DOI:10.2214/AJR.05.0930
AJR 2006; 187:1442-1447
© American Roentgen Ray Society
Iterative Decomposition of Water and Fat with Echo Asymmetry and Least-Squares Estimation (IDEAL) Fast Spin-Echo Imaging of the Ankle: Initial Clinical Experience
Samuel Fuller1,
Scott Reeder1,
Ann Shimakawa2,
Huanzhou Yu1,
Jane Johnson2,
Christopher Beaulieu1 and
Garry E. Gold1
1 Department of Radiology, Grant Building S0-68B, Stanford University, 300
Pasteur Drive, Stanford, CA 94305-5105.
2 GE Global Applied Science Laboratory, Menlo Park, CA.
Received June 1, 2005;
accepted after revision November 17, 2005.
Address correspondence to G. E. Gold
(gold{at}stanford.edu).
Abstract
OBJECTIVE. Reliable, uniform fat suppression is important. Multiple
approaches currently exist, many of which suffer from either suboptimal
signal-to-noise ratio (SNR), or the inability to obtain consistent fat
suppression around the ankle joint. Our purpose was to test iterative
decomposition of water and fat with echo asymmetry and the least-squares
estimation (IDEAL) method in combination with fast spin-echo imaging, which is
able to achieve reliable high SNR images with uniform fat-water
separation.
SUBJECTS AND METHODS. We compared IDEAL fast spin-echo with
conventional fat-suppressed fast spin-echo imaging in 33 ankles in 32
patients. Quantitative measurements of SNR and contrast-to-noise ratio
efficiency were made, and qualitative diagnostic image quality and
fat-suppression scores were determined.
RESULTS. We found that the SNR efficiency for both cartilage and
fluid was similar for both techniques, and fluid/cartilage contrast-to-noise
ratio efficiency was higher with IDEAL fast spin-echo imaging. Fat suppression
and diagnostic quality scores using the IDEAL method were superior (p
< 0.01) to fat-suppressed fast spin-echo imaging.
CONCLUSION. IDEAL fast spin-echo imaging is a promising technique
for MRI of the ankle.
Keywords: ankle • cartilage • foot • MR technique • musculoskeletal imaging
Introduction
MRI is useful in assessing soft-tissue and cartilaginous disorders of the
ankle in musculoskeletal radiology
[1]. In particular, the status
and integrity of hyaline and articular cartilage are far better evaluated by
MRI than by any other imaging method
[2]. This is very important in
the setting of both acute disorders (e.g., osteochondral injuries), and
chronic joint pathology, such as cartilage degeneration associated with
osteoarthritis [3].
Robust and uniform fat suppression is extremely important in
musculoskeletal imaging, and the detection of subtle foci of fluid and/or bone
marrow edema is critical to accurate diagnoses. Inhomogeneous fat suppression
on T2-weighted images may lead to erroneous findings that mimic pathology
[4]. In cases of extreme field
inhomogeneity, fat-suppression pulses can even saturate water signal, and
severely degrade image quality.
Many methods exist for obtaining fat-suppressed images. Fast spin-echo
imaging with chemically selective fat suppression pulses is commonly used in
ankle MRI. However, achieving uniform fat suppression in the ankle joint with
fat-suppressed fast spin-echo methods is difficult. This is a consequence of
the presence of Bo and B1 inhomogeneities, both of which
worsen with increasing field strength, off-isocenter imaging, and the
unfavorable geometry of the ankle and foot that worsens susceptibility
effects. STIR techniques provide uniform fat saturation; however, STIR is
sensitive to B1 inhomogeneity, requires additional time for
inversion pulses, and suffers from a low signal-to-noise (SNR) ratio
[5]. Furthermore,
contrast-enhanced T1-weighted images cannot be obtained using STIR sequences,
because enhancing tissue with T1 similar to fat would be erroneously
suppressed. Spectral-spatial pulses are insensitive to B1
inhomogeneities, but are relatively sensitive to magnetic field
(Bo) inhomogeneities and are complex pulses of long duration
[6].
