| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 All authors: Department of Radiology, Duke University Medical Center, Duke North, Rm. 1417, Erwin Rd., Durham, NC 27710.
Received September 20, 2004;
accepted after revision January 31, 2005.
Address correspondence to E. M. Merkle
(elmar.merkle{at}duke.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. A standardized MR cholangiography sequence protocol was used for 15 healthy male volunteers (mean age ± SD, 32.4 ± 4.3 years) who underwent both 1.5- and 3.0-T MRI within 2 hr in an alternating fashion. Dedicated circular polarized torso coils (1.5 and 3.0 T) were used. The sequence protocol included breath-hold single-slice rapid acquisition with relaxation enhancement (slice thickness, 50 mm; orientation, coronal and ± 20° oblique coronal); breath-hold multislice HASTE (slice thickness, 3 mm; coronal only); and a non-breath-hold, respiratory-triggered 3D turbo spin-echo (TSE) T2-weighted sequence (slice thickness, 1 mm; 60 slices per slab; coronal only). Maximum intensity projections were generated from each multislice data set. Bile duct (common bile duct, right posterior segmental branch, and left hepatic duct) to periductal tissue contrast-to-noise ratios were compared at 1.5 and 3.0 T. Qualitative image analysis was performed by three independent reviewers. Qualitative analysis included delineation of the extra- and intrahepatic biliary anatomy, with specific attention given to the presence (or absence) of cystic or intrahepatic ductal variants, using a 4-point confidence scale. Statistical analysis consisted of the paired Student's t test and the signed rank test.
RESULTS. Contrast-to-noise ratios between the bile duct and the periductal tissue were higher at 3.0 T in all three locations (common bile duct, right posterior segmental branch, and left hepatic duct). In each magnet class, the 3D TSE sequence offered the best contrast-to-noise ratio and qualitative analysis. Superiority of the 3D TSE sequence was statistically significant in all analyses. Five of the 15 volunteers had intrahepatic biliary variants that were detected with a higher level of confidence (p < 0.01) on the 3.0-T system than on the 1.5-T system.
CONCLUSION. Compared with MR cholangiography at 1.5 T, MR cholangiography at 3.0 T offers improved contrast-to-noise ratio and a higher level of confidence for depicting intrahepatic variants.
Keywords: biliary system • cholangiography • MRI • MR technique
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The introduction of whole-body 3.0-T MR systems in combination with dedicated circular polarized receive-only torso coils is an appealing concept with the potential to overcome these limitations. We attempted to determine the potential of MR cholangiography, including visualization of the intrahepatic biliary ductal system, at a field strength of 3.0 T compared with the standard field strength of 1.5 T.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MR Cholangiography Protocol
The first sequences (sequences 1-3) were a single-slice T2-weighted rapid
acquisition with relaxation enhancement (RARE) with fat saturation, coronal
and ± 20° oblique coronal orientation. The parameters for the 1.5-T
sequence were a TR/TE of 2,800/1,100; flip angle, 180°; number of signal
averages, 1; matrix, 256 x 256; slice thickness, 50 mm; field of view,
30 x 30 cm2; and acquisition time (TA), 3 sec. The parameters
for the 3.0-T sequence were 2,120/964; flip angle, 150°; number of signal
averages, 1; matrix, 256 x 256; slice thickness, 50 mm; field of view,
30 x 30 cm2; and TA, 2 sec.
The fourth sequence was coronal T2-weighted HASTE with fat saturation. The parameters for the 1.5-T sequence were 1,100/95; flip angle, 180°; 1 signal average; matrix, 192 x 256; 13 slices; slice thickness, 3 mm; field of view, 30 x 30 cm2; and TA, 14 sec; and for the 3.0-T sequence were 3,000/106; flip angle, 123°; 1 signal average; matrix, 256 x 256; 16 slices; slice thickness, 3 mm; field of view, 35 x 35 cm2; 2 concatenations; and TA, 2 x 24 sec.
