Ultrasound (US) is the screening method of choice for prenatal imaging; however, there are cases in which US is equivocal. In those cases, ultrafast prenatal magnetic resonance imaging (MRI) is an important complementary means of investigation.
Dr. Berggruen
received her medical degree from Loyola Stritch School of Medicine,
Maywood, IL, in 1998. She is a fourth-year resident at the
University of Arizona Health Sciences Center, Tucson, AZ. Next year
she will begin an MRI Body fellowship at Northwestern Memorial
Hospital, Chicago, IL.
Ultrasound (US) is the screening method of choice for
prenatal imaging; however, there are cases in which US is
equivocal. In those cases, ultrafast prenatal magnetic resonance
imaging (MRI) is an important complementary means of
investigation. Prenatal MRI is helpful at improving the anatomic
definition, clarifying the diagnosis, and identifying associated
abnormalities. This information facilitates family counseling for
prenatal and perinatal management.
Ultrasound (US) continues to be the screening method of choice
for prenatal imaging because of its widespread availability,
real-time capability, low cost, and safety. The quality of the
examination is dependent on the equipment and skill of the
examiner. In most cases, the findings on US are decisive. There are
cases, however, in which the findings are equivocal or diagnostic
information is insufficient, including pregnancies complicated by
maternal obesity, oligohydramnios, or an unusual fetal lie. In such
cases, the availability of another imaging technique, such as
magnetic resonance imaging (MRI), may prove beneficial.
Until recently, T2-weighted MRI acquisition times of several
minutes made it necessary to sedate the fetus directly through the
umbilical cord or sedate the mother. The development of ultrafast
T2-weighted MRI makes it possible to obtain images of good quality,
enhancing fetal anatomic evaluation without sedation. Prenatal MRI
is helpful in improving the anatomic definition, clarifying the
diagnosis, and identifying other associated abnormalities,
particularly before fetal surgical intervention.
Techniques
MRI gradient systems with higher slew rates and gradient
amplitude have allowed newer pulse sequences, including single-shot
half-Fourier T2-weighted images and echo planar imaging.
Single-shot fast-spin echo (or HASTE [half-Fourier acquisition
single-shot turbo-spin echo]) images were the first type of
sequence to obtain high quality T2-weighted spin-echo images in a
single breath-holding period. Single-shot fast-spin echo sequences
achieve relatively motion-free breath-hold images in the pelvis or
abdomen with T2-weighted tissue contrast similar to conventional
fast-spin echo images. Obtaining planes that represent anatomic
cross sections relative to the fetus facilitates examination of the
fetus with HASTE imaging.
1
Gadolinium is not used in fetal MRI as it crosses the placenta and
is not cleared from fetal circulation. The effect of gadolinium on
the fetus is not known.
2
Fetal brain
In a study by Laveaucoupet et al,
3
the usefulness of fetal MRI in ischemic brain injury was
demonstrated in seven cases of fetal brain ischemia that was
suspected on US and confirmed by fetal MRI. The abnormalities
suspected on US included ventricular dilatation (n = 3),
microcephaly (n = 1), twin pregnancy with in utero death of a twin
and suspected cerebral lesion in surviving twin (n = 3). MRI
diagnosed hydranencephaly (n = 1), multicystic encephalomalacia (n
= 2) (figure 1), unilateral capsular ischemia (n = 1), porencephaly
(n = 2) (figure 2), and corpus callosum and cerebral atrophy (n =
1). US and MR images were compared with pathologic findings or
postnatal imaging. In comparison to US, visualization of fetal
brain anomalies was superior with MRI, which provided a more
precise diagnosis. These seven cases demonstrated that MRI could be
an important complementary means of investigation when pathology is
suspected during prenatal US.
Many pathologic conditions throughout gestation may compromise
materno-fetal circulation and lead to hemorrhagic or ischemic
injuries in the fetal brain. In the second trimester, the fetal
brain responds to trauma by liquefaction of destroyed parenchyma
resulting in cavities: multicystic encephalomalacia, porencephaly,
and hydranencephaly are terms that describe patterns of cavitation.
The more mature brain reacts to an insult by astrocytic
proliferation or gliosis.
