| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Diagnostic Radioology, Chonbuk National University Medical
School and Hospital, 634-18 Keumam dong, Jeonju, Chonbuk, South Korea.
2 Department of Radiology and Institute of Radiation Medicine, Seoul National
University College of Medicine, Seoul, South Korea.
3 Department of Internal Medicine, Chonbuk National University Medical School
and Hospital, Jeonju, Chonbuk, South Korea.
Received February 16, 2005;
accepted after revision July 15, 2005.
Address correspondence to Y. K. Kim
(jmyr{at}dreamwiz.com).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. During a 30-month period, 148 patients suspected of having biliary stones and who underwent MRI including MRCP, a fat-suppressed T1-weighted fast low-angle shot (FLASH) sequence, and an axial HASTE sequence were enrolled in this study. The biliary stones were confirmed by ERCP, surgery, and percutaneous transhepatic cholangiography. Of these 148 patients, 73 had extrahepatic stones, 45 had intrahepatic stones, 20 had both extrahepatic and intrahepatic stones, and 10 had no biliary stones. Two separate sets of images, the MRCP set (composed of MRCP and axial HASTE) and the combined interpretion of the MRCP set and the T1-weighted image, were analyzed independently and separately by two observers. The diagnostic accuracy was evaluated using the receiver operating characteristic method. The sensitivity and specificity were also calculated.
RESULTS. For common duct stones, the diagnostic accuracy and the sensitivity of both image sets showed similar values without any significant difference (0.998 [97.8%] for the combined interpretation; 0.988 [97.8%] for observer 1 and 0.995 [96.8%] for observer 2 for the MRCP set). However, for the intrahepatic stones, the diagnostic accuracy (0.993) and the sensitivity (98.5%) of the combined interpretation were significantly higher than those of the MRCP set for the two observers (0.926 [83.8%] for observer 1 and 0.922 [85.3%] for observer 2) (p < 0.05). No significant difference was seen in the specificity of the two image sets for both the intrahepatic and the common duct stones.
CONCLUSION. Combining the axial T1-weighted image with MRCP is valuable for detecting intrahepatic stones.
Keywords: biliary system • MR cholangiopancreatography • MRI • MR technique • T1-weighted image
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
On MRCP, which uses heavily T2-weighted sequences [12], the diagnosis of bile duct stones is based on round, ovoid, or irregularly shaped signal voids being identified within the bile ductal lumen with hyperintensity [13]. However, it is occasionally difficult to detect bile duct stones based only on the T2-weighted technique of MRCP and on the T2-weighted axial images, especially when the calculi are only partially outlined or they are not outlined at all by the hyperintense bile. In many clinical practices, we have experienced the complementary role of T1-weighted axial images for depicting bile duct stones on MRCP and on axial T2-weighted images. Yet to date, the preliminary studies have focused mainly on the T2-weighted MRCP techniques for the diagnosis of bile duct stones or biliary stricture, except for some studies that have reported on MRCP using mangafodipir as the T1 contrast agent [14-17].
In this study, we retrospectively evaluated whether the T1-weighted images have a complementary role to MRCP in the diagnosis of intrahepatic and extrahepatic bile duct stones.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The remaining 148 patients (95 men and 53 women; age range, 33-70 years; mean age, 58 years) underwent ERCP, surgery, both ERCP and surgery, or PTC, and they were included in this study. Of these 148 patients, 73 had extrahepatic stones, 45 had intrahepatic stones, and 20 had both extrahepatic and intrahepatic stones. The remaining 10 patients were suspected of having biliary calculi on the basis of clinical or sonographic findings, but were proved to have no biliary calculi on the ERCP and MR examination. The biliary stones were confirmed by surgery with intraoperative cholangiography in 24 patients, by ERCP with or without stone removal in 94 patients, by PTC in 10 patients, and by a combination of surgery and ERCP in 10 patients. Of the 34 surgically confirmed patients, 18 underwent hepatic segmentectomy with surgical exploration of the bile duct (left lateral segmentectomy, n = 12; right posterior segmentectomy, n =6) for intrahepatic stones with biliary stricture, whereas the remaining 16 patients (intrahepatic and extrahepatic stones, n = 13; extrahepatic stones, n = 3) underwent surgical exploration of the bile duct with intraoperative cholangiography.