The fifth sequence was respiratory triggered coronal 3D T2-weighted turbo spin echo (TSE) with fat saturation using a navigator technique for detection of the diaphragm position. The parameters for the 1.5-T sequence were TR/TE, 1 x respiratory cycle/648; flip angle, 180°; 1 signal average; matrix, 222 x 256; 60 slices; slice thickness, 1 mm; field of view, 30 x 30 cm2; and TA, 3-6 min, depending on the respiratory frequency; and for the 3.0-T sequence were 1 x respiratory cycle/645; flip angle, 180°; 1 signal average; matrix, 240 x 256; 60 slices; slice thickness, 1 mm; field of view, 30 x 30 cm2; and TA, 3-6 min, depending on the respiratory frequency.
Data Analysis
The initial step of the data analysis consisted of the selection of the
best single-slice RARE image (sequences 1-3) with the least superimposition
from the duodenum. This solitary image was used for the RARE portion of the
quantitative data analysis. In addition, maximum intensity projections (MIPs)
in an analogous orientation (either coronal or ± 20° oblique
coronal) were generated from each multislice data set (sequences 4 and 5,
HASTE and 3D TSE sequences) on a separate MRI workstation (Siemens Leonardo).
In all cases, these steps were performed by the same fellowship-trained senior
abdominal radiologist.
Quantitative Data Analysis
Signal amplitudes of the common bile duct, the right posterior segmental
branch, the left hepatic duct, and the periductal tissue were measured for
each MR cholangiography sequence using the MIP images only by defining regions
of interest (ROIs) on the MRI workstation. ROIs for the signal intensity of
bile were at least 5 mm2 and were chosen in homogeneous,
artifact-free areas of the common bile duct adjacent to the insertion of the
cystic duct, in the left hepatic duct next to the hepatic bifurcation, and in
the proximal right posterior segmental branch
(Fig. 1). ROIs for signal
intensity of the periductal tissue were at least 20 mm2 and were
chosen in homogeneous, artifact-free areas adjacent to the common bile duct at
the level of the cystic duct insertion or in the liver
(Fig. 1). SD of the noise was
also measured with ROIs of at least 200 mm2
(Fig. 1). Quantitative
measurements were again performed by the same fellowship-trained senior
abdominal radiologist in all cases.
|
Statistical analysis was performed using the paired Student's t test to test the null hypothesis that the CNR was identical at 3.0 and 1.5 T on version 8.2 of the SAS software system (SAS Institute). A p value of less than 0.05 was considered statistically significant.
Qualitative Data Analysis
Qualitative data analysis was performed by three fellowship-trained
abdominal radiologists. They evaluated the three RARE images (sequences 1-3,
data set 1), the source and MIP images from sequence 4 (data set 2), and the
source and MIP images from sequence 5 (data set 3). All 90 data sets were made
anonymous and were evaluated in a randomized order in two sessions and
separately by each reviewer. The evaluation protocol consisted of scoring the
following: clear depiction of the common bile duct (yes or no); clear
depiction of the cystic duct insertion (yes or no); clear depiction of the
right posterior segmental branch insertion (yes or no); presence (or absence)
of a biliary ductal variant using a 4-point confidence scale (0 = relevant
structures not seen; 1 = equivocal; 2 = probably seen; 3 = definitely seen);
presence (or absence) of a cystic duct variant using a 4-point confidence
scale (0 = relevant structures not seen; 1 = equivocal; 2 = probably seen; 3 =
definitely seen).
Classification of aberrant bile ducts and cystic ducts was based on the reports by Huang et al. [4] and Hirao et al. [5]. A consensus interpretation by three senior abdominal radiologists who were otherwise not involved in the qualitative data analysis served as the gold standard for the presence or absence of a ductal biliary variant. This consensus was based on all the imaging data sets available per volunteer. Statistical analysis was performed using the signed rank test to test the null hypothesis that the reviewer confidence scores were the same at 1.5 and 3.0 T on version 8.2 of the SAS software system (SAS Institute). A p value of less than 0.05 was considered statistically significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Within each field strength, the MIP from the 3D TSE sequence offered the best CNR in each biliary ductal location (Table 1). Superiority of 3D TSE was statistically significant in all analyses in all three biliary locations (common bile duct, left hepatic duct, and right posterior segmental branch). The RARE sequence offered the worst CNR in each biliary ductal location (Table 1). Inferiority was statistically significant in all but one analysis (CNR measured at the level of the common bile duct, HASTE vs RARE sequence, at 3.0 T: p = 0.12).
