Pediatr Radiol (2007) 37:133–140
DOI 10.1007/s00247-006-0353-5
ORIGINAL ARTICLE
Water imaging (hydrography) in the fetus: the value
of a heavily T2-weighted sequence
Beth M. Kline-Fath & Maria A. Calvo-Garcia &
Sara M. O’Hara & Judy M. Racadio
Received: 23 June 2006 / Revised: 26 September 2006 / Accepted: 10 October 2006 / Published online: 29 November 2006
# Springer-Verlag 2006
Abstract
Background Since the development of fast imaging se-
quences, MR has proved to be a helpful tool in the evaluation
of fetal pathology. Because of the high water content of fetal
tissues and pathology, hydrography imaging (MR fetography)
can provide additional diagnostic information.
Objective To demonstrate the benefit of MR fetography in
fetal imaging.
Materials and methods From 2004 to 2005, 126 fetal MR
examinations were performed for evaluation of an abnor-
mality depicted on an antenatal sonogram. Single-shot fast
spin-echo MR imaging and MR fetography were performed
through the area of fetal pathology. The two studies were
retrospectively compared.
Results The primary diagnosis was not changed with the
addition of MR fetography. New findings, particularly in
the kidneys and spine, were identified in 9% of the patients.
When fetal pathology was of high water content (80%
patients), the MR fetography imaging increased diagnostic
confidence. In 11% of the patients, those with cardiovas-
cular or low water pathology, the MR fetography was not
beneficial.
Conclusion The mainstay of fetal imaging is currently the
HASTE and SSFSE sequences. However, MR fetography is
an excellent adjunct that highlights fetal pathology by
reinforcing the diagnosis, identifying additional findings,
and providing high-contrast high-resolution images that are
helpful when counseling clinicians and patients.
B. M. Kline-Fath (*) : M. A. Calvo-Garcia : S. M. O’Hara :
J. M. Racadio
Department of Radiology,
Cincinnati Children’s Hospital Medical Center,
University of Cincinnati Medical Center,
Cincinnati, OH, USA
e-mail: beth.kline-fath@cchmc.org
Keywords Fetal imaging . MR hydrography
Introduction
Fetal MR was first described in 1983. At that time,
conventional imaging sequences were utilized to evaluate
maternal and placental pathology beyond the capabilities of
sonography but without the use of ionizing radiation. With
this technique, fetal imaging was limited to secondary fetal
and amniotic fluid motion, despite sedation of the mother
and fetus [1]. Fast gradient echo and echoplanar imaging
sequences were attempted less than a decade later, but
proved poor because of low signal-to-noise and suscepti-
bility artifacts [2]. In the 1990s, MR technology improved
with faster and stronger gradients and higher readout
bandwidth. With these advances, fast imaging sequences
were developed. These new sequences provided high-
quality T2-weighted images that could be obtained indi-
vidually in less than a second, thus freezing fetal motion.
On Siemens scanners, this sequence is known as half-
Fourier acquired single shot turbo spin echo (HASTE); on
General Electric scanners, it is single-shot fast spin-echo
(SSFSE); on Philips scanners, it is ultrafast spin echo
(UFSE); and on Toshiba scanners, it is fast advanced spin
echo (FASE).
In the presence of many sonographic fetal abnormalities,
MR imaging of the pregnancy is becoming common
practice. Because MR can provide imaging with a large
field-of-view, multiplanar construction and high soft-tissue
contrast while avoiding image deterioration caused by
maternal habitus or presence of oligohydramnios, this
modality is extremely helpful. Fetal MR has repeatedly
been shown to define equivocal findings, boost diagnostic
confidence and elucidate additional pathology not visual-
ized on sonography [1, 3–5]. Fetal MR is a valuable
134 Pediatr Radiol (2007) 37:133–140

