DOI:10.2214/AJR.04.1636
AJR 2005; 185:1060-1062
© American Roentgen Ray Society
Optimization of Acquisition Time for MRI of Fetal Head: The Eyes Have It
Keyanoosh Hosseinzadeh1,2 and
Erma Owens1
1 Diagnostic Imaging, University of Maryland School of Medicine, University of
Maryland Medical Center, Baltimore, MD 21201.
2 Present address: Department of Radiology (Abdominal Imaging), UPMC Health
Systems (Presbyterian Campus), 200 Lothrop St., Pittsburgh, PA 15213.
Received October 19, 2004;
accepted after revision December 6, 2004.
Patent pending: A provisional patent has been granted by the United States
Patent and Trademark Office for the concept being described in the
manuscript.
Address correspondence to K. Hosseinzadeh
(keyanooshh{at}gmail.com).
Abstract
OBJECTIVE. This article describes a technique to minimize MRI time
and obtain true orthogonal T2-weighted projections of the fetal head. The
technique takes advantage of the symmetry of fetal orbits to establish a line
of reference through the orbits to obtain true sagittal, coronal, and axial
projections of the intracranial anatomy.
CONCLUSION. This technique results in a 50% reduction in imaging
time and thus decreases fetal exposure to the electromagnetic field.
Introduction
MRI is increasingly used to assess the fetus after complicated or
nonspecific sonographic evaluation
[1-4].
The introduction of ultrafast T2-weighted sequences has enabled safe and rapid
imaging of the fetus with high spatial resolution. The most common indication
for fetal MRI is evaluation of intracranial anatomy. In an ideal setting,
imaging of the head requires acquisition of three orthogonal planes relative
to the fetus, where the planes represent true anatomic planes of the fetal
head. Depending on fetal position, several attempts are initially made to
obtain an image along an anatomic plane, after which the remaining two
orthogonal sequences can be acquired.
However, performing these orthogonal sequences is challenged by fetal
motion during or between the acquisitions, which can result in sequence
repetition and prolongation of the study
[4]. Thus, it is important to
minimize the delay between the localizer and final diagnostic images. We
describe a technique that, to the best of our knowledge, has not been reported
in the literature. Our technique enables rapid acquisition of three orthogonal
planes from the addition of a single sequence in which the acquisition plane
is aligned to the fetal orbits.
Technique
The mother fasts 4 hrs before the MRI to reduce bowel peristalsis and
prevent postprandial fetal motion artifacts. Before MRI, routine sonographic
screening is performed to document fetal position. MRI of the fetus is
performed with a 1.5-T superconductive system (Eclipse, Philips Medical
Systems) and a 4-element phased-array surface coil is used for signal
reception. At our institution, fetal imaging is performed either in a
paralyzed state or a nonsedated state. Patients are positioned supine and feet
first to minimize claustrophobia. Three plane scout images are obtained. A
single-shot fast spin-echo rapid acquisition T2-weighted (EXPRESS, Siemens
Medical Solutions) localizer image is then obtained of the fetal head either
axial or coronal to the maternal pelvis with the following parameters: TR/TE,
infinite/90; matrix, 256 x 192-256; echo-train length, 100-140; section
thickness, 5 mm; gap, 0 mm.
The position of the fetus within the uterus is independent of the maternal
pelvis. If the mother chooses a decubitus position, a coronal localizer of the
maternal pelvis can be chosen from the axial scout image. An axial localizer
of the maternal pelvis would be rotated; however, there would be no impact on
subsequent imaging because the fetal orbits are the regions of interest. From
the localizer acquisition, fetal orbits are identified from the separate
sections. The technologist or radiologist will then place a slice orientation
line of reference, which can undergo both rotation and translation on all
sections of the sequence. A slice orientation line of reference overlay is
available on all MRI systems.