Qualitative Data Analysis
Delineation of common bile duct, cystic duct insertion, and right
posterior segmental branch insertion—Three reviewers independently
evaluated the delineation of the common bile duct, the cystic duct insertion,
and the right posterior segmental branch insertion in each MR cholangiography
sequence on both MR systems (Table
2). Although the delineation of the common bile duct and the
cystic duct insertion was highest for the 3D TSE T2-weighted sequence,
statistical analysis using the signed rank test did not reveal any significant
differences. Delineation of the right posterior segmental branch insertion
using the 3D TSE sequence, however, was significantly better for two of the
three reviewers on the 3.0-T scanner when compared with the 1.5-T MR scanner
(reviewer 1: p = 0.19; reviewer 2: p < 0.01; and reviewer
3: p < 0.01).
|
Presence (or absence) of ductal biliary variants—A consensus interpretation by three senior abdominal radiologists revealed intrahepatic biliary ductal variants in five of the 15 volunteers (trifurcation of the left hepatic duct, right posterior segmental duct, and right anterior segmental duct: n = 2; drainage of the right posterior segmental duct into the left hepatic duct: n = 2; and drainage of an aberrant right segmental branch into the cystic duct: n = 1).
The detection rate of the intrahepatic biliary ductal variants varied in the three independent reviewers. Sensitivity and specificity for each study reviewer and for the three reviewers combined are provided in Table 3. Overall, the imaging data sets acquired at the 3-T magnet provided better results in terms of sensitivity and specificity in the detection of intrahepatic biliary variants. Confidence levels for the detection of biliary ductal variants were also significantly higher for all three reviewers for the 3.0-T magnet than for the 1.5-T magnet for all sequences combined (reviewer 1: 1.73-2.31; reviewer 2: 1.6-1.96, and reviewer 3: 2.0-2.27). These differences were statistically significant (confidence level for biliary ductal variants at 1.5 T vs at 3.0 T, p < 0.01).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The introduction of 3.0-T MR systems combined with dedicated circular polarized receive-only torso coils offers an alternative solution to overcome the limitations of signal-to-noise ratio and spatial resolution. Although 3.0-T MR systems have already been shown to be advantageous for various indications in the brain compared with standard high-field 1.5-T MR systems, only a few 3.0-T MR studies of the abdomen have been published [14-16]. In particular, to our knowledge, no comparison data are yet available to evaluate the potential of 3.0-T MR systems compared with standard high-field 1.5-T MR systems for MR cholangiography.
Quantitative Data Analysis
CNRs between the biliary ductal system in three locations (common bile
duct, left hepatic duct, and right posterior segmental branch) and the
periductal tissue were higher at 3.0 T than at 1.5 T. These differences were
most evident when using the 3D TSE sequence (3D TSE at 1.5 T vs 3D TSE at 3.0
T: p < 0.01 in all three locations) but were also obvious when
using the HASTE and RARE sequences (Table
1). These results were expected because twice the number of free
water protons per voxel contribute to the MR signal at 3.0 T as at 1.5 T.
However, increasing the magnetic field strength results in an increase of the
T1 relaxation time and a slight decrease of the T2 relaxation time; these two
factors combine to reduce the net increase in MR signal to less than twofold
[14]. If twice the net MR
signal is sought, longer T1 relaxation times at 3.0 T need to be
compensated—for example, by increasing the sequence TR. This may explain
why the gain in signal using the breath-hold multislice HASTE sequence was
only marginal because of image cross-talk and not significantly higher despite
the fact that the TR had been adjusted from 1,000 msec at 1.5 T to 3,000 msec
at 3.0 T.
