| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, Gifu University School of Medicine, 40 Tsukasamachi, Gifu 500-8705, Japan (M.K., H.H., H.K., R.Y.); the Department of Radiology, Kyoto University Faculty of Medicine, Japan (K.I.); and the Department of Radiology, Osaka University Medical School, Japan (T.M., M.H., H.N.). Received March 10, 1998; revision requested May 7; final revision received August 6; accepted October 26. Address reprint requests to M.K.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Fat-suppressed T2-weighted MR images obtained with the four sequences in 55 patients with 81 solid and 129 nonsolid lesions were retrospectively analyzed. Image review was conducted on a segment-by-segment basis; a total of 440 liver segments were reviewed separately for solid and nonsolid lesions by three independent radiologists. Diagnostic accuracy was evaluated with receiver operating characteristic analysis.
RESULTS: The mean lesion-to-liver contrast-to-noise ratio was highest on the multishot SE echo-planar images of both solid and nonsolid lesions. Fat-suppressed respiratory-triggered fast SE images had significantly better (P < .05) or comparative detectability of both solid and nonsolid lesions compared with the other types of images. Image quality was best on the respiratory-triggered fast SE images.
CONCLUSION: Fat-suppressed respiratory-triggered fast SE imaging should replace fat-suppressed conventional SE imaging as a standard T2-weighted imaging examination in the detection of focal hepatic lesions.
Index terms: Angioma, gastrointestinal tract, 761.3194 • Liver neoplasms, 761.312, 761.3194, 761.323, 761.332 • Liver neoplasms, MR, 761.121411, 761.121414, 761.121415, 761.121416 • Magnetic resonance (MR), comparative studies, 761.121411, 761.121414, 761.121415, 761.121416 • Magnetic resonance (MR), pulse sequences, 761.121411, 761.121416
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Previous researchers (7–13,17) have described the use of a fat suppression technique for T2-weighted MR imaging to increase lesion detectability. Gaa et al (11) reported that the sensitivity in the detection of hepatic lesions was greatest with inversion-recovery fast SE imaging, followed by fat-suppressed fast SE imaging, inversion-recovery SE echo-planar imaging, and conventional SE imaging, although nearly half of the lesions evaluated by them were nonsolid lesions.
When a RARE sequence (ie, fast or turbo SE) is used in T2-weighted MR imaging, the signal intensity of fat is strengthened owing to the J-coupling effect and magnetization transfer contrast effect (18,19) and thereby leads to the increased signal intensity of the normal liver parenchyma where lipids physiologically exist (20). This may result in decreased contrast between the liver parenchyma and focal hepatic tumors, where the T2 relaxation time is usually longer than that of the nontumorous liver parenchyma. In contrast, the use of a fat suppression technique may increase the lesion-to-liver contrast by decreasing the signal intensity of the normal liver parenchyma.
The chemical shift–selective fat suppression method is one of the most common techniques available for fat suppression MR imaging (21). However, because of the inhomogeneity of the static magnetization and radio-frequency excitation field, fat suppression is not always achieved effectively. The magnetization transfer contrast effect from successive non–section-selective fat saturation pulses may darken the entire image. More practically, incomplete fat suppression in areas close to the body surface coil can degrade the image quality, and the definition of the margins of extrahepatic organs such as the adrenal gland or pancreas is suboptimal with fat suppression images (8). The net advantage of using fat suppression in the detection and characterization of focal hepatic lesions is still controversial.
We (15) previously found that the diagnostic accuracy in the detection of solid malignant lesions was statistically significantly better with non–fat-suppressed conventional SE MR imaging sequences than with non–fat-suppressed respiratory-triggered fast SE (P < .05), breath-hold fast SE (P < .001), and breath-hold multishot SE echo-planar (P < .01) sequences, although it is largely believed that fast SE imaging can replace conventional SE in the detection of focal hepatic lesions. However, to our knowledge, it has not been determined whether the addition of fat suppression to the respiratory-triggered fast SE imaging technique results in an improved diagnostic accuracy that exceeds the accuracy with conventional SE.