3
MRI allows visualization of the entire fetal brain, which
sometimes is not possible by US, especially in fetuses with
ventricular dilatation. MRI provides better soft-tissue contrast
than US, and therefore evaluates extension of lesions into the
cerebral parenchyma. However, MRI may not detect subtle ischemic
lesions such as gliosis, microcalcifications, and microthrombi
associated with isolated ventricular dilatation.
3
In a study by Wagenvoort et al,
4
MR images were reviewed to determine if MRI could give additional
information in prenatal diagnosis of congenital anomalies in cases
in which US analysis was not conclusive. In this study, in which MR
images were reviewed with US knowledge, MRI supplied additional
information in 10 of 19 cases referred due to equivocal findings of
brain anomalies on US. MRI was provided a confirmative diagnosis
for six patients (31.6%). One exam was false negative; MRI did not
show periventricular leukomalacia (PVL) or hemorrhage while vaginal
US did suggest PVL. In 5 of 10 cases referred for ventriculomegaly,
MRI discovered other anomalies (figure 3).
Sonography of posterior fossa anatomy may be difficult because
of fetal skull shadowing or engaging of the fetal head in the third
trimester of pregnancy. In three patients referred for cerebellar
anomalies, a correct diagnosis was made with fetal MRI.
4
In a study by Quinn et al,
5
MRI excluded a suspected Dandy-Walker malformation in a fetus with
agenesis of the corpus callosum, a difficult diagnosis to make
accurately by US. In contrast to US, MRI allowed for the correct
identification of posterior fossa anatomy in third-trimester
fetuses.
6
Evaluation of the posterior fossa on US often depends on a single
angled axial view through the cerebellar hemispheres and region of
the cisterna magna. The cisterna magna may appear artificially
enlarged or the shape of the hemispheres may appear abnormal. MRI
markedly improves visualization of the posterior fossa because
direct coronal, sagittal, and axial images can be obtained, thereby
confirming anatomy of the vermis and fourth ventricle. MRI usually
can identify the typical findings of a Chiari malformation.
2
Holoprosencephaly is a malformation of the prosencephalon with
failure of normal midline cleavage and incomplete midfacial
development. MRI is most helpful in trying to distinguish a mild or
lobar form of holoprosencephaly from other forms of
ventriculomegaly and hydranencephaly, which is a severe failure of
brain formation, secondary to ischemic event.
7,8
In a study evaluating central nervous system (CNS) anomalies by
Ohgiya et al,
6
US findings were suggestive of, but not definitive, for
anencephaly. MRI was helpful in defining a large amount of
angiomatous stroma and absence of normal calvaria. The information
about the CNS deformity was helpful to the parents considering a
pregnancy termination.
Normal gyral patterns, ventricular size, and patterns of
neuronal migration have all been documented by MRI.
9
Sulcal development is a marker of cortical maturation and fetal
maturity. In a MRI study of 93 fetuses, Levine and Barnes
10
found that sulcation appeared in the order predicted by anatomic
studies, but with a lag of up to 8 weeks (mean 1.9 weeks). The
first to form is the interhemispheric fissure, present in all
normal fetuses studied at 14 weeks. The Sylvian fissure becomes
grooved by 16 weeks. Beginning at approximately 29 weeks in utero,
myelination progresses cephalad from the spinal cord and is not
complete until about the second year of life.
11,12
Myelination accounts for signal intensity differences between the
different regions of the brain. In a study by Lan et al,
13
evaluation of 25 normal fetuses showed that only two layers could
be discerned after 29 weeks: an inner hyperintense layer (the white
matter) and an outer hypointense layer (the gray matter). The
corpus callosum is fully formed by 20 weeks. The basal ganglia and
thalami are best seen after 26 weeks.
In a review of 128 fetuses referred for non-CNS indications, no
fetus was found with an atrial diameter exceeding 10 mm on MRI.
14
The 10-mm rule, described on a sonographic axial view of the fetal
atrium, is therefore the measurement used as the upper limit of
normal on MRI. The cavum septum pellucidum is seen in all normal
fetuses.
14
Previously, abnormalities of neuronal migration were believed to
be rare; however, with MRI, they are seen in up to 20% of cases of
CNS anomalies. In one report, MRI was able to visualize areas of
neuronal heterotopias in 54% of fetuses with a definite diagnosis
of migrational disorder. MRI demonstrated 80% of lissencephaly, 73%
of polymicrogyria, and 100% of schizencephaly prenatally.