Potential solutions to avoid cross-talk are to acquire images in an interleaved order or with a small gap between contiguous images. However, both attempts may increase artifacts in MIPs and therefore have not been implemented. On the other hand, the prolonged T1 relaxation times at 3.0 T do not affect the signal gain when using the RARE and 3D TSE sequences because the former sequence is a single-shot, single-slice sequence and the latter sequence is respiratory triggered, with mean sequence TRs in the range of 5 sec, which allows a further recovery of the longitudinal magnetization.
Within each field strength, the MIPs from the 3D TSE sequence offered the best CNR between the biliary ductal structures and the periductal tissue (Table 1). This is in accordance with the results presented by Papanikolaou et al. [17]. In that study, using a 1.0-T MR system, the respiratory-triggered 3D TSE sequence also provided statistically higher CNRs than the breath-hold single-shot techniques [17]. The reason for this increased CNR is not necessarily obvious because the smaller voxel size of the TSE sequence would generally be associated with a lower CNR, but the respiratory triggering permits a dramatic increase in acquisition time that more than compensates for the smaller voxel size.
Qualitative Data Analysis
Statistical analysis did not reveal any significant differences between the
two MR systems regarding delineation of the common bile duct or delineation of
the cystic duct insertion. On both MR systems, the extrahepatic biliary duct
anatomy was seen similarly, with the 3D TSE sequence providing the best
results. The insertion of the right posterior segmental branch, however, was
significantly better delineated according to two of the three reviewers on the
3D TSE sequence on the 3.0-T MR system when compared with the 1.5-T system,
but also within each magnet class. This is of clinical importance because
right-to-left crossover variants have an incidence of approximately 30% and
can be detected with certainty only if the insertion of the right posterior
segmental branch is clearly visualized.
This finding contradicts the results reported by Papanikolaou et al. [17] using a 1.0-T MR system. In that study, single-shot breath-hold sequences were slightly superior to the respiratory-triggered 3D TSE sequence in the depiction of the intrahepatic biliary anatomy. A potential explanation may be differences in the analysis used in our study and the one performed by Papanikolaou et al. However, it is more likely that the higher magnetic field strengths of 1.5 and 3.0 T in our study, which offer a higher MR signal and a better signal-to-noise ratio from the beginning, are the main reason for the better depiction of the intrahepatic biliary ductal system. This increased MR signal seems to be particularly helpful in the 3D TSE sequence with an effective slice thickness of 1 mm.
Accurate preoperative radiologic imaging with a high level of confidence is essential to assess the vascular and biliary anatomy of living liver lobe donor candidates. MRI has the unique ability to delineate the biliary system without the need for bile duct-excreted contrast agents, allowing a fast preoperative workup of potential donors using a single imaging approach. Visualization and classification of the intrahepatic biliary variants are the most challenging parts of the preoperative evaluation, with variants in as many as 45% of cases [9]. A correct understanding of donor biliary anatomy is crucial to achieve safe donor hepatic resections and to minimize recipient biliary complications [8].
Overall, in our study, imaging data sets acquired on the 3.0-T magnet provided better sensitivity in the detection of biliary ductal variants in 15 volunteers, whereas the specificity was only marginally better. Although this was initially disappointing, three aspects must be addressed in more detail: First, the number of volunteers may not be large enough to obtain significant differences. Second, the results regarding specificity provided by the 1.5-T MR system are already very good, and a specificity of 97% as seen on the 3D TSE sequence on the 1.5-T MR system will be difficult to improve further. Finally, the biggest change was seen in the sensitivity in the detection of intrahepatic biliary variants when using the 3D TSE sequence (from 53% at 1.5 T to 87% at 3.0 T). This finding may reflect the true advantage of 3.0-T MR systems, in which the gain in MR signal is particularly helpful in sequences with thin slices of approximately 1 mm.
