| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, MRI Section, University of Michigan Medical Center,
1500 E. Medical Center Dr., Ann Arbor, MI 48109-0300.
2 Department of Diagnostic Imaging, Rhode Island Hospital—Brown University
School of Medicine, 593 Eddy St., Providence RI 02912-9706.
Received September 18, 2001;
accepted after revision December 18, 2001.
Presented at the annual meeting of the American Roentgen Ray Society,
Atlanta, April-May 2002.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Ten patients prospectively underwent T2-MRCP and gadolinium-EOB—enhanced MR cholangiography (EOB-MRC). Three radiologists reviewed the unpaired, then the paired, examinations, rating biliary visualization using a 5-point scale. The common bile, right and left hepatic ducts, and second-order branches were evaluated. Improved biliary visualization using paired over unpaired tests indicated the added value of contrast media. Kappa values measured interobserver reliability. A regression model controlling for fixed effects due to reviewer and subject correlation quantified improvement in ratings attributable to paired review.
RESULTS. Average visualization ratings for unpaired review of
EOB-MRC were the following: common bile duct, 3.3; right hepatic duct, 2.7;
left hepatic duct, 2.5; second-order branches, 1.4. Average visualization
ratings for unpaired review of T2-MCRP were the following: common bile duct,
3.4; right hepatic duct, 1.8; left hepatic duct, 2.2; second-order branches,
0.9. Ratings improved using paired tests over EOB-MRC and T2-MRCP for all
structures (p < 0.001) except for T2-MRCP common bile duct ratings
(p > 0.05). Agreement was moderate to good except for EOB-MRC
common bile duct ratings. Paired review improved ratings (
2
< 0.0001) over T2-MRCP alone by 1.05 and over EOB-MRC alone by 0.68.
Despite significant improvement, reviewers preferred unpaired T2-MRCP (53%)
over unpaired EOB-MRC (17%) or paired tests (30%). Reviewers were willing to
pay $25 (median) for gadolinium-EOB.
CONCLUSION. Combining T2-MRCP and EOB-MRC significantly improved biliary visualization over each test alone. However, improvement was small, and the perceived added value of gadolinium-EOB was modest.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Recent development of liver-specific contrast agents with excretion into the biliary tree, such as mangafodipir trisodium (Teslascan; Nycomed Amersham, Princeton, NJ) or gadolinium -ethoxybenzyl-diethylenetriamine pentaacetic acid (gadolinium-EOB) (Eovist; Berlex Laboratories, Montville, NJ) has renewed interest in contrast-enhanced MR imaging of the biliary tree that may counter the limitations of standard MR cholangiopancreatography. Use of intrabiliary contrast media is likely to have the greatest impact on better visualization, particularly of undilated biliary systems, and may potentially increase reliability of the examination or decrease the occurrence of a nondiagnostic or equivocal interpretation. These gains in interpretive reliability come at an increased cost of MR imaging. To date, to our knowledge, information regarding the value of gadolinium-EOB in depiction of the biliary tree is unavailable in the literature. We present data comparing gadolinium-EOB—enhanced MR cholangiography (EOB-MRC) with the standard T2-weighted MR cholangiopancreatography (T2-MCRP). Our objectives were twofold: to assess the added efficacy of gadolinium-EOB in depicting biliary structures and to measure reviewer preference for the added value of biliary contrast media, including willingness to pay for it.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
MR Imaging Technique
A 1.5-T magnet was used for all imaging (Signa; General Electric Medical
Systems, Milwaukee, WI). The torso phased array coil was used for signal
reception. The MR cholangiopancreatographic portion of the examination
comprised two pulse sequences. Breath-held heavily T2-weighted single-shot
fast spin-echo MR cholangiopancreatography, representing the T2-MRCP, was
performed in the coronal plane and oblique projections, equivalent to the
radiographic right anterior oblique and left anterior oblique projections,
before administration of gadolinium-EOB. Sequence parameters for T2-MRCP were
the following: TR/effective TE, infinite/180; bandwidth, 31.2 kHz,; slice
thickness, 5 mm; matrix, 256 x 256; excitations with fat saturation,
0.5; average, 15 slices; average scan duration, 20-30 sec. A subsequent
breath-hold three-dimensional (3D) gadolinium-EOB—enhanced MR
cholangiogram was obtained 20 min after gadolinium-EOB administration
[4] using a spoiled
gradient-echo sequence: TR range/TE, 5-8/minimum; flip angle, 50°; 20
partitions 2.6-5.0 mm thick with interpolation to 40 slices 1.3-2.5 mm thick;
matrix, 256 x 128-160.
Image Processing and Analysis
Using a commercially available workstation (Advantage Windows, General
Electric Medical Systems), the T2-MRCP images were windowed and printed. The
EOB-MRC 3D data set was reconstructed in planes of imaging analogous to the
T2-MRCP using multiple operator-defined algorithms. The reconstructed images
were similarly windowed and printed. A single operator processed all
images.