diagnostic tool, guiding counseling, fetal therapy and Table 1 Imaging parameters for SSFSE and MR fetography
postnatal care [4, 6].
Parameter Technique
The fast T2-weighted imaging sequences are the main-
stay of most fetal examinations. T1-weighted, generally SSFSE MR fetography
gradient echo imaging is performed to evaluate brain
TR (ms) 4,000 16,000
myelination, hemorrhage, and meconium in bowel or liver
TE (ms) 88 180
position, but is otherwise of limited value because of the FOV (cm) 25–32 30–34
low contrast in the fetus. T2-weighted imaging is the most Slice (mm) 3–5 5/10
diagnostic because the majority of fetal structures are of Matrix 192/224 256/512
high water content. The bright signal of water on the T2-
weighted images provides excellent contrast and allows
better separation of adjacent organs. Heavily T2-weighted cava compression and secondary vasovagal reaction. In
hydrography sequences have been utilized successfully to addition, if the mother was claustrophobic, the supine
delineate fluid-filled structures, particularly in the urinary position accentuated this symptom. A phased array torso
system and biliary tract [7, 8]. Because normal and coil was placed concentrically around the patient centered
pathologic fetal tissues tend to be fluid-filled or surrounded at the level of the gravid uterus.
by fluid, our thought was that hydrography imaging (MR Imaging of the entire fetus was obtained but directed to
fetography) would be helpful to accentuate and further the area of suspected pathology. Sequence T2 parameters
define fetal pathology. are shown in Table 1. Both T2-weighted sequences were
performed through the area of pathology with imaging
anatomically in the axial, sagittal and coronal planes of the
Materials and methods fetus.

From July 2003 to January 2005, 126 pregnant women

were referred to our institution for MR imaging because of
a fetal sonographic abnormality. The gestational age ranged
from 18 to 39 weeks. Of these imaging studies, 91 were
singleton gestations, 34 were twin gestations, and 1 was a
triplet gestation. MR studies were performed on a 1.5-T
scanner (Horizon, General Electric Medical Systems,
Milwaukee, Wis.) using SSFSE T2- and heavily T2-
weighted high-resolution MR fetography sequences. Al-
though these were the primary imaging sequences, one or
two planes of imaging with a gradient echo T1-weighted
sequence with imaging parameters of TR 160 ms, TE min,
flip angle 70°, FOV 30–35 cm and slice thickness 5 mm
was performed. Informed consent was obtained prior to the
procedure because in our institution, we believe it is
important to acknowledge that, although there are no
known risks to the fetus imaged after the first trimester,
the clinical and research data are limited. No maternal
sedation was utilized and no gadolinium was administered,
as it is not FDA-approved for this procedure. Mothers had
no solids by mouth for 3 h and no liquids for 2 h prior to
imaging, except for the administration of water. In our
experience, there is less fetal motion when the maternal/
fetal glucose level is diminished. In addition, there is less
maternal bowel artifact when the mother has nothing by
mouth. Imaging was performed with the mother in a supine
or decubitus position. Most of our mothers preferred the Fig. 1 Sagittal SSFSE T2 image to the left of the midline through a
male fetus at 22 weeks of gestation with a diagnosis of posterior
decubitus position for reasons of comfort. We also
urethral valves and decreased amniotic fluid. The left kidney shows
encouraged the decubitus position as we have found that mild distention of the central collecting system and a small perinephric
the supine position predisposes to maternal inferior vena urinoma (arrow)
Pediatr Radiol (2007) 37:133–140 135

Fig. 3 Axial SSFSE image at the level of spinal dysraphism of a fetus

Fig. 2 Sagittal MR fetography image of the left kidney in the same
fetus as in Fig. 1. The high-contrast technique clearly demonstrates a
cyst arising from the superior pole of the kidney (arrow) that is not
delineated on the SSFSE image
The studies were retrospectively reviewed. The results
were separated into groups according to whether the
addition of MR fetography (1) changed the primary
diagnosis, (2) provided additional findings, (3) increased
confidence in the diagnosis, or (4) was not beneficial. This
study qualified for expedited review IRB approval.
at 23 weeks of gestation with a Chiari II malformation and a mid-
lumbar myelomeningocele shows a neural tube defect and the spinal
cord centrally (arrow)
thorax (two) and brain (one). The new renal findings
identified were renal cysts and areas of cystic dysplasia
(Figs. 1 and 2). In the spine, diastematomyelia was
discovered in three fetuses at or above a neural tube defect
(Figs. 3 and 4). In two lung lesions, cysts were detected