Dixon fat-water separation methods are proven to have several advantages in
musculoskeletal imaging [7,
8]. These techniques are
insensitive to both Bo and B1 inhomogeneities and
obviate fat-suppression pulses because separate water and fat images are
obtained. In addition, the separated water and fat images can be recombined
and corrected for chemical shift, facilitating acquisition of images with
lower bandwidth. Dixon fat-water separation methods may be useful for
producing images at higher field strength that are free from chemical shift
artifact [9]. We previously
described an iterative least-squares decomposition of water and fat with echo
asymmetry and least-squares estimation water-fat separation method (IDEAL) in
combination with fast spin-echo imaging that allow superior water-fat
separation, and comparable signal-to-noise and contrast-to-noise ratios in
comparison with traditional fast spin-echo techniques
[10,
11]. Advantages of IDEAL with
respect to conventional Dixon methods include its compatibility with multicoil
applications, and its optimized SNR through the use of asymmetric echoes
[11,
12].
View larger version (99K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1D —24-year-old woman with ankle pain. Fat-suppressed fast
spin-echo image. Note homogeneous, excellent fat-water separation in IDEAL
fast spin-echo image (B) as compared with fat-suppressed fast spin-echo
image (D). Multiple areas of failed fat saturation (arrows)
are seen in both bone and soft tissues.
|
|
The purpose of our current study was to apply IDEAL to imaging of the
ankle, an area in which it has been difficult to achieve optimal fat
suppression. Fat suppression in the ankle is more difficult than the knee
because of differences in bulk susceptibility. This article compares the IDEAL
fast spin-echo method to conventional fat-suppressed fast spin-echo in imaging
33 ankles of 32 patients.
Subjects and Methods
Imaging
We studied 32 consecutive patients (17 men, 15 women; age range, 19-78
years) who had been referred for routine ankle MRI by their primary care
physicians. Patients were taken from the routine referral for ankle imaging at
our institution, which includes, primarily, indications for ankle pain and
trauma. Two subjects had lateral metallic fibula fixation plates with screws.
Imaging was performed on a GE Signa 1.5-T scanner (TwinSpeed, GE Healthcare)
using a quadrature extremity coil. Informed consent from each patient and
approval of our institutional review board were obtained before imaging, and
all aspects of the study were Health Insurance Portability and Accountability
Act (HIPAA) compliant.
Images were obtained as a part of the routine ankle protocol at our
institution. Comparison images between the two sequences were obtained in the
coronal plane only. IDEAL fast spin-echo image parameters included TR/TE,
4,600/22 milliseconds; bandwidth, +20 kHz; echo-train length, 8; field of
view, 12 cm; slice thickness, 3 mm; 24-27 coronal images; 512 x 192
matrix; and one signal average, for a total scanning time of 5:31 minutes.
Reeder et al. [11] recently
published detailed information about the IDEAL fast spin-echo sequence.
Conventional fat-suppressed fast spin-echo images were obtained at the same
slice locations with identical scan parameters except TE (17 milliseconds) and
the acquisition of two signal averages, for a total imaging time of 3:41
minutes. Frequency encoding was performed in the superior-to-inferior
direction.
View larger version (104K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B —20-year-old man with foot pain. Coronal iterative
decomposition of water and fat with echo asymmetry and least-squares
estimation (IDEAL) fast spin-echo water-only image shows cyst
(arrow).
|
|
View larger version (115K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2D —20-year-old man with foot pain. Coronal fat-suppressed fast
spin-echo image shows cyst (solid arrow). Note superior fat-water
separation in region of lateral malleolus in IDEAL fast spin-echo image
(B) as compared with failure of fat saturation that might be mistaken
for bone contusion in fat-suppressed fast spin-echo image (dashed
arrow).
|
|
The TE values used for the two sequences differed because of the slight
differences in echo spacing required for adjustment of the readout gradient
with respect to the spin-echo for IDEAL
[11]. Images were
reconstructed with an online implementation of the IDEAL algorithm that uses a
robust, region-growing reconstruction method
[13]. The online
reconstruction consisted of a button on the scanner pressed by the
technologist after the IDEAL images were acquired, and took approximately 2
minutes. Fat, water, and combined images were then transferred to the PACS for
display.
Image Evaluation
Cartilage signal was measured for both IDEAL fast spin-echo and
fat-suppressed fast spin-echo imaging, from a 2-mm circular region of interest
(ROI) in the articular cartilage on the talar dome. The signal from joint
fluid was also measured using a 3-mm ROI from the subtalar joint, and the
signal from noise was measured from a 9-mm ROI that was inferior in relation
to the ankle in a region free of phase ghosting artifact. ROIs were placed to
avoid regions of inhomogeneous fat saturation and metal artifact.
