| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Mayo Clinic Jacksonville, 4500 San Pablo Rd.,
Jacksonville, FL 32224.
2 Department of Transplantation, Division of Transplant Medicine, Mayo Clinic
Jacksonville, Joe Adams 1100 Transplant Center, Jacksonville, FL 32216.
Received October 14, 2003;
accepted after revision November 13, 2003.
Address correspondence to M. D. Bridges
(Bridges.Mellena{at}mayo.edu).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Our study group consisted of 25 consecutive patients who were referred for MR evaluation of clinically suspected biliary complications after orthotopic liver transplantation. Conventional MRC in the axial and coronal planes was performed in each patient, followed by fat-suppressed volumetric gradient-echo imaging in the same planes both before and after the IV administration of mangafodipir trisodium. Imaging was performed in all patients until the contrast agent was seen in the bowel. Images were then graded for quality, visualization of bile ducts and anastomoses, presence of significant stricture or leak, and level of diagnostic confidence.
RESULTS. Mangafodipir trisodium–enhanced MRC tended to outperform conventional MRC in overall image quality and extrahepatic duct visualization; it was also more effective in delineating biliary anastomoses, and the difference was statistically significant (p < 0.001). All 25 enhanced examinations were considered diagnostic. Diagnostic confidence was scored as poor or lacking in 14 of the conventional MRC examinations for biliary stenosis and in 12 examinations for biliary leak.
CONCLUSION. Enhancement with mangafodipir trisodium improves the performance of MRC for the detection and exclusion of biliary abnormalities after orthotopic liver transplantation. Future investigations should compare the performance of mangafodipir trisodium–enhanced MRC with the performance of more invasive techniques.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
As the frequency of liver transplantation procedures rises, so do the associated complications and, consequently, the need for accurate, rapid diagnosis [2, 3]. Biliary complications, especially anastomotic stricture and leak, are an important and common cause of morbidity and graft failure. Frequencies vary among centers, with recent larger cohorts ranging from 11% to 30% and reported ranges overall of 9–50% [4–8].
Because biliary complications do not present specific clinical or laboratory findings, the posttransplantation patient is often referred for diagnostic imaging. Analysis of the postoperative biliary system has traditionally been the purview of endoscopic retrograde cholangiography (ERC) or percutaneous transhepatic cholangiography [9–14]. These methods have the advantage of being both diagnostic and potentially therapeutic techniques. Unfortunately, they are also invasive, expensive, and subject to their own inherent complications. Furthermore, ERC is of little value in the patient with a biliary–enteric anastomosis.
MR cholangiography (MRC), a technique dependent on the high T2 signal intensity of bile, is recognized as a noninvasive alternative to these methods for evaluation of the biliary system [15–17]. In our experience, however, conventional MRC has been less than satisfactory. For example, although the biliary anastomosis is easily seen, most reconstructions show some degree of narrowing. Furthermore, donor–recipient mismatches in duct size occur commonly, which makes analysis of anastomotic narrowing more difficult. At issue in both of these situations is not only morphologic narrowing but also functional significance, especially because significant donor duct dilatation is not seen often enough in our population to be a useful predictive sign. As a technique that depends on the high signal intensity of bile to depict ducts, conventional MRC can also be limited by conflicting signals from ascites, perihepatic fluid collections, and soft-tissue edema, all of which are common in the postoperative period. Similarly, conventional MRC cannot easily distinguish between a biloma and a simple perihepatic collection.
Mangafodipir trisodium (Teslascan, Amersham) is an IV-administered, T1-shortening contrast agent developed for hepatic imaging. Its safety has been shown in multicenter trials [18]. The agent is primarily excreted in the bile and has been applied to T1-weighted biliary imaging, with investigators addressing its potential in several recent studies [19–22].
In combination, these investigations suggested that mangafodipir trisodium–enhanced MRC could provide the high-quality anatomic imaging that is necessary for the evaluation of small-caliber ductal structures, even in the setting of intraabdominal fluid and soft-tissue edema. Detection of strictures and intraductal debris, and direct visualization of anastomotic leaks, might be possible. Furthermore, if a correlation between delay in contrast excretion into the bile ducts and the physiologic significance of an anastomotic stricture could be shown, and if exclusion of excretion delays caused by synthetic dysfunction could be accomplished, then the technique could add information previously unavailable with MRI.