7
In a study by Enomoto et al,
15
121 fetuses were imaged to determine the predictive value of
prenatal MRI for evaluation of fetal brain anomaly and other
pathology. The degree of cortical convolution and ventricular size
were used as landmarks to assess the fetal brains of different
gestational age, together with the degree of myelination and
presence of parenchymal signal abnormality in the group of late
gestation. The subjects were divided into 99 with normal findings
and 22 with abnormalities. In normal subjects, the ventricles
ceased to dilate after 25 weeks. The convolutional development
increased after 30 weeks and a primitive form of the cerebral
surface was detected at 32 weeks. The myelination started in the
brain stem and internal capsules between 22 and 25 weeks. The
abnormalities found included anencephaly, holoprosencephaly, Chiari
malformation, ventriculomegaly, cephalocele, Dandy-Walker syndrome,
and hypogenesis of the corpus callosum. Five autopsy cases and 17
postnatal imaging cases confirmed the abnormalities. Focal
encephalomalacia, abnormal configuration of the ventricles,
heterotopia, and callosal and posterior fossa anomalies were
delineated clearly on MRI.
Neck masses
In a study of 3 cases, fetal MRI was instrumental in evaluating
the anatomic extent of giant neck masses with potential airway
obstruction.
5
In contrast to US, anatomic details about location and impingement
of the fetal airway and neck vessels by lesions that rivaled the
fetus in volume were better seen with MRI. Cervical teratoma could
be differentiated from cystic hygroma on MRI. In each case, fetal
MRI helped in the prenatal assessment of the mass before cesarean
delivery using an ex-utero intrapartum therapy (EXIT) procedure to
establish an airway. Compared with US exams in which only a small
part of the lesion may be present in the acoustic window, the whole
lesion and its anatomic relations could be reviewed on MRI.
Lungs
Lungs show moderately high signal intensity on T2-weighted
images.
4
Congenital diaphragmatic hernia (CDH) is the most common fetal
chest mass, usually occurring on the left side. US differentiation
of bowel loops from other chest masses is sometimes difficult. The
US determination of liver herniation is not always possible because
it relies on indirect signs.
6
Evaluating liver position on US may be difficult, as the echo
texture of the liver and lung is similar. Estimations of the
residual lung area are not always possible by US. With MRI, it can
be easy to depict the herniated bowel and the position of the liver
and to distinguish meconium-filled bowel from normal lung because
of the difference in signal intensity. On T1-weighted images,
meconium-filled bowel is very high in signal intensity, making the
position of bowel above or below the diaphragm easy to evaluate.
2
In a study by Quinn et al,
5
fetal surgery candidates undergoing evaluation for a congenital
anomaly that was potentially a correctable lesion were selected for
MRI. The diagnoses included: CDH (n = 14) and lung masses (n = 4).
The results of the US exam were known at the time of the MRI study
on the same day. All diagnoses were confirmed after delivery by
radiographic imaging, operation, or autopsy. With CDH, MRI
demonstrated liver herniation into the chest in 11 of 14 cases. The
information included the amount and location of bowel, stomach, and
liver herniated into the fetal thorax. The degree of liver
herniation was substantial in six cases, in which a third or more
of the fetal liver was intrathoracic. MRI was able to identify
bilateral CDH in one fetus. In four cases, US findings had not been
definitive. In two cases of CDH detected by MRI, the primary
diagnosis by US was congenital cystic adematoid malformation (CCAM)
based on the presence of nonperistaltic bowel in the chest that may
appear as a multicystic lesion similar to a CCAM on US.
Despite advances in sonography and color Doppler imaging, the
prenatal diagnosis of anomalies such as CDH is difficult and has
been missed in 41% of cases in one recent retrospective study.
16
The mortality in CDH relates to the degree of pulmonary hypoplasia
and associated pulmonary hypertension and has been estimated to be
as high as 58%. Differentiating between chest masses, such as CCAM
and bronchopulmonary sequestration (BPS), by US can be difficult.
Fetal MRI provided more anatomic detail than US regarding extent
and location of the chest masses that included BPS (n = 2), and
CCAM (n = 2). In many cases, the lobe of the lung involved with the
lesion could be identified.