Cartilage SNR efficiency, fluid SNR efficiency, and cartilage/fluid
contrast-to-noise ratio efficiency were computed. SNR was defined as the
measured signal divided by the square root of the noise. Contrast-to-noise
ratio was defined as the difference between SNR for the two tissues. SNR and
contrast-to-noise ratio efficiencies were calculated to provide a fair measure
of SNR and contrast-to-noise ratio that corrects for differences in scanning
time between the two sequences. Specifically, IDEAL fast spin-echo SNR and
contrast-to-noise ratio measurements were multiplied by the square root of the
quotient of fat-suppressed fast spin-echo and IDEAL fast spin-echo scan
times.
Subjective scoring of the images by two experienced musculoskeletal
radiologists (5 and 12 years' experience, respectively) was done by consensus.
The two radiologists evaluated fat suppression or fat-water separation and
diagnostic quality on a 4-point scale: 0, poor; 1, fair; 2, good; 3,
excellent. The same two radiologists made all SNR measurements. Observers were
not blinded as to the type of acquisition they were grading, but comparisons
between images for image quality and fat suppression or fat-water separation
were done with the water frequency IDEAL fast spin-echo images and the
fat-suppressed fast spin-echo images only.
Statistical Analysis
Statistical analysis was performed using Excel 11.1.1 (Microsoft). The two
imaging sequences were compared with respect to SNR and contrast-to-noise
ratio efficiency using a paired sample Student's t test. The mean
scores of the two radiologists for IDEAL fast spin-echo and fat-suppressed
fast spin-echo sequences also were compared with respect to image quality and
uniformity of fat suppression or fat-water separation using a Wilcoxon's
signed rank test. Results were considered significant for both quantitative
and qualitative analyses at p < 0.05.
View larger version (115K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —Images of ankle in 40-year-old woman with metallic plate in
fibula after fracture. Recombined fat-water iterative decomposition of water
and fat with echo asymmetry and least-squares estimation (IDEAL) fast
spin-echo image.
|
|
View larger version (115K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3D —Images of ankle in 40-year-old woman with metallic plate in
fibula after fracture. Fat-suppressed fast spin-echo image. Note superior
fat-water separation of IDEAL fast spin-echo image (B) in region of
metal plate compared with failure of fat saturation in fat-suppressed fast
spin-echo image (D, arrows).
|
|
Results
Cartilage SNR efficiency was similar for IDEAL fast spin-echo (37.8
± 8.9) and fat-suppressed fast spin-echo (39.3 ± 9.2)
(p > 0.05). Fluid SNR efficiency was also similar for IDEAL fast
spin-echo (77.0 ± 19.4) and fat-suppressed fast spin-echo imaging (78.9
± 22.0) (p > 0.05). However, fluid-cartilage
contrast-to-noise ratio was significantly higher for IDEAL fast spin-echo
(41.1 ± 17.5) compared with fat-suppressed fast spin-echo imaging (37.7
± 14.5; p <0.05).
Overall image quality was higher (p <0.05) using IDEAL fast
spin-echo (2.9) as compared with to fat-suppressed fast spin-echo (2.5)
imaging. Fat suppression was scored as fair to good on fat-suppressed fast
spin-echo images (2.0), but fat-water separation was deemed excellent (2.9) on
all but one of the IDEAL fast spin-echo images (Figs.
1A,
1B,
1C, and
1D). The difference between fat
suppression and fat-water separation using IDEAL was statistically significant
(p < 0.005). In several instances, areas of poor fat suppression
on fat-suppressed fast spin-echo images were seen mimicking bone marrow edema
(Figs. 2A,
2B,
2C, and
2D), but superior fat-water
separation in the corresponding IDEAL fast spin-echo water image showed that
this was an artifact caused by poor fat saturation. Fat-water separation with
IDEAL fast spin-echo imaging performed particularly well in the presence of
metal, showing fewer artifacts than fat-suppressed fast spin-echo imaging
(Figs. 3A,
3B,
3C,
3D,
4A,
4B,
4C, and
4D).