The purpose of this study was to examine the usefulness of incorporating mangafodipir trisodium–enhanced imaging into our MRC protocol in cases of suspected posttransplantation biliary complications. Specifically, our intent was to determine whether this approach could improve ductal visualization or diagnostic confidence over that provided by conventional MRC alone, to evaluate the effect of ascites or edema and the type of anastomotic reconstruction on diagnostic confidence for the two techniques, and to determine whether delay in contrast excretion could be correlated with functional significance of ductal narrowing.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
MRI was performed on a 1.5-T scanner (Symphony or Sonata, Siemens Medical Solutions) using a phased array torso coil. Imaging sequences included axial and coronal HASTE sequences through the entire abdomen for ductal localization as well as for a general survey of the abdominal structures and potential fluid collections. MRC sequences were then obtained as follows: axial and coronal multislice HASTE imaging through the central biliary tree (TR/TE, 1,000/89; refocusing angle, 180°; slices, 20; slice thickness, 3 [coronal] or 4 [axial] mm with a 10% gap; matrix, 168–192 x 256; field of view, 220–340 mm [depending on body habitus]). We also obtained six projectional HASTE images in the coronal and coronal oblique planes (TR/TE, 2,800/1,100; refocusing angle, 150; single slice per breath-hold; slice thickness, 40 mm; matrix, 256 x 256; field of view, 240–340 mm). Images were obtained either during suspended respiration (20- to 25-sec breath-hold) or using a free-breathing navigator echo-gated technique, depending on patient ability.
Axial and coronal volumetric spoiled gradient-echo imaging with chemical fat-suppression pulses was then performed (TR range/TE range, 3.2–3.6/1.3–1.5; flip angle, 12°; partitions, 64; slice thickness, 1.5 mm; matrix, 156 x 256; field of view, 260–340 mm; phase oversampling on the coronal images). Each of these acquisitions was obtained in a breath-hold (20–25 sec). Except that the asymmetric echo parameter was enabled to decrease TR and slice thickness was diminished, the parameters were identical to our protocol for gadolinium-enhanced imaging of the liver.
Total scanning time averaged 45 min in uncomplicated cases, which is the time allotted in our practice for a single MRI examination. All 25 examinations were technically adequate. No adverse events clearly attributable to the administration of mangafodipir trisodium were encountered.
After a standardized IV injection of 10 mL of mangafodipir trisodium administered slowly over several minutes and followed by a 20-mL saline flush, the axial and coronal 3D imaging sequences were repeated at 5 and 15 min, with the delay interval calculated from the completion of injection. After 15 min of scanning, the radiologist was consulted. If the study was considered adequate for evaluating ductal and anastomotic anatomy, excretion into bowel, and bile leak, imaging was ended and maximum-intensity-projection (MIP) reconstructions of the best coronal data set were created. In cases in which excretion had not yet been visualized, or when anastomotic leak remained in question, delayed imaging was performed up to 1 hr. If visualization was still not documented, further scanning was done the next morning.
Each MRI examination was stored in digital archives as two sets of images, with the conventional MRC sequences constituting one set and the volumetric mangafodipir trisodium–enhanced sequences the second set. The latter included source images as well as MIP reconstructions. All image review was performed on a diagnostic interpretation workstation (MagicView 1000, Siemens Medical Solutions), and each set of images was evaluated separately during a different consensus interpretation session.
Two radiologists who were experienced in the interpretation of abdominal MR images, interpreting in consensus and unaware of the specifics of the patient's clinical status, subjectively evaluated image quality and visualization of anastomoses, intrahepatic bile ducts, and extrahepatic bile ducts using a 5-point scale (with 1 being poor and 5, excellent). The radiologists also determined the presence or absence of significant anastomotic stenoses or leaks (yes, no, indeterminate) and graded their confidence (none, poor, good) regarding each diagnostic determination. Decisions regarding stenotic significance were subjectively based, drawing from experience with ERC and percutaneous cholangiography. In our endoscopy practice, the biliary reconstruction is considered within normal limits if an 8.5-mm occlusion balloon can be passed easily across the anastomosis. On the other hand, dilatation is routinely attempted if the anastomosis will not permit passage of a balloon larger than 4 mm. Intermediate results are considered indeterminate. Diameter measurements were not attempted on the MR images because the ERC standard refers more to ease of passage than to exact measurements and because no accepted measurement standards exist for percutaneous cholangiography.