We convened a panel of three radiologists with subspecialty training in body MR imaging. To limit reviewer bias, none of the reviewers were involved in the ongoing institutional gadolinium-EOB research protocol described previously. Images from the 10 unpaired examinations were arranged in a random order, and the three observers reviewed the images from 20 examinations. Subsequently, the 10 paired sets of T2-MRCP and EOB-MRC were arranged in a random order. After an 8-week delay, all members of the panel reviewed all the paired sets of images. All reviewers completed a standardized data sheet.
We divided the evaluation of the images into three domains: diagnostic quality, structure visualization, and added value of intrabiliary contrast media. For diagnostic quality, reviewers noted the presence or absence of artifacts and rendered a judgment as to whether the test was diagnostic or not.
Each reviewer evaluated the quality with which the structures of the biliary tree were depicted, which we termed "structure visualization." The following structures were evaluated: common bile duct, right and left hepatic ducts, and second-order division of the intrahepatic ductal branches. The quality of structure depiction was graded on a 5-point Likert-type scale, with 0 representing no visualization and 4 representing excellent visualization. Structure visualization was evaluated for the unpaired review of the T2-MRCP and EOB-MRC images and the paired review of T2-MRCP and EOB-MRC images.
The added value of intrabiliary contrast media indicated the degree to which using paired tests improved structure visualization over each of the unpaired tests. Measures of this domain included the improvement of structure visualization ratings, selection of the best single sequence for evaluating the biliary tree from the paired test review, subjective assessment of the degree of overall improvement added by the secondary sequence, and the willingness-to-pay. The improvement of structure visualization ratings was assessed by deriving the change in visualization ratings from the unpaired examination compared with the paired examination.
Determination of assessment of the degree of overall improvement added by the secondary sequence is illustrated by the following example: If the T2-MRCP was judged by a reviewer to be the best single sequence, then she or he subjectively assessed the degree of improvement in structure visualization that adding an EOB-MRC conferred. The degree of improvement was graded on a 5-point Likert-type scale: 0 for no improvement and 4 for excellent improvement.
Willingness-to-pay was determined at the time of the paired review by asking the reviewer to select a dollar value (range, $0-100) that she or he was willing to pay to obtain a paired examination for each patient evaluated. Because all patients had liver lesions that might become more conspicuous after gadolinium-EOB administration, we instructed the panel to ignore increased liver lesion visualization attributable to gadolinium-EOB when considering sequence preference willingness-to-pay, to isolate the value of gadolinium-EOB as an intrabiliary contrast agent.
Statistical Analysis
All statistical analyses were performed using Intercooled Stata 6.0
software (College Station, TX). Regarding analysis of structure visualization
data, we evaluated a weighted kappa statistic to measure interreviewer
agreement in structure visualization. Data were segregated and analyzed
according to test and anatomic structure. Serial weighted pair-wise
comparisons across all three reviewers were performed, and an average kappa
value across all comparisons was calculated. Typical values for defining
agreement on the basis of the kappa value are summarized in
Table 1. The Wilcoxon's signed
rank test evaluated the presence of a statistically significant improvement in
observer ratings comparing unpaired to paired test review, expressed as a
p value. A regression model constructed to control for fixed effects
due to reviewer and subject correlation quantified the improvement in ratings
attributable to addition of a second sequence, isolating the effects of the
paired review. Average improvement was expressed as the difference between the
paired test rating and the single test rating. Statistical significance was
expressed using the chi-square statistic.
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
Of the two unpaired tests, standard T2-MRCP and EOB-MRC, reviewers preferred T2-MRCP (53%; 95% confidence interval [CI], 34-72%) to EOB-MRC (37%; 95% CI, 20-56%); however, this preference is not statistically significant (Table 2). In 3% of the cases, no preference was expressed. In 7% of the cases, reviewers indicated that either T2-MRCP or EOB-MRC was acceptable.
|
Average visualization ratings (Table 3) are the following for unpaired review of EOB-MRC: common bile duct, 3.3; right hepatic duct, 2.7; left hepatic duct, 2.5; second-order division of the intrahepatic ducts (second-order branches), 1.4. Ratings for unpaired review of T2-MRCP were the following: common bile duct, 3.4; right hepatic duct, 1.8; left hepatic duct, 2.2; second-order branches, 0.9 (Figs. 1A,1B and 2A,2B,2C). The preceding visualization ratings improved using paired tests over EOB-MRC and T2-MRCP for all structures (p < 0.001) except for T2-MRCP evaluation of the common bile duct (p > 0.05). Overall, reviewer agreement for the degree of structure visualization was moderate to good (Table 4) with the exception of the common bile duct ratings by EOB-MRC.
|
|
If we controlled for reviewer and subject, paired review improved assessment of structure visualization over T2-MRCP alone (Table 3) an average of 1.05 (95% CI, 0.81-1.28%; p < 0.0001). Paired review improved visualization over EOB-MRC alone an average of 0.68 (95% CI, 0.49-0.86%; p < 0.0001). Overall, improvement due to paired review was modest. Although improvement using paired review was greater over T2-MRCP alone compared with EOB-MRC alone, the difference is not significant.