View larger version (114K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4D —44-year-old woman with ankle pain. Fat-suppressed fast
spin-echo image. Osteochondral lesion is present in medial talar dome
(large arrow). The IDEAL fast spin-echo image (B) displays
superior fat-water separation in soft tissues on lateral aspect of ankle and
medial aspect of foot, and reduced artifact and superior fat saturation
associated with metallic hardware in distal fibula (small
arrows).
|
|
Discussion
IDEAL fast spin-echo imaging showed many advantages over fat-suppressed
fast spin-echo imaging in our study. Among these advantages were superior
fat-water separation, improved diagnostic image quality, and higher
fluid-cartilage contrast-to-noise ratio. Of particular importance is that
IDEAL fast spin-echo imaging reliably achieved excellent fat suppression in
areas susceptible to consistent failure of fat suppression on the
fat-suppressed fast spin-echo images—near the medial and lateral
malleoli, in the distal tibia, and in the calcaneus. The IDEAL method even
achieved excellent fat-water separation despite the presence of metallic
hardware that caused failure of fat saturation on conventional fat-suppressed
fast spin-echo images.
Another benefit of IDEAL imaging is that in addition to water-only images,
fat-only and combined water-fat images are generated, and the combined images
are corrected for chemical shift artifact
[9]. Chemical shift artifact is
increased at lower bandwidth or higher field strength, so bandwidth is
typically increased at 3.0 T to offset this artifact. Correction of chemical
shift with IDEAL may make increased bandwidth at 3.0 T unnecessary, preserving
the full benefit of the higher field SNR. Additional studies are required to
determine the full benefits of IDEAL fast spin-echo at higher field
strengths.
That IDEAL yields both fat and combined images permitted elimination of
coronal T1-weighted images from our protocol, resulting in a net examination
time reduction of approximately 2 minutes. IDEAL fast spin-echo imaging may
eliminate the need to acquire more than one sequence in the same plane in an
imaging protocol. Although our protocol does not acquire coronal T1-weighted
images, the combined IDEAL fast spin-echo images are of high quality and
provide anatomic reference similar to a T1-weighted image. Further study is
required to determine if removing the T1-weighted coronal images from an ankle
protocol results in the loss of important information. T1-weighted images were
still acquired in the axial plane, however, which was sufficient to evaluate
the bone marrow in our study.
It is important to note that it is possible to obtain T1-weighted IDEAL
fast spin-echo images. These could be of particular benefit for performing
studies using intraarticular or IV gadolinium, such as with MR arthrography.
This may provide a means for using fat-suppressed MR arthrography in areas of
metal or an inhomogeneous field. This could also be useful in visualizing
methemoglobin, melanin, or other species seen on fat-suppressed T1-weighted
images.
The scanning time of IDEAL fast spin-echo imaging was longer than that for
conventional fat-suppressed fast spin-echo imaging, because IDEAL requires
three acquisitions to separate water from fat. However, there are several
strategies to reduce scanning time, including parallel imaging
[14,
15], reduced sampling
strategies [16,
17], and the use of half
k-space acquisitions and homodyne reconstruction
[18]. Another limitation of
this study was that phase wrap was not available with IDEAL fast spin-echo.
IDEAL fast spin-echo imaging was performed in the coronal plane to avoid phase
wrap but still visualize the cartilage in the talar dome. Further study is
required to determine the optimal scan parameters. One final limitation of our
study was the lack of comparison of our technique with conventional Dixon
imaging methods. Although scanning time limitations did not permit us to add
conventional Dixon imaging methods to our protocol, future studies will
include such a comparison.
In conclusion, IDEAL fast spin-echo imaging is a promising technique for
ankle MRI because it provides superior fat suppression compared with
conventional fat-suppressed fast spin-echo imaging while maintaining high SNR
and image quality. Ankle studies performed with more patients are needed to
verify these results. Additional studies also must be done to assess the
performance of IDEAL in other musculoskeletal imaging applications, such as
evaluation of the shoulder, knee, and wrist, and high field imaging at 3.0
T.
References
- Kerr R. MRI of soft tissue disorders of the ankle. Clin
Podiatr Med Surg 2002; 19:285
-307[CrossRef][Medline]
- Disler DG, Recht MP, McCauley TR. MR imaging of articular
cartilage. Skeletal Radiol 2000;29
: 367-377[CrossRef][Medline]
- Gold GE, McCauley TR, Gray ML, Disler DG. What's new in cartilage?
