| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, Osaka University Medical School, 2-2, Yamadaoka,
Suita-city, Osaka 565-0871, Japan
2
Department of Radiology, Gifu University School of Medicine, 40, Tsukasamachi,
Gifu 500-8705, Japan.
Received June 30, 1999;
accepted after revision October 13, 1999.
Address correspondence to M. Hori.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. T2-weighted MR images were retrospectively reviewed from 51 patients with 85 solid and 59 nonsolid lesions using the following four sequences: conventional spin-echo, respiratory-triggered fast spin-echo, single-shot fast spin-echo, and multishot spin-echo echoplanar imaging. Images were reviewed on a hepatic segment-by-segment basis; T2-weighted images of a total of 408 hepatic segments were reviewed separately and independently for solid and nonsolid lesions by four radiologists. Quantitative, qualitative, and receiver operating characteristic analyses were performed.
RESULTS. For solid lesions, no significant differences were seen among the lesion-to-liver contrast-to-noise ratios with the four sequences. In terms of solid lesion detection, no significant difference was seen between the diagnostic accuracy of multishot spin-echo echoplanar (Az = 0.90) and respiratory-triggered fast spin-echo (Az = 0.91) imaging, which showed the best performance of the four sequences. For nonsolid lesion detection, respiratory-triggered fast spin-echo and single-shot fast spin-echo imaging were judged the best (Az = 0.94).
CONCLUSION. Breath-hold single-shot fast spin-echo and multishot spin-echo echo-planar sequences can be substituted for conventional spin-echo and respiratory-triggered fast spin-echo T2-weighted sequences.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Recent improvements in gradient technology have made it possible to obtain T2-weighted single breath-hold images of the entire liver. Single-shot fast spin-echo sequences and half-Fourier single-shot turbo spin-echo sequences are single-shot fast spin-echo techniques that use a single 90° excitation pulse followed by multiple 180° pulses to cover half of the K-space. This technique is often used for MR cholangiopancreatography [12] and MR urography [13]. Single-shot fast spin-echo sequences can cover the entire liver in 30 sec, which is a distinct advantage over other breath-hold fast spin-echo techniques that require more than one breath-hold to image the entire liver. The spin-echo echoplanar pulse sequence is also capable of obtaining T2-weighted single breath-hold MR images of the entire liver [7]. This technique is available at many facilities. Two main types of spin-echo echoplanar imaging are available: single-shot echoplanar and multishot echoplanar. With single-shot echoplanar imaging, the entire liver can be studied in 2 sec. However, one of the major problems of this sequence is the presence of magnetic susceptibility artifacts, particularly at the air—tissue interface. With multishot echoplanar imaging, a reduction in susceptibility artifacts can be achieved, although the imaging time is longer than that for single-shot echoplanar imaging.
As mentioned before, single breath-hold imaging of the entire liver is expected to be useful because of the short imaging time and the reduction in motion artifacts. Some reports [7, 14, 15] have indicated the potential of single breath-hold imaging. However, to our knowledge, no reports have clearly shown the usefulness of single breath-hold imaging, including the multi-shot echoplanar technique, for the detection of solid or nonsolid hepatic tumors separately by means of receiver operating characteristic analysis. Thus, the purpose of this study was to evaluate the diagnostic usefulness of single breath-hold single-shot fast spin-echo and multishot spin-echo echoplanar imaging in comparison with conventional spin-echo and respiratory-triggered fast spin-echo imaging.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MR Imaging Techniques
MR imaging was performed using a 1.5-T superconducting magnet with blipped
echoplanar capability (Signa; General Electric Medical Systems, Milwaukee,
WI). The system provides a maximum gradient strength of 23 mT ·
m-1 with a peak slew rate of 77 mT · m-1 ·
msec-1. All MR images were obtained in the axial plane. A body coil
and a quadrature surface coil (Quad Abdominal Flex Coil; Medical Advances,
Milwaukee, WI) were used for transmission and reception, respectively. The
section thickness was 8 mm, with a 2-mm intersection gap for all pulse
sequences. The imaging protocol included T1-weighted spin-echo imaging (TR/TE,
500/28; two signals acquired) and T2-weighted imaging with (1) conventional
spin-echo sequences (2000/80, two signals acquired, 256 x 160 matrix,