The purpose of our study was to compare the diagnostic accuracy in the detection of focal hepatic lesions depicted on T2-weighted MR images obtained with a conventional SE sequence, respiratory-triggered fast SE sequence, and breathhold fast SE sequence with a chemical shift–selective fat suppression technique by means of quantitative and receiver operating characteristic (ROC) analyses. We also correlated these pulse sequences with a breath-hold multishot SE echo-planar sequence obtained with a chemical shift–selective water excitation technique that was an alternative fat suppression technique.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Of the 55 patients, seven with HCC, four with colorectal cancer metastases, two with cholangiocarcinoma, and one with gallbladder carcinoma subsequently underwent definitive surgery with intraoperative US within 2 weeks. Seven patients with HCC and eight with metastases (three with colon carcinoma, one with gallbladder carcinoma, one with pancreatic carcinoma, one with uterine cervical carcinoma, one with pulmonary carcinoma, and one with intestinal leiomyosarcoma) underwent US-guided needle-core aspiration biopsy of at least one liver tumor for histologic proof of malignancy within 2 weeks of US, CT, or MR imaging; in these patients, other tumors with imaging findings similar to those of the histologically evaluated lesions were considered to be of the same disease. In three patients with cirrhosis in whom histopathologic evidence of disease was not obtained, the visible hepatic tumors were considered to be HCC on the basis of tumor growth during 3–5 months (mean, 4.2 months) at follow-up US, CT, and MR imaging and on the basis of increased levels of serum 
-fetoprotein and protein induced in vitamin K absence (PIVKA)-II.
Twelve patients had cavernous hemangiomas, and 24 patients had liver cysts. These benign diseases were diagnosed on the basis of pathognomonic findings at US, contrast material–enhanced CT, and dynamic gadolinium-enhanced gradient-recalled-echo MR imaging. Further confirmation of all benign lesions was obtained with findings at follow-up US, contrast-enhanced CT, and dynamic MR imaging.
The presence or absence of liver lesions was determined by consensus opinion between three radiologists (M.K., H.K., H.H.) on the basis of findings at US, CT, T1- and T2-weighted MR imaging, and dynamic MR imaging; follow-up US, CT, and MR imaging; serologic examination; biopsy; and definitive surgery. Thus, in the 55 patients, confirmation was obtained of 31 HCC lesions (diameter, 12–80 mm; mean, 26.5 mm), 45 metastatic lesions (diameter, 5-70 mm; mean, 21.9 mm), five cholangiocarcinoma lesions (diameter, 5–115 mm; mean, 47.4 mm), 30 cavernous hemangiomas (diameter, 3–35 mm; mean, 12.3 mm), and 99 liver cysts (diameter, 3–80 mm; mean, 10.0 mm).
MR Imaging Techniques
MR imaging was performed with a superconducting imager at 1.5 T with blipped echo-planar capability (Signa Horizon; GE Medical Systems, Milwaukee, Wis). 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 with a phased-array multicoil for the body. The section thickness was 8 mm, with a 2-mm intersection gap for all pulse sequences. The imaging protocol consisted of T1-weighted SE imaging (repetition time [TR] msec/echo time [TE] msec, 500/9; two signals acquired) and T2-weighted imaging with (a) a fat-suppressed conventional SE sequence (2,000/80, two signals acquired, 256 x 192 matrix, 16-kHz receiver bandwidth, 32 x 32-cm or 29 x 29-cm field of view, ordered-phase encoding, gradient moment nulling in the frequency-encoding direction, 13.8-minute acquisition time); (b) a fat-suppressed respiratory-triggered fast SE sequence (3,333–8,571/77–88 [effective TR/effective TE], echo train length of 8–18, four signals acquired, 512 x 256 matrix, 62.5-kHz receiver bandwidth, 32 x 24-cm or 29 x 22-cm field of view, 20% respiratory trigger point, 40% trigger window, gradient moment nulling in the frequency-encoding direction, 4.4–6.2-minute acquisition time); (c) a fat-suppressed breath-hold fast SE sequence (2,000–2,100/80–81 [TR/effective TE], echo train length of 18, one signal acquired, 256 x 224 matrix, 62.5-kHz receiver bandwidth, 32 x 24-cm or 29 x 22-cm field of view, gradient moment nulling in the frequency-encoding direction, acquisition time of nine locations per 25 seconds, resulting in two data acquisitions to cover the entire liver in all patients); and (d) a breath-hold multishot SE echo-planar sequence (2,400-3,692/60 [effective TR/effective TE], eight shots, one signal acquired, 256 x 256 matrix, 62-kHz receiver bandwidth, 32 x 24-cm or 29 x 22-cm field of view; 50%–65% peripheral gating trigger point, 10%–20% trigger window, gradient moment nulling in the frequency-encoding direction, acquisition time of 12–18 locations per 18–30 seconds).