For the enhanced image sets, the reviewers also examined the time delay between completion of mangafodipir trisodium administration and its appearance in the bowel; failure of copious excretion by 15 min was considered definitely abnormal and suggestive of significant stenosis. Type of anastomotic reconstruction, presence of ascites or focal fluid near the anastomosis, and presence of soft-tissue edema at the liver hilum were also recorded. The data gathered during these interpretations was then entered into a spreadsheet.
To allow correlation of laboratory measures of hepatic dysfunction with the imaging results, concurrent liver function test values were obtained from the medical records for all patients and were added to the spreadsheet data.
To compare the two techniques, we extracted the patients for whom quality or confidence scores differed and examined the proportion of those for whom the mangafodipir trisodium–enhanced images had the better score. An exact binomial test of the null hypothesis was performed on the basis of the observed proportion. A p value of less than 0.05 was considered significant. Finally, to investigate whether diagnostic confidence was associated with the presence or absence of a specific feature, we performed a Fisher's exact test.
Six months after completion of data collection, examination of the medical records of all 25 patients was undertaken, and note was made of interval imaging studies (including cross-sectional and cholangiographic examinations), biliary interventions, and repeated transplantation procedures. Eight patients had undergone subsequent ERC, four had percutaneous transhepatic cholangiography, six had abdominal CT, and three had repeated MRC. Six patients have had no further abdominal imaging other than sonography. Of these, four have had at least one liver biopsy, and the other two have been followed up clinically. Also noted for each patient were relevant laboratory values and subsequent clinical status.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
The data for diagnoses of anastomotic stenoses and leaks are summarized in Tables 2 and 3. For each patient, the pairs of image sets were consistent with one another, in the sense that on no occasion was there a clearly positive diagnosis for one set contradicted by a negative diagnosis for the other set. However, diagnoses were made with greater confidence on the basis of the contrast-enhanced MRC (p < 0.001 for both stricture and leak): in all 25 patients, the reviewers had good confidence in their exclusion or inclusion of significant stenosis or leak on the basis of the mangafodipir trisodium–enhanced images. In contrast, 56% and 48% of the unenhanced MRC image sets were found to be either poorly diagnostic or nondiagnostic for stenosis (Fig. 2A, 2B, 2C) and leak (Fig. 3A, 3B, 3C), respectively.
|
|
|
|
|
|
|
|
The data regarding the diagnostic impact of type of anastomosis are outlined in Table 4, and for ascites and edema in Table 5. Anastomotic type had no statistically significant effect on diagnostic confidence for either technique, nor did the presence of ascites or edema for the enhanced technique. However, ascites or edema did tend to have a negative impact on the confidence of reviewers interpreting the conventional MRC sequences. Although this tendency did not achieve statistical significance for diagnosis of stenosis (p = 0.51), the evidence was clear for a negative impact on confidence in diagnosing leak (p = 0.004).
|
|
Several other observations regarding the enhanced image sets are of interest. First, in 21 of the 25 patients, appearance of the contrast agent in the recipient's common duct and small bowel was documented within 15 min after completion of the IV injection. This group included two patients in whom the reviewers diagnosed mild strictures because of the degree of narrowing, an impression confirmed on subsequent ERC. Of the rest of this group, none was diagnosed with morphologic strictures on mangafodipir trisodium–enhanced MRC, and none was subsequently proven to have a stricture during a follow-up period of 4–6 months. Alternative diagnoses have included biopsy-proven cellular rejection, biliary necrosis, and drug-related hepatitis.
In two of the remaining four patients, visualization was delayed to 45 min and 1 hr, respectively; subsequent ERC showed high-grade anastomotic strictures in both patients (Figs. 4A, 4B and 5A, 5B, 5C). In neither of these patients was postoperative synthetic dysfunction thought to be responsible for the delay because their transplantation procedures had been performed 4 and 8 months previously. Direct serum bilirubin levels were only mildly elevated (0.4 and 0.6 mg/dL, respectively). In the final two patients, one with hemorrhagic graft necrosis and the other with extensive bile duct necrosis (Fig. 6A, 6B, 6C), both surgically proven, next-day imaging finally documented visualization.
|
|
|
|
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
On the other hand, depiction of the intrahepatic biliary tree was not usually improved by the enhanced technique. In fact, in a number of cases conventional MRC received higher scores, results at odds with the impression of Lee et al. [18] that T2-based MRC was often inadequate to define anatomic variants in the central intrahepatic biliary radicles. Several possible explanations exist for this discordance. First, patient motion—primarily respiratory—is a significant challenge in our postsurgical population. Also worth considering would be the hepatic synthetic dysfunction present to one degree or another in our patients: excretion of the contrast agent might not be as copious as in healthy prospective donors. Finally, our reviewers were evaluating the small peripheral as well as the central intrahepatic ducts.