The time interval between MRCP and ERCP or PTC was 2-7 days (mean interval, 3 days), and the time interval between MRCP and surgery was 8-17 days (mean interval, 14 days). None of these patients had coexistent bile duct cancer. Of the 65 patients with intrahepatic bile duct stones, 21 were diagnosed as having recurrent pyogenic cholangitis based on surgery (for 18 patients) and the following typical imaging findings [18]: bile duct stones, disproportionately severe dilatation of the extrahepatic ducts with mild or no dilatation of the intrahepatic ducts, focal strictures, acute peripheral tapering. decreased arborization, an increased branching angle of the intrahepatic bile ducts, and segmental hepatic atrophy.
Imaging Techniques
MRI was performed on a 1.5-T superconducting scanner (Magnetom Symphony,
Siemens Medical Solutions) with a combination of a phased-array body coil and
a spine array coil for the signal reception. The patients were asked to fast
for a minimum of 6 hours before imaging to promote gallbladder filling.
Approximately 10 minutes before the examination, the patients drank a mixture
of 8.4 mg of superparamagnetic iron oxide (Feridex, Advanced Magnetics) and
300 mL of water to prevent overlapping or concealing of the biliary system by
the high signal intensity fluid in the duodenum and stomach. The axial
fat-suppressed T1-weighted fast low-angle shot (FLASH) sequence images and the
axial HASTE images were obtained before MRCP was performed. The imaging
parameters for the fat-suppressed FLASH sequences were TR/TE, 159/2.6; flip
angle, 70°; matrix, 120 x 256; slice thickness, 7 mm; intersection
gap, 10%; receiver bandwidth, 260 Hz/pixel; field of view, 35-40 cm; signal
average, 1. Twenty images were obtained using a concatenation technique during
37 seconds in two breath-holds. Presaturation bands that were positioned
superior and inferior in relation to the imaging volume were applied once
during each TR to reduce the signal intensity from the inflowing blood. The
HASTE images were performed using the following parameters: TR/TE,
infinite/85; refocusing flip angle, 180°; echo-train length, 192; matrix
146 x 256; field of view, 35 cm; slice thickness, 6-mm; scanning time,
16 seconds for 23 slices. Two kinds of MRCP techniques were performed,
single-shot single-projection thick-section RARE and multisection half-Fourier
RARE. The single-projection thick-section MRCP images using the RARE technique
were obtained at various angles (± 25°, 15°, 0°) to the
coronal plane and at two axial planes with a slice thickness of 50-80 mm. The
imaging parameters for the single-shot single-projection thick-section RARE
sequences were TR/TE, infinite/1,100; flip angle, 150°; echo-train length,
320; matrix, 320 x 320; field of view, 350 mm; signal average, 1;
acquisition time, 3 seconds. The multisection half-Fourier RARE images were
obtained at an angle of ± 15° to the coronal plane for inclusion of
both the common bile duct and the pancreatic duct. The imaging parameters for
the multisection half-Fourier RARE sequences were TR/TE, infinite/84; flip
angle 150°; echo-train length, 218; matrix, 218 x 256; field of
view, 350 mm; section thickness, 4 mm with no gap; slices, 19; signal average,
1; acquisition time, 15 seconds.
One gastroenterologist with 6 years' experience performed all the ERCP procedures with technologist-assisted fluoroscopy. The standard ERCP technique was used. Most of the common duct stones were extracted by ERCP together with sphincterectomy. All the PTCs were performed by one experienced interventional radiologist using a standard technique. The cholangiograms acquired during ERCP or PTC were jointly reviewed by a study coordinator and by the operator of each procedure.
Image Analysis
Two gastrointestinal radiologists experienced in interpreting liver imaging
in their daily clinical practice for at least 3 years reviewed the images both
independently and separately. They were unaware of the results of the clinical
data, the laboratory data, and the final diagnosis. Two separate sets of
images were analyzed—the MRCP set, which included the single-projection
thick-section RARE images, the multisection half-Fourier RARE images, and the
axial HASTE images and a combination of the MRCP set and the T1-weighted FLASH
images. To minimize any learning bias, a 2-week interval was scheduled between
each interpretation session. All the images were reviewed on a monitor of a
2000 x 2000 PACS (PACS, Marotech).
Biliary calculi were defined as nodular areas of signal void or low signal intensity within a bile duct lumen that were partially or completely surrounded by the high signal intensity of bile on MRCP or on the axial T2-weighted image. Biliary calculi on the T1-weighted axial image were defined as nodular filling defects within the bile duct lumen, showing low signal intensity or high signal intensity relative to the surrounding bile duct lumen. The standard of reference regarding the diagnosis of bile duct stone included the surgical findings of biliary surgery with intraoperative cholangiography, the interpretation of the ERCP with and without stone removal, and the PTC by consensus of a study coordinator and the operator of each procedure.