The chemical shift–selective fat suppression technique was used for conventional SE and fast SE imaging, and the chemical shift–selective water excitation technique was used for echo-planar imaging. Automated shimming was always used before data acquisition. We subsequently performed triphasic dynamic contrast-enhanced MR imaging in all patients. With all MR imaging, spatial presaturation pulses were used superiorly and inferiorly to the imaging volume.
Quantitative Image 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 of the liver, spleen, hepatic lesions, and background noise. The signal intensities of the liver and spleen were measured in areas devoid of focal changes in signal intensity, large vessels, and prominent artifacts. To minimize the difference in signal intensity due to the near-field effect when surface coils are used, the region of interest in the liver and spleen were located so that the vertical distances from the ventral side surface coils to the region of interest were the same. For measurement purposes, a maximum of five large lesions in each patient that were seen on images obtained with all pulse sequences were chosen for analysis. Consequently, 43 solid lesions (25 HCCs, 14 metastases, and four cholangiocarcinomas) and 42 nonsolid lesions (15 cavernous hemangiomas and 27 cysts) 10 mm in diameter or larger were studied. In the liver lesions, a circular region of interest was drawn to encompass as much of the lesion as possible. When the hepatic lesion was too small to draw a region of interest, the image was magnified up to three times. Whenever possible, at least 35-mm2 regions of interest were used.
The SD of background noise, SDB, was measured in the phase-encoding direction outside the anterior abdominal wall to calculate the following: liver signal-to-noise ratio (SNR) = SIliver/SDB, liver-to-spleen contrast-to-noise ratio (CNR) = (SIspleen - SIliver)/SDB, and lesion-to-liver CNR = (SIlesion - SIliver)/SDB, where SIliver, SIspleen, and SIlesion are the signal intensities of the liver, spleen, and lesion, respectively.
Qualitative Image Analysis
The review procedure was performed in the same manner as that in our previous study (15). Three readers (K.I., T.M., M.H.), who have served mainly as gastrointestinal radiologists for 6–16 years and interpreted MR images of the liver as part of their daily clinical and research practice, were invited from other institutions to the image review. They knew that the patients were referred for assessment of a possible liver tumor but did not know any other information about the patients' histories.
The image review was conducted on a segment-by-segment basis. To prevent mislocation of the lesions by the readers, the hepatic segment numbering system of Couinaud (22) was drawn on the images by the study coordinator (M.K.). A total of 440 liver segments were reviewed, including 77 segments with 81 solid lesions (31 HCCs, 45 metastases, and five cholangiocarcinomas) and 111 segments with 129 nonsolid lesions (30 cavernous hemangiomas and 99 cysts). The review procedure was performed in four separate sessions. The images were reviewed in alphabetic order according to patient names, but the order in which the images obtained with the four pulse sequences were reviewed was randomized. To minimize learning bias, the name, age, identification number, and imaging parameters of each patient were masked, and the four reviewing sessions were performed in 2-week intervals.