Perhaps most important, in a number of cases (56% for stenosis, 48% for leak) reviewer diagnostic confidence increased significantly when they were given the mangafodipir trisodium–enhanced images. In no case was diagnostic confidence poor or lacking for significant stenosis or for leak at review of the enhanced sequences. However, in roughly half the cases, conventional MRC images resulted in poor or absent diagnostic confidence.
In our study group, prompt excretion (within 15 min) of mangafodipir trisodium into distal duct and bowel accurately indicated the absence of significant ductal stenosis. Of the four patients in whom excretion was delayed beyond 15 min, two subsequently required intervention for anastomotic strictures, and two required retransplantation because of graft failure. These results suggest a potential for establishing a time threshold for contrast visualization below which the ductal reconstruction is predictably normal.
Our experience with the mangafodipir trisodium–enhanced MRC protocol since the study period has suggested several drawbacks of the technique. Although acquired in less than 25 sec, the 3D gradient-echo sequences we use for enhanced imaging are motion-sensitive and depend heavily on the patient's ability to breath-hold, and image quality has suffered in a few patients. For these, single-shot T2-weighted images have been especially valuable. Clearly, if these patients had been included in our study group, diagnostic confidence for the enhanced technique would not have been 100%. To address this issue, we are working with parallel imaging techniques to shorten imaging times for very ill patients, as well as experimenting with gated sequences, which show some promise, especially for detecting leaks.
This imaging protocol can be time-consuming. Because mangafodipir trisodium is not injected until after the conventional sequences are acquired, an imaging hiatus of several minutes occurs. When that delay is added to the time needed to acquire the postinjection sequences and to create the MIP reconstructions, another half hour may be added to an approximately 30-min MRC protocol. However, because mangafodipir trisodium decreases the T2 of bile, leading to signal loss in the ducts [23], this order of imaging is appropriate. Additionally, we believe that T2-weighted imaging will remain important because of its ability to depict fluid collections, because of the relative motion insensitivity of its single-shot variants, and because of its good performance for the intrahepatic ducts.
Another potential issue is the patient with a long delay in ductal contrast visualization. Hopefully, as more experience accumulates concerning patients with cellular hepatic dysfunction, a threshold will be established beyond which only conventional MRC would be indicated, obviating unproductive repeated imaging.
Limitations of our study derive from its focus on comparison between two MRI techniques and from its lack of a gold standard. Consequently, although some patients did have correlative studies in a short period of time, those diagnosed as normal on MRI often did not undergo ERC or percutaneous transhepatic cholangiography. On the other hand, examination of the medical records for the subsequent 6 months revealed no patient who later required biliary intervention. Additionally, all but two patients either had crosssectional imaging follow-up (often multiple examinations) or underwent diagnostic liver biopsy. None of these examinations was interpreted as suggesting biliary obstruction or extravasation. At least four patients had follow-up MRC, with no change in findings.
Only one biliary leak was diagnosed among our cohort, but it was well depicted by mangafodipir trisodium–enhanced MRC. This is congruent with the excellent results reported by Vitellas et al. [20] in detecting leaks after cholecystectomy.
In conclusion, the addition of mangafodipir trisodium–enhanced T1-weighted MRC to our MRC protocol for the detection and characterization of biliary complications after orthotopic liver transplantation provided an improvement in visualization of extrahepatic ducts and biliary reconstructions when compared with conventional MRC. Mangafodipir trisodium–enhanced MRC also improved diagnostic confidence significantly. Further study is needed to determine when and if this approach can supplant more invasive procedures.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Novellas, T. Caramella, M. Fournol, J. Gugenheim, and P. Chevallier MR Cholangiopancreatography Features of the Biliary Tree After Liver Transplantation Am. J. Roentgenol., July 1, 2008; 191(1): 221 - 227. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Hoeffel, L. Azizi, M. Lewin, V. Laurent, C. Aube, L. Arrive, and J.-M. Tubiana Normal and Pathologic Features of the Postoperative Biliary Tract at 3D MR Cholangiopancreatography and MR Imaging RadioGraphics, November 1, 2006; 26(6): 1603 - 1620. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
