus, should be reserved for cases in which the disproportion between occipital and frontal horns is subjectively marked and/or the frontal horns have abnormal orientation. While not all fetuses with colpocephaly will have dysgenesis of the corpus callosum, this finding necessitates a careful evaluation of the entire corpus callosum.
Angular Ventricles
A pointed appearance of the frontal horns in neonates with NTDs has been described in studies in which computed tomography (18) and US (19) were used to image the fetal head. While the prenatal intracranial findings of ventriculomegaly and the Chiari malformation associated with NTDs have been well documented in the literature, the importance of prenatally visualized angular configuration of the frontal horns has not, to our knowledge, been previously assessed.
In this study, an angular configuration was seen only in association with NTDs. This finding can therefore serve as a valuable secondary sign of the presence of NTDs. These defects can be small and difficult to detect on MR images (6). This kind of information is important, since not all MR images are evaluated in conjunction with US images, and will allow MR imaging to be specifically tailored to finding NTDs if such a ventricular configuration is noted. Therefore, an angular appearance of the ventricles on US or MR images, as well as findings of the Chiari malformation, should trigger a close evaluation of spinal NTDs and encephaloceles.
Study Limitations
One limitation of our study is that it was performed retrospectively. The MR images of the ventricles were not standardized in the normal population because of the manner in which the images were obtained. For cases with abnormal findings, MR images were standardized to the extent possible with a moving fetus. US images were standardized for views of the lateral ventricles, although images of the frontal horns were not obtained in all cases. Since much of the information obtained at US comes from real-time examination, it must be emphasized that this study was not performed to assess the sensitivity of US in the detection of CNS abnormalities. Only the appearance of the ventricles was assessed. For example, all spinal NTDs in this study were visualized with US, but only one of these fetuses had a documented angular appearance of the frontal horns at US. Since the diagnosis of a NTD can be made at US without the demonstration of the abnormal appearance of the ventricles, this finding is unlikely to be helpful in US examinations. On MR images, however, spinal NTDs can be subtle and difficult to detect. In these cases, the secondary sign of angular ventricles can be a valuable clue to aid in diagnosis. Although the retrospective nature of this study limits a true comparison of MR imaging and US in the assessment of ventricular morphology, it does not negate our study, since our results emphasize the importance of evaluating the entirety of the lateral ventricles with US. Since performing this study, we have changed our US protocol to include views of the frontal horns in an effort to improve our screening abilities.
One important finding in this study was that ventricular configuration on MR images occasionally varies, depending on the level and angulation of the imaging plane (Fig 9). This is due to lack of standardized angulation of fetal CNS imaging planes and is a caveat in MR imaging assessment of ventricular morphology. Errors in subjective impression could have occurred if the entire ventricles were not assessed, an error we hope to have avoided by having three reviewers.
Measurement error could have occurred on MR images as a result of either selection of appropriate imaging plane or placement of calipers; however, lack of difference between ventricular measurements obtained on US and MR images is reassuring.
Another possible bias could have been introduced into our categorization of ventricular morphology if one author failed to recuse herself for recognizing a fetus.
A further limitation of our study is selection bias, since most of our patients were referred for imaging of a CNS anomaly. This limits the applicability of our results to a healthy, low-risk population. It should be noted, however, that all fetuses defined as having a normal CNS had either normal ventricular morphology or primitive fetal ventricular configuration.
A final limitation of our study is that our reference standard was imaging alone for some patients. Lack of a true reference standard for final diagnosis is a general problem in prenatal imaging research because tissue diagnosis is rarely available (20).
In conclusion, it must be emphasized that US is the screening modality of choice in the assessment of the pregnant patient and fetus. However, there are cases in which additional information is needed and MR imaging can be helpful. Multiple articles (4–11) have demonstrated the increased amounts of information available with MR imaging in the evaluation of the CNS. The purpose of the current study was not to compare the effect of MR imaging on prenatal diagnosis but to clarify further the normal and abnormal appearances of the fetal cerebral ventricles on MR images. The results of this study show that the fused appearance of the frontal horns is associated with agenesis of the septi pellucidi; that even with normal orientation of the frontal horns, colpocephaly is associated with dysgenesis of the corpus callosum; and that angular ventricles are associated with NTDs. Ventricular contours differ with differing diagnoses of CNS abnormalities. Although not proven in this study, we believe it is likely that routine use of frontal horn views at US and recognition of the configuration of the lateral ventricles at fetal MR imaging will aid in the diagnosis of a number of CNS abnormalities.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, D.L.; study concepts, D.L., P.D.B.; study design, D.L.; literature research, I.T., D.L.; clinical and experimental studies, D.L., T.S.M.; data acquisition, D.L., I.T., T.S.M.; data analysis/interpretation, D.L., I.T.; statistical analysis, D.L., I.T.; manuscript preparation, D.L., I.T.; manuscript definition of intellectual content, D.L.; manuscript editing, D.L., P.D.B.; manuscript revision/review, all authors; manuscript final version approval, D.L.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. R. Wax, M. G. Pinette, A. Cartin, J. Michaud, and J. Blackstone Fetal Cerebral Ventricular Pointing as a Marker of Spina Bifida: Incidence and Observational Agreement J. Ultrasound Med., March 1, 2009; 28(3): 317 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Callen and R. A. Filly Supratentorial Abnormalities in the Chiari II Malformation, I: The Ventricular "Point" J. Ultrasound Med., January 1, 2008; 27(1): 33 - 38. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. R. Benacerraf, T. D. Shipp, B. Bromley, and D. Levine What Does Magnetic Resonance Imaging Add to the Prenatal Sonographic Diagnosis of Ventriculomegaly? J. Ultrasound Med., November 1, 2007; 26(11): 1513 - 1522. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O.A. Glenn and A.J. Barkovich Magnetic resonance imaging of the fetal brain and spine: an increasingly important tool in prenatal diagnosis, part 1. AJNR Am. J. Neuroradiol., September 1, 2006; 27(8): 1604 - 1611. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. V. Valsky, L. Ben-Sira, S. Porat, N. Yanai, A. Lewin, M. Nadjari, J. M. Gomori, and S. Yagel The Role of Magnetic Resonance Imaging in the Evaluation of Isolated Mild Ventriculomegaly J. Ultrasound Med., April 1, 2004; 23(4): 519 - 523. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Benveniste, J. S. Fowler, W. D. Rooney, D. H. Moller, W. W. Backus, D. A. Warner, P. Carter, P. King, B. Scharf, D. A. Alexoff, et al. Maternal-Fetal In Vivo Imaging: A Combined PET and MRI Study J. Nucl. Med., September 1, 2003; 44(9): 1522 - 1530. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