For each pulse sequence, the readers recorded the size and site (ie, Couinaud segment) of visible abnormalities and indicated for each segment whether the presence of solid or nonsolid lesions separately could be ascertained. The readers assigned one of five confidence levels as follows: 1, definitely absent; 2, probably absent; 3, equivocal; 4, probably present; and 5, definitely present. When a lesion was located in two or more segments, the reader was asked to consider only the segment that was mainly involved and to assess the probability of another lesion in the other segments. When a gallbladder tumor directly invading the hepatic parenchyma was seen, the reader was asked to disregard the lesion and assess the probability of the presence of other separate lesions in all the segments. The readers were instructed to indicate a score of 1 when no focal signal intensity change was seen; of 3 when the signal intensity change was subtle, ill-defined, and not circular or oval; and of 5 when the signal intensity change was discrete, well-circumscribed, and circular or oval. Scores of 2 and 4 were assigned on the basis of the reader's subjective judgment. Because readers simultaneously evaluated each hepatic segment for the presence of solid and nonsolid lesions at each reading, two confidence levels were assigned for each individual hepatic segment.
Furthermore, each reader evaluated the degree of image quality degradation due to motion, ghosting, susceptibility artifacts, and imperfect fat suppression by using the following five-point scale: 1, severe; 2, moderate; 3, mild; 4, minimal; and 5, absent. A score of "severe" was assigned when the image could not be interpreted because of an artifact; a score of "mild" was assigned when the artifact was present but did not markedly preclude interpretation. Scores of "moderate" and "minimal" were assigned according to the reader's subjective judgment.
Statistical Analyses
The Wilcoxon sign-rank test was used to compare the liver SNR, liver-to-spleen CNR, and lesion-to-liver CNR on images obtained with the four pulse sequences. For each pulse sequence, a binomial ROC curve was fitted to each reader's confidence rating by using a maximum-likelihood estimation (23). The diagnostic accuracy of each pulse sequence determined by each reader was estimated by calculating the area under the ROC curve (Az) (24). Composite ROC curves that combined the performance of all readers into a single curve were obtained for each pulse sequence by using the maximum-likelihood curve-fitting algorithm to rate the pooled data of the three independent readers (25).
The relative sensitivities of each pulse sequence for the detection of solid and nonsolid lesions by three individual readers and composite data were determined by using the number of segments that were assigned a score of 3 or greater (ie, equivocal to definitely present) of a total of 77 and 111 segments that harbored solid and nonsolid lesions, respectively. The score of 3 (equivocal) was regarded as being indicative of the presence of lesions in a given segment to determine the relative sensitivity, because the segments scored 3 were often found to have true lesions after definitive surgery, biopsy, or follow-up evaluation. These relative sensitivities were compared among the four pulse sequences by using the McNemar test.
To assess interobserver variability in interpreting images, the 
statistic was used to measure the degree of agreement between each combination of two observers. We used the nonweighted statistic, with binary data defined in terms of the less-than-50% cutoff level. The degree of disagreement was not factored into the calculation. A value of up to 0.40 indicated positive but poor agreement, a value of 0.41–0.75 indicated good agreement, and a value of greater than 0.75 indicated excellent agreement.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
values determined by the three readers, which were calculated on the basis of each reader's confidence levels for the ROC analysis, are summarized in Table 5. Good to excellent agreement was obtained among the readers with regard to the presence of lesions in a given segment. The total degree of image degradation caused by artifacts and imperfect fat suppression was highest with the breath-hold multishot SE echo-planar sequence and lowest with the respiratory-triggered fast SE sequence (Table 6).
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
There are some explanations for the improved diagnostic accuracy of respiratory-triggered fast SE imaging, especially for the detection of solid malignant lesions, with the fat suppression technique. First, the liver parenchyma turned darker with fat suppression than without fat suppression (26), because the physiologically existing lipid in the liver was suppressed, and this led to better tumor conspicuity. Second, fast SE imaging with a phased-array multicoil produces prominent high signal intensity of the abdominal wall adjacent to the coils because of the high signal intensity of fat and the near-field effect; this causes serious ghosting artifacts over the liver, which may obscure the focal hepatic lesions. Contrarily, the use of a chemical shift–selective fat suppression pulse could decrease such ghosting artifacts, and it might have led to improved detectability of hepatic lesions with respiratory-triggered fast SE imaging in the present study (9,12,26) (Figs 1, 2). The results of our assessment of image degradation supported this postulation (Table 6).