RadioGraphics 2003;23
: 1227-1242[Abstract/Free Full Text]
- Leffler S, Disler DG. MR imaging of tendon, ligament, and osseous
abnormalities of the ankle and hindfoot. Radiol Clin North
Am 2002; 40:1147
-1170[CrossRef][Medline]
- Bydder GM, Pennock JM, Steiner RE, Khenia S, Payne JA, Young IR.
The short TI inversion recovery sequence—an approach to MR imaging of
the abdomen. Magn Reson Imaging 1985;3
: 251-254[CrossRef][Medline]
- Meyer CH, Pauly JM, Macovski A, Nishimura DG. Simultaneous spatial
and spectral selective excitation. Magn Reson Med1990; 15:287
-304[Medline]
- Bredella MA, Losasso C, Moelleken SC, Huegli RW, Genant HK, Tirman
PF. Three-point Dixon chemical-shift imaging for evaluating articular
cartilage defects in the knee joint on a low-field-strength open magnet.
AJR 2001; 177:1371
-1375[Abstract/Free Full Text]
- Kornaat PR, Reeder SB, Koo S, et al. MR imaging of articular
cartilage at 1.5 T and 3.0 T: comparison of SPGR and steady state free
precession (SSFP) sequences. Osteoarthritis Cartilage2005; 13:338
-344[CrossRef][Medline]
- Yu H, Reeder S, Shimakawa A, Gold G, Pelc N, Brittain J.
Implementation and noise analysis of chemical shift correction for fast spin
echo Dixon imaging. (abstr) Proceedings of the 12th annual meeting
of ISMRM. Kyoto, Japan: International Society of Magnetic
Resonance in Medicine, 2004:2686
- Reeder SB, Wen Z, Yu H, et al. Multicoil Dixon chemical species
separation with an iterative least-squares estimation method. Magn
Reson Med 2004; 51:35
-45[CrossRef][Medline]
- Reeder S, Pineda A, Wen Z, et al. Iterative "Dixon"
water-fat separation with echo asymmetry and least squares estimation (IDEAL):
application with fast spin-echo imaging. Magn Reson
Med 2005; 54:636
-644[CrossRef][Medline]
- Pineda AR, Reeder SB, Wen Z, Pelc NJ. Cramer-Rao bounds for
three-point decomposition of water and fat. Magn Reson
Med 2005; 54:625
-635[CrossRef][Medline]
- Yu H, Reeder S, Shimakawa A, Brittain J, Pelc N. Robust field map
estimation in a Dixon water-fat separation algorithm with short echo time
increments. (abstr) Proceedings of the 12th annual meeting of
ISMRM. Kyoto, Japan: International Society of Magnetic Resonance
in Medicine, 2004: 345
- McKenzie C, Reeder S, Shimakawa A, Pelc N, Brittain J. Abdominal
three point Dixon imaging with self calibrating parallel MRI. (abstr)
Proceedings of the 12th annual meeting of ISMRM.
Kyoto, Japan: International Society of Magnetic Resonance in Medicine,2004
: 917
- Reeder S, McKenzie C, Markl M, Yu H, Pelc N, Brittain J. Parallel
cardiac CINE imaging: application to Dixon water-fat separation and
steady-state free precession. (abstr) Proceedings of the 12th
annual meeting of ISMRM. Kyoto, Japan: International Society of
Magnetic Resonance in Medicine, 2004:267
- Brau A, McKenzie CA, Shimakawa A, et al. Accelerated IDEAL
water-fat separation techniques for single- and multi-coil applications.
(abstr) Proceedings of the 13th annual meeting of
ISMRM. Miami, FL: International Society of Magnetic Resonance in
Medicine, 2005: 491
- Yu H, McKenzie CA, Shimakawa A, et al. Parallel imaging accelerated
single acquisition water-fat separation for dynamic imaging. (abstr)
Proceedings of the 13th annual meeting of ISMRM.
Miami, FL: International Society of Magnetic Resonance in Medicine,2005
: 2390
- Reeder SB, Hargreaves BA, Yu H, Brittain JH. Homodyne
reconstruction and IDEAL water-fat decomposition. Magn Reson
Med 2005; 54:586
-593[CrossRef][Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
G. E. Gold, C. A. Chen, S. Koo, B. A. Hargreaves, and N. K. Bangerter
Recent Advances in MRI of Articular Cartilage
Am. J. Roentgenol.,
September 1, 2009;
193(3):
628 - 638.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