5
Pulmonary hypoplasia is a condition that causes high mortality
and morbidity in neonates. The prenatal diagnosis of pulmonary
hypoplasia assists with the postnatal care. In a study by Kuwashima
et al,
17
the diagnostic capabilities of fetal MRI in pulmonary hypoplasia
were evaluated. Subjects included 23 fetuses. A diagnosis of
pulmonary hypoplasia was made on the basis of perinatal physical,
surgical, and/or autopsy findings. With regard to pulmonary
hypoplasia, the subjects were divided into three groups--1a:
infants with pulmonary hypoplasia and perinatal death (n = 8),
including 5 who underwent autopsy; 1b: infants with pulmonary
hypoplasia and longer survival over the neonatal period (n = 2);
and 2: infants without pulmonary hypoplasia (n = 13). On MRI, all
fetuses in groups 1a and 1b showed low signal intensity of the
lung, obscured pulmonary vessels, and a small thorax. All fetuses
in group 2, except for one fetus at 24 weeks gestation, showed
homogeneously high signal intensity in the lung and well-defined
pulmonary vessels.
In another study, Shinmoto et al
18
investigated normal fetal lung growth and correlated the estimated
lung volume with clinical outcome in 90 fetuses (58 with normal
lungs and 32 with suspected pulmonary hypoplasia) who underwent MRI
including volumetric lung measurement. The gestational age ranged
from 21 to 38 weeks. The lung volume was calculated by multiplying
total lung area and slice thickness. The measured fetal lung
volumes were correlated with gestational age and other parameters
of fetal size such as crown-rump length (CRL). The calculated
percentages of lung volume were well correlated with neonatal
respiratory condition and Apgar scores. The volumetric measurement
of fetal lung was valuable in that the relative fetal lung volume
was correlated with neonatal respiratory conditions.
Abdomen/pelvis
Fetal renal anomalies often are associated with oligohydramnios,
a condition that makes fetal US difficult. MRI is not greatly
affected by decreased amniotic fluid.
2
A study by Caire et al
19
was designed to evaluate the usefulness of MRI in suspected
congenital fetal genitourinary anomalies.
16
Assessments were made of amniotic fluid; location, size and
presence of kidneys; bladder appearance; and evaluation of the
perineum. MRI findings were compared with US and neonatal
diagnosis. MRI changed the diagnosis in 4 of 15 cases.
Oligohydramnios or anhydramnios was present in 8 of 15 cases; this
did not hinder MRI visualization of genitourinary anomalies.
Visualization of bladder and kidneys was possible after 18 weeks.
MRI also provided better anatomic detail of the perineum.
19
Multicystic dysplastic kidney (MCDK) may be confused with
hydronephrosis on US. MRI, confirming the absence of normal renal
parenchyma and renal pelvis, can diagnose MCDK.
14
Although primary hepatic tumors are uncommon, hamartomas,
hemangoendotheliomas, and hepatoblastomas do occur in the fetus.
MRI helps differentiate these masses based on morphology. In
addition, MRI can be used to differentiate abdominal masses and
other cystic structures, such as mesenteric cysts, lymphangiomas,
and ovarian cysts, from dilated loops of bowel.
2
With huge abdominal masses, such as mesenteric cyst and
lymphangioma, MRI clarified the diagnosis in 14 of 31 (45%)
patients who underwent fetal treatment after US and MRI.
5
Twin pregnancy
When assessing twin anomalies by US, the problems of multiple
fetuses in the field, fetal lie, and aberrant anatomy can create
diagnostic confusion. MRI allowed anatomic refinement of the US
diagnosis of twin-twin transfusion syndrome. The placental anatomy
and cord insertion sites were delineated for preoperative planning
for cord legation or placental vessel laser ablation.
5
Safety
To date, no known harmful effects to the developing fetus have
been documented using clinical MR scanners. One long-term study
evaluated 20 children who had undergone fetal MRI after 21 weeks
gestation. After 3 years, there was no demonstrable increase in
occurrence of disease or disabilities.
20
In 1991, the safety committee of the Society for Magnetic Resonance
Imaging issued guidelines for patient safety.
21
The guidelines recommended that patients be informed that there has
been no indication that the use of clinical MRI during pregnancy
has produced deleterious effects; however, safety has not been
proven to date.
2
Conclusion
As further improvements in MRI technology yield faster scan
times and higher resolution, the applications for fetal imaging
will increase. Presently, MRI is an adjunct to good prenatal US. It
can provide significant additional information that can affect
prenatal counseling, prenatal intervention, and delivery
planning.