Results

In 100% of the examinations reviewed, the MR fetography

imaging did not change the primary diagnosis rendered by
the SSFSE sequence. Fetal MR diagnosis was confirmed in
56 by postnatal imaging, 20 at autopsy and 31 (mostly
cases of twin–twin transfusion syndrome and twin-reversed
arterial transfusion syndrome) by intrauterine therapy.
Nineteen of the pregnancies were lost to follow-up or did
not have autopsy after intrauterine or postnatal demise.
In 11 examinations, or 9% of the studies, additional
findings were noted on the MR fetography sequence that
Fig. 4 MR fetography image of the same fetus as in Fig. 3 delineates
were not evident on the SSFSE sequence. The findings two separate spinal cords (arrows), consistent with diastematomyelia,
were in investigations of the kidneys (five), spine (three), at the site of the neural tube defect
136
Fig. 5 Sagittal SSFSE image of a fetus at 30 weeks of gestation with
a lung lesion and hydrops. The lung lesion (long arrow) shows a
homogeneously high signal when compared to normal compressed
lung (short arrow) (small thin arrow intraabdominal ascites)
within the mass solely with MR fetography (Figs. 5 and 6).
This helped confirm the presence of two congenital cystic
adenomatoid malformations. Finally, in the brain, MR
fetography sequence defined an area of nodularity in the
cortex of a frontal lobe, consistent with an area of cortical
malformation.
In 101 examinations (80%), the addition of the MR
fetography sequence subjectively increased the radiologist’s
confidence in the findings. As expected, it improved
visualization of structures with fluid-filled boundaries (14
bowel, 5 renal collecting system, 3 biliary and 2 airway;
Figs. 7 and 8). The MR fetography imaging defined the
cyst border and architecture in 10 thoracic and 5 abdominal
lesions (Figs. 9 and 10). This imaging also sharpened the
detail of structures surrounded by fluid (52 cerebral, 7
ventral wall defects and 5 spine lesions; Figs. 11 and 12).
In 14 examinations (11%), MR fetography was not
beneficial. Seven of these studies were examinations
Pediatr Radiol (2007) 37:133–140
Fig. 6 Sagittal MR fetography image of the same fetus as in Fig. 5
demonstrates a homogeneous high-signal lung lesion (thin arrow) but
also depicts a cyst (thick arrow) along the superior aspect of the
lesion. The findings are consistent with a congenital cystic adenoma-
toid malformation
performed for cardiac or vascular abnormalities. In three
fetuses with a situs abnormality, there was no benefit. The
MR fetography images were poor in two fetuses with renal
agenesis and in two normal anatomy fetuses with premature
rupture of membranes.
Discussion
Water is the largest constituent of the human body. Total
body water content varies between individuals but is known
to decrease with age. The human fetus is made up of
approximately 90% water, while adult water content
normally ranges from 50% to 70% [9]. Prenatal sonography
utilizes the high water content of the fetus and amniotic
fluid environment to monitor normal and abnormal fetal