Each reviewer was asked to record the presence, number, and location of the biliary calculi (right or left intrahepatic duct and the common duct), assigning each a confidence level on a 5-point scale: 1 for definitely absent, 2 for probably absent, 3 for possibly present, 4 for probably present, and 5 for definitely present. Before the image interpretation, all the reviewers were aware that the sensitivity calculations were made based only on those lesions that were awarded a confidence rating of 3 or more. Comparisons between the scored lesions by each reviewer and the standards of the reference examinations were performed by a study coordinator.
Statistical Analysis
For each imaging set, a binomial receiver operating characteristic (ROC)
curve was fitted to each observer's confidence rating data using a maximum
likelihood estimation program (ROCKIT version 0.9B, courtesy C. E. Metz,
University of Chicago, IL, 1998)
[19]. The diagnostic accuracy
of each imaging set for each observer was calculated by measuring the area
under the ROC curve (A index or Az). The differences
between the imaging sets in terms of the mean area under the ROC curves were
statistically analyzed using the two-tailed Student's t test for
paired data. The sensitivity and specificity for each image set for the
diagnosis of stones in the intrahepatic ducts and the common duct were then
calculated.
|
|
|
To assess the interobserver agreement for the evaluation of the two imaging sets, we calculated the kappa statistic for the two observers [21]. Agreement between the blinded observers is reported in terms of the kappa values, with those greater than zero indicating a positive correlation. Kappa values less than 0.20 indicated positive but poor agreement; those of 0.21-0.40, fair agreement; those of 0.41-0.60, moderate agreement; those of 0.61-0.80, good agreement; and those greater than 0.81, excellent agreement. The level for statistical significance was p < 0.05. The statistical analyses were calculated using SPSS version 8.0 (Statistical Package for the Social Sciences) for Windows (Microsoft).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The signal intensities of the biliary calculi on MRCP or on the T2-weighted axial images were low relative to the high signal intensity of the bile in 146 of 158 stones (92.4%); the residual stones were invisible or had isosignal intensities. The signal intensities of biliary calculi on the T1-weighted axial images were high relative to the adjacent bile or liver parenchyma in 134 of 158 stones (84.8%), invisible or isosignal in 20 (12.7), and low in 4 (2.5%) (Figs. 1A, 1B, and 1C).
The calculated individual Az values for the interpretation of each image set in terms of the presence of biliary stones are shown in Table 1. For the common duct stones, the individual Az values with the combined interpretation of the MRCP set and the T1-weighted image were similar to those of the MRCP set alone. For the diagnosis of intrahepatic stones, the Az values with the combined interpretation of the MRCP set and the T1-weighted image were significantly higher than those of the MRCP set alone for both observers (Az value for the combined interpretation, 0.993; Az value for MRCP, 0.926 for observer 1, 0.922 for observer 2; p = 0.001) (Figs. 2A, 2B, 2C, and 2D).
|
|
|
|
|
The sensitivities and specificities for each observer and for each image set are listed in Table 2. For the common duct stones, the sensitivities of the combined interpretation of the MRCP set and the T1-weighted image (97.8%) were similar to those of the MRCP set alone for both observers (97.8% for observer 1, 96.8% for observer 2) (Figs. 1A, 1B, and 1C). Although two tiny common duct stones (3 mm in diameter) in two patients were not found by any observer on the MRI, these stones were proved by ERCP on stone removal. To detect the intrahepatic stones, the sensitivities of the combined interpretation of the MRCP set and the T1-weighted image (98.5%) were significantly higher than those with the MRCP set alone for both observers (83.8% for observer 1, 85.3% for observer 2; p = 0.002, 0.004, respectively) (Figs. 2A, 2B, 2C, 2D, 3A, 3B, 3C, 3D, and 3E). No significant differences were seen in the MRCP set, the combined interpretation of the MRCP set, and the T1-weighted image in specificity for evaluating both the common duct stones and the intrahepatic duct stones (p > 0.05). For evaluating the common duct stones, one false-positive finding was found on both the MRCP set and on the combined MRCP set and T1-weighted image interpretation. For evaluating the intrahepatic stones, two false-positive findings were found by each observer on the MRCP set. On MRCP, the false-positive findings were attributed to the partial volume averaging effect or to pneumobilia. One reviewer interpreted the high signal intensity seen in the distal common duct as a common duct stone in one patient, but this was later regarded as reflux of the superparamagnetic iron oxide agent in the stomach and duodenum (Fig. 4).