The center axis of rotation is placed on a single orbit, preferably on the
low-signal-intensity lens or orbit center
(Fig. 1A). During image
scrolling on the MRI console, the line of reference is rotated to coincide
with the same location through the second orbit on a neighboring section
(Figs. 1B and
1C). This line of reference
simulates the projection of a line interconnecting the orbits in 3D on a 2D
plane (Fig. 2). Infrequently,
the orbits superimpose on one another during scrolling of the images on the
workstation, which implies that a true-sagittal image has already been
obtained. Alternatively, if the orbits are positioned symmetrically in the
same section, either a true-axial or a true-coronal image has been
obtained.
View larger version (103K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A —Ultrafast T2-weighted images obtained coronal to maternal
pelvis in 23-year-old woman. Single section containing one fetal orbit is
shown. Line of reference (line) is placed with center of axis on
orbit (crosshair). Sections are subsequently scrolled through on MRI
console with superimposed line of reference.
|
|
View larger version (106K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —Ultrafast T2-weighted images obtained coronal to maternal
pelvis in 23-year-old woman. Single section of fetus shows second orbit and
line of reference (line). Line of reference is rotated
counterclockwise (arrow) about center of axis (crosshair) to
cross second orbit at same point (dashed line).
|
|
View larger version (99K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1C —Ultrafast T2-weighted images obtained coronal to maternal
pelvis in 23-year-old woman. Single section of fetal head shows orbit
(highlighted) with second orbit (highlighted and
arrowhead) superimposed on displayed section to show relative
position of both orbits during image scrolling. Rotated line of reference is
seen to course through both orbits, which denotes plane of acquisition of
subsequent sequence.
|
|
View larger version (135K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2 —Three-dimensional representation of ultrafast T2-weighted
image acquisition coronal to maternal pelvis in 23-year-old woman. Sections
containing orbits have been shown and intervening sections have not. Sections
have been masked except for fetal heads, which have been highlighted.
Three-dimensional representation of line of reference connecting both orbits
(line) is displayed. Subsequent sequence yields image containing both
orbits symmetrically aligned about midline (asterisk).
|
|
Once accurate placement of the line of reference has been completed through
both orbits, which takes on average 30 sec, ultrafast T2-weighted imaging of
the head is performed in this designated plane. The resulting image of the
fetal head will always include the orbits centered symmetrically about a
midline sagittal plane (Fig.
2). The section that includes both orbits serves as the scout for
the subsequent orthogonal true-sagittal sequence, after which imaging of the
remaining orthogonal coronal and axial sequences can be performed in a similar
fashion, using each sequence as the scout for subsequent imaging. Thinner
section thicknesses (3-4 mm) are used for these three orthogonal sequences. If
necessary, T1-weighted gradient-echo sequence is performed in a desired plane
with the following parameters: TR/TE, 140/4.2; flip angle, 70°; matrix 256
x 160; section thickness, 5 mm; gap, 1 mm; one signal acquired.
To date our modified technique has been used successfully in five patients,
with a total imaging time of 19.5 ± 4.1 min (range, 16-24 min). When
compared with the old "hit or miss" approach in 10 patients, the
total imaging time was 39.2 ± 10.1 min (range, 22-54 min). The total
numbers of ultrafast T2-weighted sequences completed to achieve orthogonal
projections of the fetal head for the modified and old techniques were 5
± 0.5 (range, 4-5) and 8 ± 2 (range, 5-11), respectively.
An unpaired, one-tailed t test assuming unequal variance was
performed comparing the means of the total study time and the number of
T2-weighted sequences between the two patient groups (old technique and
modified technique). A p value of less than 0.05 was considered to be
statistically significant. Our study showed statistically significant
p values of 0.0001 and 0.0004, respectively. In the modified
technique, no fetal motion was encountered. In the old technique, three
patients required repetition of the scout imaging because of fetal motion.