development. Fetal MR, which provides superior soft-tissue
contrast to sonography, is primarily obtained utilizing a T2-
Pediatr Radiol (2007) 37:133–140
Fig. 7 Sagittal SSFSE image of a fetus at 29 weeks of gestation with
a large omphalocele containing liver (arrow)
weighted sequence, as this imaging highlights differences
between soft tissue and high water content structures and
the amniotic fluid environment. In addition, because most
fetal pathology is central nervous system or renal in origin,
T2-weighted imaging is the modality of choice.
The workhorses for fetal imaging have been the SSFSE,
HASTE, UFSE and FASE sequences. These sequences are
also known as the half-Fourier single-shot RARE (rapid
acquisition with relaxation enhancement) technique. With
this imaging, phase encoding for sections are obtained after
a single 90° radiofrequency excitation and multiple 180°
refocusing pulses. Because only half of the K-space is
sampled, the imaging time is decreased nearly twofold [10].
Data obtained from this sequence are of excellent quality
with good T2 weighting. More important, this imaging
allows acquisition of a single slice at a time with each
image being obtained in milliseconds, significantly mini-
mizing motion artifact. Thus, if the fetus moves, only the
slice or slices at the time of the motion in a group of images
are blurred.
Although SSFSE is excellent for fetal imaging, the half-
Fourier single-shot RARE technique has also been utilized
for other imaging purposes. In particular, if the TR is long
137
Fig. 8 Sagittal MR fetography image of the same fetus as in Fig. 7
demonstrates liver (large arrow) in a large omphalocele (long thin
arrow gallbladder within the defect, thin black arrow membrane; these
findings were best depicted on this sequence)
and the TE is increased to more than 140–160 ms, then a
heavily T2-weighted hydrography image is obtained [11].
With the hydrography method, the T2 effects of aqueous
liquids are amplified because water has a long transverse
and longitudinal magnetization when compared to other
tissues [12]. When utilizing this technique, solid organs and
flowing blood or fluids have low signal intensity, whereas
static fluids, such as cerebrospinal fluid, urine in the
collecting system, bile in the hepatobiliary system and
cysts in any organ, have high signal intensity [13].
Hydrography imaging, designated by the anatomical area
of investigation, provides high-quality imaging contrast
between fluid and background and is invaluable in studying
static fluid pathology. In the biliary system, MR cholangi-
ography is commonly used to evaluate bile duct obstruc-
tion, postsurgical alterations and congenital anomalies [14].
MR urography is an asset in the evaluation of the renal
parenchyma and perinephric processes and mimics excre-
tory urography, defining urinary tract obstruction, filling
defects in the urinary tract, and congenital anomalies [15,
16]. MR myelography produces images with excellent
spatial resolution of exiting nerve roots, the conus medul-
laris and the cauda equina [17].
138 Pediatr Radiol (2007) 37:133–140