Reviewers were asked to categorize their certainty of the intrahepatic biliary anatomy by selecting from the following choices: 0 = relevant structures not seen; 1 = equivocal; 2 = probably seen; 3 = definitely seen. The statistical improvement in confidence levels seen with the 3.0-T studies is evidenced by a shift from an interpretation of "equivocal" to one of "probably seen" (confidence levels of 1.73, 1.6, and 2.0 for the three reviewers) and in the direction of "probably seen" to "definitely seen" (confidence levels of 2.31, 1.96, and 2.27). With the data sets from the 3.0-T system, we are able to shift reviewer confidence toward a more definitive interpretation, thus providing a more persuasive impression that may in the future allow better preoperative planning without further invasive imaging or intervention.
Limitations
The major limitations of our study are the small number of volunteers and
the lack of a true gold standard with regard to the presence and type of
biliary variants. Obviously, no invasive biliary imaging, such as endoscopic
retrograde cholangiography or percutaneous transhepatic cholangiography, can
be justified in volunteer studies. However, even contrast-enhanced CT or MRI
using bile duct-excreted contrast agents seems to be excessive in a volunteer
study and would not be approved by our institutional review board for human
investigation. Therefore, the consensus interpretation performed, on the basis
of all the available imaging data, by three senior abdominal radiologists who
otherwise were not involved in the study, appears to be the best gold standard
obtainable.
A minor limitation of our study may be the different sequence parameters used on the 1.5- and 3.0-T MR systems. However, these are the default protocol settings currently recommended by the vendor. This causes TRs and TEs that are sometimes the same, sometimes longer, and sometimes shorter. Thus, CNR differences could theoretically be attributable simply to comparing an optimal 3.0-T sequence to a suboptimal 1.5-T sequence. However, it is more likely that the 1.5-T sequences are closer to optimal and that any advantages of the 3.0-T sequences are real. It is also possible that the differences might be greater in future comparisons with more refined 3.0-T sequences. The same holds true regarding the radiofrequency transmitter and receiver technology, because we are comparing a relatively mature 1.5-T technology to a much newer technology at 3.0 T.
Summary
In conclusion, MR cholangiography at 3.0 T is a reliable means of providing
all relevant information of the biliary ductal system. Compared with MR
cholangiography at 1.5 T, MR cholangiography at 3.0 T offers improved CNR and
a higher level of confidence for depicting intrahepatic variants. In addition,
the 3D TSE sequence has the highest CNR at both 1.5 and 3.0 T and is more
sensitive for detecting intrahepatic biliary ductal variants at 3.0 T than at
1.5 T.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. J. Chang, I. R. Kamel, K. J. Macura, and D. A. Bluemke 3.0-T MR Imaging of the Abdomen: Comparison with 1.5 T RadioGraphics, November 1, 2008; 28(7): 1983 - 1998. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. I. Sainani, O. A. Catalano, N.-S. Holalkere, A. X. Zhu, P. F. Hahn, and D. V. Sahani Cholangiocarcinoma: Current and Novel Imaging Techniques RadioGraphics, September 1, 2008; 28(5): 1263 - 1287. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. K. Kuhl, F. Traber, J. Gieseke, W. Drahanowsky, N. Morakkabati-Spitz, W. Willinek, M. von Falkenhausen, C. Manka, and H. H. Schild Whole-Body High-Field-Strength (3.0-T) MR Imaging in Clinical Practice * Part II. Technical Considerations and Clinical Applications Radiology, April 1, 2008; 247(1): 16 - 35. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M. Haystead, B. M. Dale, and E. M. Merkle N/2 Ghosting Artifacts: Elimination at 3.0-T MR Cholangiography with SPACE Pulse Sequence Radiology, February 1, 2008; 246(2): 589 - 595. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. M. Akisik, K. Sandrasegaran, A. M. Aisen, C. Lin, and C. Lall Abdominal MR Imaging at 3.0 T RadioGraphics, September 1, 2007; 27(5): 1433 - 1444. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. I. Sher Invited Commentary RadioGraphics, September 1, 2007; 27(5): 1462 - 1464. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