The magnetization transfer contrast effect with the RARE sequence may be more enhanced by adding the chemical shift–selective fat suppression technique. The magnetization transfer contrast effect decreases the signal intensity of solid malignant lesions but has little or no effect on the signal intensity of nonsolid benign lesions, in which there are few protons in macromolecules and their associated "bound" water (27–29). This may be a reason why the fat-suppressed respiratory-triggered fast SE sequence was best for the detection of nonsolid benign lesions (Fig 2). Another reason for the superior detectability of small-diameter nonsolid lesions with respiratory-triggered fast SE imaging was that the sequence enabled the use of a larger matrix size (512 x 256) and a larger number of signals acquired (four times) compared with the other sequences.
In the previous (15) and present studies, we found that breath-hold fast SE imaging by using a standard RARE sequence with a 16-18 echo train length and full k-space filling technique was not very accurate for the detection of solid and nonsolid lesions, regardless of the use of chemical shift–selective fat suppression. We conclude that breath-hold fast SE imaging with a standard RARE sequence is of no value for further investigation. Tang et al (14) concluded that breath-hold T2-weighted images obtained by using a faster single-shot RARE sequence with a longer (up to 64) echo train with shorter interecho space and a partial k-space filling (half-Fourier) technique were free of motion artifacts, and the observer diagnostic performance was similar to that with a respiratory-triggered RARE sequence that is analogous to the respiratory-triggered fast SE sequence used in our study. Respiratory-triggered fast SE and breath-hold fast SE imaging with use of such a faster RARE sequence should be compared further.
There are some limitations to this study. Only 14 patients underwent definitive surgery with intraoperative US; the remaining 18 patients with malignant hepatic tumors did not. Fourteen patients underwent biopsy of at least one liver tumor for histologic proof and underwent imaging within 3–6 months for confirmation that all other lesions were malignant as well from their growth. Some HCC lesions undergo fatty metamorphosis (30), and the use of a fat suppression technique may obscure such lesions; however, our study population did not include such an HCC.
In conclusion, among the fat-suppressed conventional and fast SE sequences and water excitation echo-planar sequences currently available in clinical practice, T2-weighted MR imaging with a respiratory-triggered fast SE sequence resulted in the greatest accuracy in the detection of both solid malignant and nonsolid benign hepatic lesions. Echo-planar imaging resulted in the best CNR and shortest acquisition time, although the accuracy in lesion detection with this sequence was inferior to that with respiratory-triggered fast SE imaging owing to the degraded image quality caused by artifacts. The quality of echo-planar imaging must be improved by reducing the artifacts so that operators can take advantage of the shorter acquisition time and increased contrast resolution. At present, however, we believe that fat-suppressed respiratory-triggered fast SE sequences should replace conventional SE sequences as standard pulse sequences for T2-weighted imaging in the detection of focal hepatic lesions.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantors of integrity of entire study, H.H., H.N.; study concepts and design, M.K.; definition of intellectual content, M.K.; literature research, M.K.; clinical studies, M.K., H.K.; experimental studies, K.I., T.M., M.H.; data acquisition, M.K., H.K., R.Y.; data and statistical analyses, M.K.; manuscript preparation and editing, M.K.; manuscript review, M.K., H.H., H.N.