Fig. 10 Sagittal MR fetography image of the same fetus as in Fig. 9.

Given the high water contrast, the image effectively demonstrates the
borders of the meningocele (large arrow), the cloacal anomaly (long
thin arrow) and the bladder (short thin arrow)

Fig. 9 Sagittal SSFSE image of a fetus at 18 weeks of gestation with

multiple anomalies shows a small distal meningocele (large arrow)
(short thin arrow bladder, long thin arrow cystic structure compatible
with a cloacal anomaly). The brain of the fetus was also abnormal

Because many areas studied with hydrography imaging

are common sites of fetal pathology, and because water is
such an abundant component of the fetus, MR fetography
was studied at our institution. The addition of a T2-
weighted sequence was also applied given the fact that
when performing MR examinations in children or adults, it
is common to utilize two levels of T2 weighting. For
Fig. 11 Sagittal SSFSE image of a fetus at 19 weeks of gestation with
example, in the knee and brain, proton density and T2-
a proven diagnosis of Walker-Warburg syndrome shows a dorsal kink
weighted images are usually obtained. This difference in of the brainstem at the mesencephalic-pontine junction (short arrow)
weighting often provides more information and increases and cerebellar vermian hypoplasia (long arrow)
Pediatr Radiol (2007) 37:133–140
Fig. 12 Sagittal MR fetography image of the same fetus as in Fig. 11
does not change the diagnosis but provides border edge enhancement
of the brainstem kink (short arrow), cerebellar vermian hypoplasia
(long arrow) and midline brain structures (arrowhead nuchal soft-
tissue swelling interface, accentuated by the imaging)
the confidence when describing an abnormality. With fetal
imaging at our institution, only one echo of T2-weighted
and T1-weighted sequences has been routine.
Although the addition of the MR fetography sequence
did not change the primary diagnosis, MR fetography
showed additional findings, particularly in the evaluation of
the kidneys and spine. In the kidney, because the organ has
a high water content, subtle changes in the parenchyma can
be difficult to detect on a single T2-weighted image. The
MR fetography accentuates the contrast between cyst and
normal renal parenchyma, separating these entities in
greater detail than the SSFSE. Similarly in the lung, MR
fetography was helpful in the identification of cysts in two
lung lesions by providing increased contrast to separate cyst
from lung lesions with a high T2 signal. In the spine and
brain, MR fetography provided excellent spatial resolution,
enhancing borders and elucidating architecture abnormali-
ties caused by the adjacent cerebral spinal fluid.
Subjectively, having two different T2 weightings in-
creased confidence in the findings. With MR fetography,
the ability to highlight fluid and the enhanced spatial
resolution were beneficial in the evaluation of most fetal
pathologies. Because borders of fluid-filled structures were
well-defined, it was easier to separate these structures from
adjacent soft-tissue organs. Even in the evaluation of very
small structures, such as the airway, biliary tree, or
aqueduct of Sylvius, MR fetography usually provided the
best definition of the anatomy. In a majority of the imaging,
the fluid-filled structure represented the pathology, so the
139
MR fetography exaggerated the imaging abnormality while
nulling the background. This was especially helpful in the
evaluation of organs with a high water content, as the MR
fetography was an asset in revealing normal or abnormal
signal in the brain, kidneys and lungs. In addition, if the
peripheral wall of the fluid-filled structures was abnormal,
for example bowel wall thickening or cerebral ventricle
wall nodularity, we were able to define this abnormality
better with the MR fetography sequence. Sometimes, the
fluid-filled structure was not simple, and the MR feto-
graphy provided improved knowledge of internal cyst or
fluid complexity (septation or debris). We therefore believe,
as imagers, that the MR fetography sequence is a useful
addition to our protocol, similar to a prior review of fetal
hydrography imaging [18], and our clinicians have also
found MR fetography images visually appealing and
helpful when reviewing fetal pathology. Because of high
contrast and resolution, MR fetography often provided the
best images with which to counsel families.
MR fetography was not helpful in evaluation of cardiac or
vascular abnormalities, as expected. The hydrography tech-
nique results in little or no signal in flowing blood. Therefore,
detail of vascular structures was poor. In cases where there
was little or no water in the fetal anatomy or pathology, the
MR fetography sequence was not beneficial. In pregnancies
with low amniotic fluid, there was decreased internal fetal
water content, and the images were dark with poor contrast.