±16-kHz receiver band-width, 30 x 30 cm field of view, ordered
phase encoding, gradient moment nulling in the frequency-encoding direction,
11.7-min acquisition time); (2) respiratory-triggered fast spin-echo sequences
(effective TR range/effective TE, 3157-7028 [mean, 4549]/75, echo train length
of eight, three signals acquired, 13-msec interecho spacing, 512 x 160
matrix, ±32-kHz receiver bandwidth, 30 x 30 cm field of view, 20%
respiratory trigger point, 40% trigger window, gradient moment nulling in the
slice-select direction, acquisition of 12-28 locations [mean, 16.8] per 3-7.5
min [mean, 4.9 min]); (3) breath-hold single-shot fast spin-echo sequences
(effective TE, 80; half-Fourier acquisition; echo train length, 104; one
signal acquired; 5.0-msec interecho spacing; 256 x 192 matrix;
±62.5-kHz receiver bandwidth; 30 x 23 cm field of view; gradient
moment nulling in the frequency-encoding direction; acquisition of 15-22
locations [mean, 18.1] per 21-31 sec [mean, 25.6 sec]); and (4) breath-hold
multishot spin-echo echoplanar sequences (1739-2812 [mean, 2232]/80, fat
suppression with spatial—spectral radiofrequency excitation, eight
shots, one signal acquired, 256 x 128 matrix, ±130-kHz receiver
bandwidth, 36 x 28 cm field of view, acquisition of 14-21 locations
[mean, 16.5] per 16-25 sec [mean, 20.1 sec]; gradient moment nulling was not
used). For breath-hold multishot spin-echo echoplanar imaging, a pulse
oximeter was attached to the patient's forefinger to gate data collection and
to reduce artifacts caused by phase errors due to pulsation (50-65% peripheral
gating trigger point, 10-20% trigger window). The gating trigger point is the
moment at which scanning begins and is expressed as a percentage of the
amplitude of the oximetric pulse wave. The trigger window is the specified
interval of no data acquisition to allow variations in the pulsation pattern.
This parameter is also expressed as a percentage of the cycle. With all MR
imaging sequences, spatial presaturation pulses were applied superior and
inferior to the imaging volume. The frequency selected fat-suppression
technique was not used for conventional spin-echo, fast spin-echo, or
single-shot fast spin-echo imaging.
Quantitative Analysis
The quantitative analysis was conducted with images obtained with the four
pulse sequences by using the operator-defined region-of-interest measurements
of mean signal intensity in the liver and hepatic lesions and of background
noise. The signal intensity in the liver was measured in areas selected to
avoid focal changes in signal intensity, large vessels, and prominent
artifacts. For measurement purposes, representative lesions of each of the
patients were chosen for analysis. When more than one lesion was present in a
patient, the largest lesion was identified and measured. Thirtyfour solid
(hepatocellular carcinomas [n = 17], metastases [n = 17])
and 26 nonsolid (cavernous hemangiomas [n = 18], cysts [n =
8]) lesions 8 mm in diameter or larger (equal to the imaging thickness) were
studied. One patient with metastasis, one patient with hemangioma, and two
patients with cysts were excluded from the quantitative lesion analysis
because their lesions were smaller than 8 mm in diameter. For the liver
lesions, a circular region of interest was drawn to encompass as much of the
lesion as possible. The standard deviation of background noise was measured
with the largest possible region of interest positioned in the phase-encoding
direction outside the abdominal wall to account for motion artifacts, after
which the following were calculated: liver signal-to-noise ratio =
SIlesion / SD, and lesion-to-liver contrast-to-noise ratio =
(SIlesion — SIliver) / SD, where
SIliver and SIlesion are the respective signal
intensities of the liver and the lesion and SD is the standard deviation of
background noise.
Qualitative Analysis
MR images were reviewed independently by four observers who were mainly
gastrointestinal radiologists for 5-11 years and who interpreted MR images of
the liver as part of their daily clinical and research practice. All observers
had experience in evaluating multishot spin-echo echoplanar images for more
than 1 year. Each observer independnetly evaluated the degree of image quality
degradation caused by respiratory ghost, pulsatile blood flow ghost, and
susceptibility artifacts using a four-point scale (1 = severe, 2 = moderate, 3
= mild, 4 = minimal or absent). A "severe" score was assigned when
the image could not be interpreted because of an artifact; a
"mild" score was assigned when the artifact was present but did
not markedly preclude interpretation. A "moderate" score was
assigned according to the observer's subjective judgment.