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. Hirokawa, H. Isoda, Y. S. Maetani, S. Arizono, K. Shimada, T. Okada, T. Shibata, and K. Togashi Hepatic Lesions: Improved Image Quality and Detection with the Periodically Rotated Overlapping Parallel Lines with Enhanced Reconstruction Technique--Evaluation of SPIO-enhanced T2-weighted MR Images Radiology, May 1, 2009; 251(2): 388 - 397. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Bruegel, J. Gaa, S. Waldt, K. Woertler, K. Holzapfel, B. Kiefer, and E. J. Rummeny Diagnosis of Hepatic Metastasis: Comparison of Respiration-Triggered Diffusion-Weighted Echo-Planar MRI and Five T2-Weighted Turbo Spin-Echo Sequences Am. J. Roentgenol., November 1, 2008; 191(5): 1421 - 1429. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. S. Kim, J. H. Kim, G. M. Choi, S. H. Kim, J. K. Park, B.-C. Song, and W. Kang Comparison of Three Free-Breathing T2-Weighted MRI Sequences in the Evaluation of Focal Liver Lesions Am. J. Roentgenol., January 1, 2008; 190(1): W19 - W27. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Huang, S. S. Raman, N. Vuong, J. W. Sayre, and D. S. K. Lu Utility of Breath-Hold Fast-Recovery Fast Spin-Echo T2 Versus Respiratory-Triggered Fast Spin-Echo T2 in Clinical Hepatic Imaging Am. J. Roentgenol., March 1, 2005; 184(3): 842 - 846. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kanematsu, S. Goshima, H. K. Hussain, T. D. Johnson, H. V. Nghiem, and I. R. Francis Does T2-weighted MR Imaging Really Add No Value in Detection and Characterization of Focal Lesions in Cirrhotic Liver? * Dr Hussain and colleagues respond: Radiology, February 1, 2005; 234(2): 638 - 640. [Full Text]
|
|
|
|
|
|
I.-M. Danet, R. C. Semelka, P. Leonardou, L. Braga, G. Vaidean, J. T. Woosley, and M. Kanematsu Spectrum of MRI Appearances of Untreated Metastases of the Liver Am. J. Roentgenol., September 1, 2003; 181(3): 809 - 817. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. K. Kim, S. H. Kim, W. J. Lee, H. Kim, J. W. Seo, D. Choi, H. K. Lim, S. J. Lee, and J. H. Lim Preoperative Detection of Hepatocellular Carcinoma: Ferumoxides-Enhanced Versus Mangafodipir Trisodium--Enhanced MR Imaging Am. J. Roentgenol., September 1, 2002; 179(3): 741 - 750. [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]
|
|
|
|
|
|
M. T. Keogan and R. R. Edelman Technologic Advances in Abdominal MR Imaging Radiology, August 1, 2001; 220(2): 310 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Choi, S. H. Kim, J. H. Lim, W. J. Lee, H.-J. Jang, S. J. Lee, and H. K. Lim Preoperative Detection of Hepatocellular Carcinoma: Ferumoxides-Enhanced MR Imaging Versus Combined Helical CT During Arterial Portography and CT Hepatic Arteriography Am. J. Roentgenol., February 1, 2001; 176(2): 475 - 482. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. I. Goldberg and A. R. Margulis Gastrointestinal Radiology in the United States: An Overview of the Past 50 Years Radiology, July 1, 2000; 216(1): 1 - 7. [Abstract] [Full Text]
|
|
|
|
|
|
M. Hori, T. Murakami, T. Kim, M. Kanematsu, K. Tsuda, S. Takahashi, M. Takamura, H. Hoshi, and H. Nakamura Single Breath-Hold T2-Weighted MR Imaging of the Liver: Value of Single-Shot Fast Spin-Echo and Multishot Spin-Echo Echoplanar Imaging Am. J. Roentgenol., May 1, 2000; 174(5): 1423 - 1431. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kanematsu, H. Hoshi, M. N. J. M. Wasser, M. E. J. Pijl, E. L. van Persijn van Meerten, J.-W. C. Gratama, and J. L. Bloem Can Inversion-Recovery Gradient- and Spin-Echo T2-weighted MR Imaging Be an Alternative to Fast Spin-Echo Imaging with Fat Suppression? • Dr Wasser and colleagues respond: Radiology, July 1, 1999; 212(1): 291 - 291. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
) (Az = 0.88), fat-suppressed respiratory-triggered fast SE (
) (Az = 0.90), fat-suppressed breath-hold fast SE (
) (Az = 0.84), and breath-hold multishot SE echo-planar (
) (Az = 0.86) MR images. (b) Composite ROC curves generated from the pooled data of the three independent readers for the detection of nonsolid lesions. The curves display the observers' confidence in the detection of nonsolid lesions on fat-suppressed conventional SE (