Discussion
Standard prenatal evaluation of central nervous system (CNS) anomalies is
performed with sonography. However, sonography is limited by the nonspecific
appearance of some anomalies, difficulty in evaluation of the posterior fossa
in late gestation and of the side of the brain near the transducer, subtle
parenchymal abnormalities, and poor fetal visualization
[1]. Fetal MRI using ultrafast
T2-weighted sequences enables multiplanar imaging with direct visualization of
parenchyma, and thus detailed analysis of the CNS anatomy.
Fetal MRI is usually performed in the second or third trimester, based on
abnormalities detected by sonography. Although there is no conclusive evidence
that routine clinical use of MRI procedures during pregnancy produces
deleterious effects on the unborn fetus, every effort is made to minimize the
specific absorption rate and exposure of the fetus to the electromagnetic
field. Consequently, total imaging time is kept to a minimum.
One of the challenges of fetal MRI is evaluation of the CNS because
acquisition of images in three orthogonal planes relative to the fetus has
been shown to improve depiction of the intracranial anatomy
[4]. Most MRI systems are
designed for imaging of static subjects. Minimization of delay between
localization and diagnostic image acquisition is critical, especially in the
second trimester because the fetus is most active at this gestational age.
Although preliminary studies have been undertaken to image freely moving
subjects, there are currently no real-time localization interfaces that are
appropriate for freely (random and nonperiodic) moving subjects, and a fixed
frame of reference is used to describe slice orientation
[5].
Without the benefit of an anatomic landmark, several sequence repetitions
are often made of the fetus to align a single acquisition along an anatomic
plane, after which the remaining orthogonal projections can be obtained
relative to a scout section from a preceding acquisition. In effect,
orthogonal planes of acquisition similar to adult brain imaging are
reproduced. The use of our modified imaging algorithm has decreased the total
study time by 50% and provided optimal depiction of intracranial anatomy
because a maximum of five acquisitions are performed for T2-weighted images
including the initial localizer.
The decrease in imaging time is related predominantly to a significant
decrease in the number of T2-weighted acquisitions required for orthogonal
imaging. When necessary, a T1-weighted image can be obtained in a single plane
at the conclusion of the T2-weighted acquisitions to evaluate for fat and
hemorrhage. The fetal orbits are used as the fixed landmarks from the
T2-localizer image, and the success of this algorithm is based on the concept
of symmetry of the orbits in relation to the sagittal axis of the fetal
head.
After obtaining a localizer sequence of the maternal pelvis, a line of
reference drawn through the fetal orbits in 3D is translated into a line that
courses through the orbits on overlapping 2D images. This technique is
especially useful for nonsedated and nonparalyzed cases given the minimum
delay between the localizer and the final diagnostic sequences, which is
important to lessen the risk of fetal motion. However, we recognize that any
gross fetal motion during or between each acquisition is unpredictable and may
require reinitiating the localizer sequence to retrieve the proper orthogonal
projections. The modified technique enables the technologist to complete the
study in a timely manner with minimal, if any, supervision by the radiologist.
Thus, the time spent to check and launch the next plane of acquisition is
minimized.
References
- Coakley FV, Glenn OA, Qayyum A, Barkovich AJ, Goldstein R, Filly
RA. Fetal MRI: a developing technique for the developing patient.
AJR 2004; 182:243
-252[Free Full Text]
- Coakley FV, Hricak H, Filly RA, Barkovich AJ, Harrison MR. Complex
fetal disorders: effect of MR imaging on management—preliminary clinical
experience. Radiology 1999;213
: 691-696[Abstract/Free Full Text]
- Levine D, Smith AS, McKenzie C. Tips and tricks of fetal MR
imaging. Radiol Clin North Am 2003;41
: 729-745[CrossRef][Medline]
- Levine D, Barnes PD, Sher S, et al. Fetal fast MR imaging:
reproducibility, technical quality, and conspicuity of anatomy.
Radiology 1998;206
: 549-554[Abstract/Free Full Text]
- Neustadter DM, Chiel HJ. Imaging freely moving subjects using
continuous interleaved orthogonal magnetic resonance imaging. Magn
Reson Imaging 2004; 22:329
-343[CrossRef][Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