Lesion Detection with Receiver Operating Characteristic Analysis
The image review was conducted on a segment-by-segment basis because our
objective was to assess the ability of radiologists to detect lesions on
images obtained with each pulse sequence and not to localize the lesions. To
prevent mislocation of the lesions by the observers, the hepatic segment
numbering system of Couinaud
[16] was drawn on the images
by the study coordinator. A total of 408 liver segments were reviewed,
including 69 segments with 85 solid lesions (hepatocellular carcinoma
[n = 21], metastases [n = 64]) and 52 segments with 59
non-solid lesions (cavernous hemangiomas [n = 38], cysts [n
= 21]). No images were excluded because of poor image quality.
The review was performed at four separate sessions. In one review session, images obtained with only one of the four pulse sequences for all patients were reviewed. The four reviewing sessions were performed at 2-week intervals. The observers were unaware of patient name, clinical history, and imaging sequence, although it is possible that the observers recognized the imaging sequence by distinctive image features such as field of view, fat suppression, or artifacts.
For each technique, each observer independently recorded the size and site (Couinaud segment) of visible abnormalities and indicated, for every segment, whether the presence of solid or nonsolid lesions, separately, could be ascertained. A continuous rating scale was used to represent the confidence of each of the observers. The observers were instructed to assign a confidence value between 0 and 100 (0 = definite absence, 50 = equivocal, 100 = definite presence), with a higher value indicating a higher degree of confidence. Intermediate values were assigned on the basis of the observer's subjective judgment. The observers were asked to maintain a consistent and uniform interpretation of the confidence rating scale throughout the four interpretation sessions. Before the first session, the observers underwent a training session for familiarization with the observer test. When a lesion was located in two or more segments, the observer was asked to consider only the segment that was mainly involved and to assess the probability of another lesion in the other segments. Because observers simultaneously evaluated each hepatic segment for the presence of solid and nonsolid lesions at the same session, two confidence values were assigned for an individual hepatic segment. For example, when one observer thought that a hepatic segment probably harbored a solid lesion but probably not a nonsolid lesion, confidence scores of 85 and 30 were assigned to the individual hepatic segment for the presence of solid and nonsolid lesions, respectively.
Data Analysis
The liver signal-to-noise ratio and lesion-to-liver contrast-to-noise ratio
were compared on images obtained with the four pulse sequences using the
Friedman test for the overall test and Wilcoxon's signed rank test for
comparison between two groups. To assess interobserver variability in
interpreting images, the nonweighted kappa statistic with binary data was used
to measure the extent of agreement among the four observers. The binary data
were calculated from the confidence rating value assigned to a segment for
each of the observers. Binary values of 0 and 1 were assigned to segments that
had confidence ratings of less than 50 and equal to or greater than 50,
respectively. The extent of disagreement was not factored in the calculation.
Kappa values larger than zero were considered to indicate a positive
correlation; values up to 0.40, positive but poor agreement; values of
0.41-0.75, good agreement; and values greater than 0.75, excellent
agreement.
For each of the four pulse sequences, a binomial receiver operating characteristic curve was fitted to a given observer's confidence rating data by means of a maximum-likelihood estimation. Composite receiver operating characteristic curves were used to represent the performance of the four observers as a group and were calculated using the maximum-likelihood curve-fitting algorithm to rate the pooled data of the four independent observers [17]. A computer program, ROCKIT 0.9B (Metz C, University of Chicago, IL), was used. The diagnostic accuracy for a given technique was estimated by calculating the area under the receiver operating characteristic curve (Az). Differences between receiver operating characteristic curves were tested for statistical significance using a univariate z score test of the difference between the two Az values.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Qualitative Analysis
Results of the four observers are summarized in
Table 2. With breath-hold
single-shot fast spin-echo imaging, we observed the lowest degree of image
degradation caused by respiratory ghost, pulsatile blood flow ghost, and
susceptibility artifacts (Fig.
1A,1B,1C,1D).
Although breath-hold multishot spin-echo echoplanar imaging showed the lowest
score for susceptibility artifacts, severe distortion or signal deterioration
of the liver caused by susceptibility artifacts was not observed even at the
dome of the liver (Fig.
1A,1B,1C,1D).