|
|
|
|
|
|
|
For evaluating the common duct stones, the kappa values for the two observers were 0.837 for MRCP and 0.987 for the combined interpretation of MRCP and the T1-weighted image, showing excellent interobserver agreement with regard to the presence of biliary calculi. For the intrahepatic duct stones, excellent agreement was also obtained between the two reviewers with regard to the presence of biliary calculi; this was shown by the kappa value of 0.854 for MRCP and 0.970 for the combined interpretation of MRCP and the T1-weighted image.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In our study, we performed ROC analysis to evaluate the diagnostic capabilities of the various imaging techniques because ROC analysis is a precise statistical method for evaluating the diagnostic performance of various imaging techniques and, therefore, more closely reflects actual clinical practice [19]. Our study results showed no significant differences between the two image sets (the MRCP set and the combination of the MRCP set and the T1-weighted image) for diagnosing common duct stones, both for the diagnostic accuracy of the ROC analysis and for the detection sensitivity. However, for intrahepatic stones, the two observers achieved a higher diagnostic accuracy and detection sensitivity with the combined interpretation of the MRCP set and the T1-weighed image set than with the MRCP set alone. Interpreting the T1-weighted image together with the MRCP set resulted in additional detection of intrahepatic stones in nine or 10 patients for each observer, respectively, compared with using only the MRCP set. A retrospective review of these nine or 10 patients found that all of them had minimal intrahepatic bile duct dilatation relative to the other study patients who had intrahepatic stones. This led to these patients being misdiagnosed as having only biliary stricture without calculi on the MRCP set (Figs. 2A, 2B, 2C, and 2D). In one patient, a single bile duct stone at the bifurcation of the segment 8 intrahepatic duct was regarded as a normal portal vein on the MRCP set, but the bright high signal intensity of the intrahepatic stone was clearly detected on the T1-weighted image by both observers (Figs. 3A, 3B, 3C, 3D, and 3E). Furthermore, most of the intrahepatic stones that were missed on the MRCP set were clearly depicted as having bright high signal intensity on the T1-weighted image; this allowed both observers to arrive at the correct diagnosis.
In our study, most of the signal intensities of the biliary calculi on the T1-weighted images were high signal intensities relative to the adjacent bile or the liver parenchyma (84.8%, 134/158), and the remaining biliary calculi were invisible or had isosignal intensities (12.7%, 20/158) and low signal intensity (2.5%, 4/158). This result is similar to that of several previous reports that showed the high signal intensities of gallstones or intrahepatic stones on T1-weighted MRI [23-27]. Based on the previous studies by Ukaji et al. [26], metal irons in the pigment of gallstones were the main cause of the high signal intensity of gallstones on T1-weighted MRI. However, this study was retrospective, so evaluation of the chemical composition of the removed biliary stone and correlation analysis of the biliary calculi signal intensity with the stone's chemical composition were not made. Based on the fact that many of the study's subjects had underlying recurrent pyogenic cholangitis (known as intrahepatic pigmented calculus disease) and based on the high prevalence of bilirubin-pigmented stones in Korea [18], most of the intrahepatic stones in our study could be regarded as bilirubin-pigmented stones.
When comparing using the T2-weighted axial image (HASTE) and the T1-weighted axial image for evaluating intrahepatic duct stones, the T1-weighted image with fat suppression might be advantageous for depicting stones compared with T2-weighted imaging. Indeed, the high signal intensity of most intrahepatic stones on the T1-weighted image clearly showed excellent contrast to the liver parenchyma and the adjacent vessels and to the bile duct lumen, regardless of the bile duct dilation. However, when the intrahepatic bile ducts are not sufficiently dilated, the lower signal intensity of the liver parenchyma using heavily T2-weighted imaging such as HASTE or RARE sequences with a large echo-train length [27-29] could be detrimental for revealing the low signal intensity of intrahepatic stones. This explanation is supported by several cases of intrahepatic stones, such as those in Figures 2A, 2B, 2C, 2D, 3A, 3B, 3C, 3D, and 3E. In addition, the intrahepatic vascular structures on the T2-weighted axial image are shown as having low signal intensity because the fast flowing liquids usually leave the acquisition volume between the radiofrequency excitation and the readout time. This occasionally makes it difficult to differentiate a vessel from a biliary stone, as can be seen in Figures 3A, 3B, 3C, 3D, and 3E.