Drawbacks of the addition of the heavily T2-weighted
sequences were noted. First, the overall length of the
examinations was increased, as more imaging was performed.
The MR fetography imaging time for each series tended to be
longer, as the acquisition of each image took approximately
four times longer than the acquisition each SSFSE image.
Image slice thickness of the fetography sequence had to be
increased to decrease the imaging time. For example, when
using SSFSE imaging, if 16 images were obtained at a slice
thickness of 4 mm, the time for acquisition was 54 s. The MR
fetography, covering a similar area of interest was performed
at thickness of 5 mm and 14 images were obtained in 97 s.
Also, as expected, the soft-tissue contrast deteriorated because
of the heavy T2 weighting. However, resolution was
increased because of the higher matrix.
With this information, we have changed our approach to
imaging. We believe that SSFSE should continue to be the
primary sequence for fetal MR imaging in all cases,
especially since the MR fetography sequence did not
change the primary diagnosis in any of our examinations.
However, in most single gestations, in addition to our
standard SSFSE sequence, we obtain MR fetography
sequences through areas of pathology. If there is a renal,
spinal, biliary or a cystic lesion, we perform multiple planes
of imaging with the heavily T2-weighted technique. If there
is severe oligohydramnios and/or absence of water in or
140
adjacent to an abnormality, then we do not perform MR
fetography. We also do not use MR fetography imaging
when evaluating a fetus with cardiac or vascular pathology.
Because multiple gestation examinations tend to be longer,
as they require imaging of more than one fetus, we usually
perform the SSFSE routine sequence and do not prolong
the examination with MR fetography imaging unless we
believe strongly that evaluation of fetal pathology would be
improved with inclusion of the sequence.
Conclusion
Fetal MR is becoming common practice for the evaluation of
many fetal disorders after the identification of an abnormality
on sonography. Fast imaging sequences implemented in the
1990s have improved fetal imaging by minimizing motion
artifact. Although SSFSE and HASTE imaging are currently
the workhorses for fetal imaging, MR fetography is an
excellent adjunct. Because of the high water content in
normal fetal anatomy and fetal pathology, this sequence can
elucidate additional findings, increase confidence in the
diagnosis, and provide high resolution and contrast images
that are well received by clinicians and families.
References
1. Coakley F, Glenn OA, Qayyum A et al (2004) Fetal MRI: a
developing technique for the developing patient. AJR 182:243–252
2. Glastonbury CM, Kennedy AM (2002) Ultrafast MRI of the fetus.
Australas Radiol 46:22–32
Pediatr Radiol (2007) 37:133–140
3. Frates CM, Dumar AJ, Benson CB et al (2004) Fetal anomalies:
comparison of MR imaging and US for diagnosis. Radiology
232:398–404
4. Quinn TM, Hubbard AM, Adzick NS (1998) Prenatal magnetic
resonance imaging enhances fetal diagnosis. J Pediatr Surg
33:553–558
5. Levine D (2001) Ultrasound versus magnetic resonance
imaging in fetal evaluation. Top Magn Reson Imaging 12:25–
38
6. Levine D (2006) Obstetric MRI. J Magn Reson Imaging 24:1–
15
7. Jara H, Barish MA, Yucel EK et al (1998) MR hydrography:
theory and practice of static fluid imaging. AJR 170:873–
882
8. Regan F, Calvaluzzi J, Nguyen B (1998) Fast MR abdominal
imaging using the HASTE sequence. AJR 170:1471–1476
9. Ellis KJ (1993) Human body composition: in vivo methods,
models and assessment. Plenum Press, New York
10. Chen Q, Levine D (2001) Fast fetal magnetic resonance imaging
techniques. Top Magn Reson Imaging 12:67–79
11. Barish MA, Soto JA (1997) MR cholangiopancreatography:
techniques and clinical applications. AJR 169:1295–1303
12. Westbrook C, Kaut C (1994) MRI in practice. Blackwell
Scientific Publications, Oxford, UK
13. Roy C, Saussine C, Jacqmin D (2000) Magnetic resonance
urography. BJU Int 86 [Suppl 1]:42–47
14. Fulcher AS, Turner MA, Capps GW (1999) MR cholangiography:
technical advances and clinical applications. Radiographics
19:25–41
15. Avni FE, Nicaise N, Hall M et al (2001) The role of MR imaging
for the assessment of complicated duplex kidneys in children:
preliminary report. Pediatr Radiol 31:215–223
16. Blandino A, Gaeta M, Minutoli F et al (2002) MR urography of
the ureter. AJR 179:1307–1314
17. Nagayam M, Watanabe Y, Okumura A et al (2002) High-
resolution single-slice MR myelography. AJR 179:515–521
18. Chaumoitre K, Wikberg E, Shojai R et al (2006) Fetal magnetic
resonance hydrography: evaluation of single-shot thick-slab
RARE (rapid acquisition with relaxation enhancement) sequence
in fetal thoracoabdominal pathology. Ultrasound Obstet Gynecol
27:537–544
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.