Sufficient fat suppression was observed in breath-hold multishot spin-echo
echoplanar imaging.
|
|
|
|
|
Interobserver Variability
The kappa values for the four observers calculated on the basis of each
observer's confidence levels for the receiver operating characteristic
analysis are summarized in Table
3. Good agreement was obtained among the observers with regard to
the presence of lesions in a given segment.
|
Lesion Detection Capability
The Az values for each observer and the pooled data of
the four MR pulse sequences in terms of solid and nonsolid lesions are shown
in Table 4. The composite
receiver operating characteristic curves generated from the pooled data of the
four observers for solid and nonsolid lesions are shown in Figures
2 and
3, respectively.
|
|
|
All four observers achieved optimal performance for solid lesion detection with respiratory-triggered fast spin-echo sequences of the four imaging techniques, although multishot spin-echo echoplanar imaging for observer 3 showed the same detection performance as did respiratory-triggered fast spin-echo imaging. In terms of mean Az values for the composite receiver operating characteristic curves, no significant differences were noted among the Az values of conventional spin-echo (Az = 0.89), respiratory-triggered fast spin-echo (Az = 0.91), and multishot spin-echo echoplanar (Az = 0.90) imaging. The diagnostic accuracy for detection of solid lesions with breath-hold single-shot fast spin-echo images (Az = 0.84) was statistically significantly lower than with conventional spin-echo (p < 0.01), respiratory-triggered fast spin-echo (p < 0.0005), and multishot spin-echo echoplanar (p < 0.05) images (Fig. 4A,4B,4C,4D). For nonsolid lesion detection, respiratory-triggered fast spin-echo (Az = 0.94) and breath-hold single-shot fast spin-echo (Az = 0.94) images showed the best performance of the four imaging techniques. The diagnostic accuracy for detection of nonsolid lesions with multishot spin-echo echoplanar images (Az = 0.88) was statistically significantly lower than with respiratory-triggered fast spin-echo (p < 0.005) and breath-hold single-shot fast spin-echo (p < 0.005) images (Fig. 5A,5B,5C,5D).
|
|
|
|
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In terms of single-shot fast spin-echo imaging, Coates et al. [18] reported that "double shot" half-Fourier single-shot turbo spin-echo imaging, which is also an implemented version of half-Fourier RARE sequences as is single-shot fast spin-echo imaging, showed a reduced lesion-to-liver contrast-to-noise ratio for solid lesions compared with conventional spin-echo and fast spin-echo techniques. They used an echo space of 8.2 msec and an echo train length of 36. Tang et al. [10] reported that fast half-Fourier single-shot turbo spin-echo imaging, which was achieved with the use of a short echo space of 4.2 msec, a rapid sampling rate, and a lower echo train length of 64, showed improved image quality and soft-tissue contrast resolution compared with conventional half-Fourier single-shot turbo spin-echo imaging, which had an echo space of 10.9 msec and an echo train length of 120. Our results show that single-shot fast spin-echo sequences with an echo space of 5.0 msec have the highest signal-to-noise ratio of the liver among the four T2-weighted sequences and have no significantly inferior lesion-to-liver contrast-to-noise ratio compared with the other sequences. These results are compatible with those of previous reports [10, 14] that evaluated the fast half-Fourier single-shot turbo spin-echo sequence with an echo space of 4.2 msec. The reason for the high signal-to-noise ratio of the liver is probably that the effect of motion-related artifacts was extremely low. Tang et al. concluded that fast half-Fourier single-shot turbo spin-echo imaging resulted in a similar diagnostic performance to that obtained with respiratory-triggered fast spin-echo imaging, although they did not show separate results of receiver operating characteristic analyses for the detection of solid or nonsolid lesions. Our separate receiver operating characteristic analysis data for solid and nonsolid lesions show that the diagnostic capability of single-shot fast spin-echo imaging for solid lesion detection was poor. The magnetization transfer contrast effect with the RARE sequences decreases the signal intensity of solid lesions [19]. As a result, this effect can decrease the diagnostic capability for solid lesion detection. However, no significant differences in lesion-to-liver contrast-to-noise ratios among the four sequences were seen in our study. We think that the magnetization transfer contrast effect cannot account sufficiently for the inferior diagnostic capability of single-shot fast spin-echo imaging for solid lesion detection in our study. One of the reasons for this is thought to be image blurring due to T2-filtering effects [18, 20,21,22]. An effective point spread function for fast spin-echo imaging is different for each tissue—that is, broader for tissues with a short T2 than for those with a long T2. A broader effective point spread function leads to more image blurring. Therefore, the signal from small lesions with a short T2 can easily be lost through this blurring mechanism [21]. Some of small lesions may be missed because of this effect, especially solid lesions, which have a shorter T2 than do nonsolid lesions. For the detection of nonsolid lesions, single-shot fast spin-echo sequences show the best diagnostic accuracy of the four sequences. The higher lesion-to-liver contrast-to-noise ratios and reduced T2-filtering effects compared with those of solid lesions may account for these results.