The sensitivity of the MRCP set for detecting common bile duct stones in this study was comparable to that of previous reports [6-10, 30, 31]. In our study, the sensitivity of only the MRCP set for detecting intrahepatic stones was 83.8% and 85.3% for reviewers 1 and 2, respectively. However, the sensitivities of the combined interpretation of the MRCP set and the T1-weighted image were slightly higher than MRCP sensitivities of the previous reports, which ranged from 96-97% [11, 32]. As previously mentioned, the inclusion of patients with insufficient duct dilatation and those with biliary stricture resulting from recurrent pyogenic cholangitis might have been responsible for the relatively lower sensitivity of the MRCP set in this study compared with the previous studies for detecting intrahepatic stones. In two patients, two common duct stones that were 2 mm in diameter were missed on all the imaging techniques. However, when considering the time interval between the MRI examination and ERCP and the underlying recurrent pyogenic cholangitis with intrahepatic stones in these two patients, the common duct stones might have been fragments of the intrahepatic stones that were most likely fragile calcium bilirubinate stones [18].
Several kinds of negative gastrointestinal MRI contrast agents, such as perfluorocarbons [33], clays [34], and iron oxide particles [35], have been used during MRCP to decrease the high signal intensities of the intraluminal bowel fluids. We used a superparamagnetic iron oxide agent to eliminate the high signal intensities of the overlapping fluid-containing organs. In our study, we obtained excellent suppression of the high signal intensities of the intraluminal fluid within the stomach and duodenum with oral administration of the superparamagnetic iron oxide agent. However, among the false-positive findings of this study on the T1-weighted image, one reviewer interpreted high signal intensity in the distal common duct as a common duct stone in one patient. This was subsequently regarded as reflux of the superparamagnetic iron oxide agent in the stomach and duodenum, and the enterobiliary reflux was then proven by ERCP (Fig. 4). The possible explanation of this false-positive finding in our study was the T1 shortening effect of the superparamagnetic iron oxide agent on the T1-gradient echo imaging with a very short TE [35]. This false-positive finding could be resolved via the combined interpretation of the MRCP set and the T1-weighted image, because the MRCP set alone without adding the T1-weighted image showed very high sensitivity and specificity for diagnosing common duct stone. Also, the difference in the shape of the low signal intensity between the stone and the reflux of the superparamagnetic iron oxide agent on MRCP could easily make it possible to differentiate between them. Even if adding the T1-weighted image to the MRCP set did not contribute to detecting additional common duct stones, we believe that the differentiation of common duct stone from blood and proteinaceous material that mimics calculi on the T2-weighted image [36] might be easier with the addition of the T1-weighted image.
Our study has some limitations. First, given that in routine clinical practice most of the patients who have no evidence of biliary stone on MRI do not undergo further invasive examinations, our study population was probably slightly biased. Therefore, the actual sensitivity of MRI for biliary stones might have been overestimated by minimizing the number of the false-negative lesions. Second, the exact number of biliary stones detected on the MRI examination was not correlated with the standard of reference; that is, calculating the sensitivity of this study was not based on the precise number of stones but rather on the presence or absence of stones in three places—the right intrahepatic duct, the left intrahepatic duct, and the common duct. However, a large number of stones may fragment during surgery or stone removal. Therefore, any exact correlation of the original number of stones between the MRI examination and the gold standard might be difficult to achieve. Third, no correlation of the biliary calculi signal intensity with their chemical composition was performed in our study. Last, the intermediately effective TE of the HASTE and RARE sequences and the relatively thick section of the HASTE image might have contributed to their low sensitivity for detecting intrahepatic stones.
In conclusion, adding interpretation of the T1-weighted image to the MRCP set was useful for detecting intrahepatic bile duct stones, but this showed no additional value for the detection of common duct stones compared with using the MRCP set alone.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. R. Nandalur, H. K. Hussain, W. J. Weadock, E. J. Wamsteker, T. D. Johnson, A. S. Khan, A. R. D'Amico, M. K. Ford, S. R. Nandalur, and T. L. Chenevert Possible Biliary Disease: Diagnostic Performance of High-Spatial-Resolution Isotropic 3D T2-weighted MRCP Radiology, December 1, 2008; 249(3): 883 - 890. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