Some reports [15, 23] have shown a better diagnostic capacity for multishot spin-echo echoplanar imaging than for non—breath-hold T2-weighted MR imaging. However, these reports did not evaluate the diagnostic accuracy of the technique for the detection of hepatic lesions by means of receiver operating characteristic analysis. Kanematsu et al. [11] studied the diagnostic accuracy of various T2-weighted MR imaging techniques, including conventional spin-echo, respiratory-triggered fast spin-echo, breath-hold fast spin-echo, and multishot spin-echo echoplanar imaging. Using receiver operating characteristic analysis, they concluded that the diagnostic accuracy of multishot echoplanar imaging was inferior to that of conventional spin-echo imaging for the detection of solid hepatic lesions and to that of respiratory-triggered fast spin-echo imaging for nonsolid lesions. These results do not agree with ours, however. Our results show that, in terms of solid lesion detection, the diagnostic accuracy of multishot spin-echo echoplanar imaging is not statistically significantly inferior to that of either conventional spin-echo or respiratory-triggered fast spin-echo imaging. We think one of the reasons for this discrepancy is that the imaging parameters for multishot spin-echo echoplanar imaging used in this study, such as matrix size and receiver bandwidth, are different from those used in the study of Kanematsu et al. To reduce susceptibility artifacts [24], we obtained multishot spin-echo echoplanar images with a comparatively smaller number of phase-encoding steps (128) compared with 256 steps in the study of Kanematsu et al. Moreover, we used a quadrature surface coil for reception instead of a phased array coil. Receiver bandwidth can be made as wide as ±130 kHz with the surface coil, compared with ±62 kHz used by Kanematsu et al. Susceptibility artifacts were suppressed with such wide receiver bandwidth, although the signal-to-noise ratio of the images was reduced. Another reason is thought to be more experience of observers in evaluating multishot spin-echo echoplanar images. In the study of Kanematsu et al. [11], they mentioned that lack of familiarity could result in less accuracy. In our study, however, the blinded observers were accustomed to interpreting echoplanar images because they had been interpreting MR images, including multishot echoplanar sequences, for more than a year. These differences may account for the incompatibility of the results of the two studies.
Although no significant differences were seen among the lesion-to-liver contrast-to-noise ratios for nonsolid lesions, the diagnostic accuracy with multishot spin-echo echoplanar images was statistically significantly lower than with respiratory-triggered fast spin-echo and breath-hold single-shot fast spin-echo images. The inferior image quality of multishot spin-echo echoplanar images caused by artifacts may account for the results.
Our study did not evaluate the diagnostic ability of single-shot spin-echo echoplanar imaging, although several authors [7, 14] have reported on this capability for the detection of focal hepatic lesions. Their reports did not show promising results for single-shot spin-echo echoplanar imaging. One of the most serious problems in echoplanar imaging is susceptibility artifacts. With the single-shot echoplanar technique, any phase error that arises from a variation in the resonating frequencies of protons tends to propagate through the entire K-space. In multishot echoplanar imaging, however, the readout is divided into multiple segments, so that phase errors have less time to build up than with single-shot echoplanar imaging. Consequently, a reduction in magnetic susceptibility artifacts can be achieved. In our preliminary experience, we observed major susceptibility artifacts with single-shot spin-echo echoplanar imaging, especially at the dome of the liver. On the other hand, no such significant image distortion or signal deterioration caused by susceptibility artifacts was observed with multishot spin-echo echoplanar imaging. Although the imaging time is prolonged from several seconds to several tens of seconds, high-quality single breath-hold images of the entire liver can be obtained with the multishot echoplanar technique.
Although using the fat-suppression technique has some advantages, we did not use this technique for conventional spin-echo, respiratory-triggered fast spin-echo, or single-shot fast spin-echo imaging because, with the use of this technique, ambiguity of organ borders occurs; fat-containing lesions may be obscured; and potential overall image quality is degraded because of incomplete suppression of fat [11]. However, the fat-suppression technique increases the signal-to-noise ratio and dynamic range of the image and reduces motion artifacts from fat of the anterior body wall [25, 26]. Moreover, a report has been published that shows superior diagnostic capability of fat-suppressed respiratory-triggered fast spin-echo sequences compared with multishot echoplanar sequences for the detection of both solid and nonsolid hepatic lesions [27]. But severe image distortion caused by susceptibility artifacts was seen in their echoplanar images, as in the figures of the report [27]. We think that further studies comparing the diagnostic ability of our improved echoplanar sequences with that of fat-suppressed respiratory-triggered fast spin-echo sequences are needed.
For most receiver operating characteristic studies to date, the observers used a discrete rating scale with five or six categories. However, this scale has several disadvantages [28]: it is more likely to yield degenerate data sets [17], it entails extensive and specific observer training, and it is only indirectly related to clinical reporting. These disadvantages are likely to be eliminated or substantially reduced by the continuous rating scale [28]. Rockette et al. [28] compared the results of two receiver operating characteristic analyses, which were obtained separately with both the discrete and continuous rating scales, for CT images for the detection of abdominal masses. They reported no significant differences between the Az values evaluated with the two scales and recommended continuous rating scales for routine use in radiologic receiver operating characteristic studies because of their potential advantages. For this study, we therefore used a continuous rating scale, as was done in another study [29].
In conclusion, breath-hold single-shot fast spin-echo and multishot spin-echo echoplanar imaging can be used as an alternative to either conventional spin-echo or respiratory-triggered fast spin-echo sequences with the MR parameters used in this study. Of the four T2-weighted MR imaging sequences, single-shot fast spin-echo imaging produced the best diagnostic performance for detection of non-solid lesions, and multishot spin-echo echoplanar imaging resulted in almost the same performance for the detection of solid lesions as respiratory-triggered fast spin-echo imaging, which was rated the best. Thus, by obtaining both types of single breath-hold T2-weighted images, we can reduce the examination time significantly without reducing the diagnostic capability for the detection of both solid and nonsolid lesions.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K Coenegrachts, J Delanote, L Ter Beek, M Haspeslagh, S Bipat, J Stoker, F Van Kerkhove, L Steyaert, H Rigauts, and J W Casselman Improved focal liver lesion detection: comparison of single-shot diffusion-weighted echoplanar and single-shot T2 weighted turbo spin echo techniques Br. J. Radiol., July 1, 2007; 80(955): 524 - 531. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. J. Pijl, J. Doornbos, M. N. J. M. Wasser, H. C. van Houwelingen, R. A. E. M. Tollenaar, and J. L. Bloem Quantitative Analysis of Focal Masses at MR Imaging: A Plea for Standardization Radiology, June 1, 2004; 231(3): 737 - 744. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Augui, O. Vignaux, C. Argaud, J. Coste, H. Gouya, and P. Legmann Liver: T2-weighted MR Imaging with Breath-hold Fast-Recovery Optimized Fast Spin-Echo Compared with Breath-hold Half-Fourier and Non-Breath-hold Respiratory-triggered Fast Spin-Echo Pulse Sequences Radiology, June 1, 2002; 223(3): 853 - 859. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Masui, M. Katayama, S. Kobayashi, S. Nakayama, A. Nozaki, H. Kabasawa, T. Ito, and H. Sakahara Changes in Myometrial and Junctional Zone Thickness and Signal Intensity: Demonstration with Kinematic T2-weighted MR Imaging Radiology, October 1, 2001; 221(1): 75 - 85. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Katayama, T. Masui, S. Kobayashi, T. Ito, H. Sakahara, A. Nozaki, and H. Kabasawa Evaluation of Pelvic Adhesions Using Multiphase and Multislice MR Imaging with Kinematic Display Am. J. Roentgenol., July 1, 2001; 177(1): 107 - 110. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||
) (Az = 0.91),
breath-hold single-shot fast spin-echo (
) (Az =
0.84), and breath-hold multishot spin-echo echoplanar (
)
(Az = 0.90) MR images. Note that no significant
differences are seen among diagnostic accuracy of conventional spin-echo,
respiratory-triggered fast spin-echo, and multishot spin-echo echoplanar
imaging.