Interestingly, when asked which examination combination reviewers preferred in a clinical setting, unpaired T2-MRCP was chosen (53%; 95% CI, 34-72%) over unpaired EOB-MRC (17%; 95% CI, 6-35%) or paired tests (30%; 95% CI,14-49%). This preference was not statistically significant.
Because both measures of test preference showed the reviewers' preference for T2-MRCP, we used the reviewers' willingness-to-pay to obtain paired tests over unpaired tests as a proxy for willingness-to-pay for gadolinium-EOB. Despite a statistically significant, albeit small, improvement in reviewer visualization ratings, reviewers were willing to pay a median of $25 to obtain paired tests over unpaired tests, a modest amount. Examining the distribution of values, we found the two most commonly selected values were $0 and $50 (Fig. 3).
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In individuals without biliary stenosis or dilated biliary systems, intrahepatic biliary anatomy depiction and lesion detection are often inadequate [10]. Other centers investigating the use of MR cholangiopancreatography in choledocholithiasis and malignant strictures have reported better visualization of the extrahepatic ducts than of the intrahepatic ducts [3, 11]. Traditional two-dimensional imaging using turbo spin-echo sequences to generate the MR cholangiopancreatographic source data results in low spatial resolution [6], whereas 3D imaging using mangafodipir trisodium improves resolution by producing a voxel size of 1 x 2 x 1.5 mm. Similar improvements in spatial resolution using 3D acquisition after gadolinium-EOB administration give EOB-MRC an advantage over conventional T2-MRCP in visualizing smaller caliber ducts. To avoid bias toward EOB-MRC, we compared filmed reconstructions in planes analogous to the conventional T2-MRCP, although 3D imaging further allowed reconstruction of the source data in various obliquities. Postprocessing and reconstruction made transparent the orthogonal relationships between the right, left, and common hepatic ducts. These spatial relationships can be difficult to assess with certainty on two-dimensional imaging [10].
The strength of MR imaging rests on lesion conspicuity because of increased contrast-to-noise. Increasing the contrast between the biliary tree and the surrounding liver can be achieved with a variety of contrast agents. Gadolinium chelate, acting as a negative contrast agent on the standard MR cholangiopancreatographic sequence, improves the visualization of both extra- and intrahepatic biliary branches in patients with suspected malignant upper abdominal disease by suppressing hepatic parenchymal signal [12]. Other data suggest that mangafodipir trisodium facilitates definition of intrahepatic ductal anatomy in healthy individuals with nondilated hepatic ducts by acting as a positive biliary contrast agent [10]. In particular, use of intrabiliary contrast media like mangafodipir trisodium (and gadolinium-EOB) facilitates differentiation of hepatic vessels from bile ducts, which can be difficult on conventional T2-weighted imaging because of slow flow.
Potential diagnostic applications for gadolinium-EOB as an intrabiliary contrast medium include delineation of aberrant duct anatomy in potential liver donors similar to that previously described with mangafodipir trisodium and improved visualization of ductal anatomy and disease in individuals with smaller-than-normal ducts that have traditionally been difficult to visualize. This population includes patients with primary sclerosing cholangitis and infants and children. Further, the use of a hepatobiliary gadolinium-based MR contrast agent such as gadolinium-EOB can potentially provide a combined evaluation of the hepatic parenchyma and vasculature (performed during the dynamic-phase imaging) and the bile ducts (performed during delayed-phase imaging). Studies evaluating the utility of gadolinium-EOB regarding the detection of liver lesions are on-going [13]. Application of gadolinium-EOB as a biliary contrast agent has, so far, been limited to animal studies [14,15,16].
To our knowledge, our study represents the first evaluation of the clinical efficacy of gadolinium-EOB in improving visualization of the biliary tree. Our data support the finding that intrabiliary contrast media significantly improves visualization of biliary structures, increasing the visualization scores by an average of 0.86, regardless of the initial examination, with more improvement in conventional T2-MRCP visualization scores. Further, improvement was reasonably reproducible in all reviewers. In evaluating the unpaired examinations, the reviewers preferred conventional T2-MRCP. Curiously, despite greater improvement seen in the T2-MRCP ratings compared with the EOB-MRC ratings after paired review, on average, the reviewers continued to prefer T2-MRCP alone as the examination of choice, suggesting that the increased conspicuity of the biliary tree with gadolinium-EOB is insufficient to change reviewer preference. We hypothesize that the continued preference for conventional T2-MRCP may be due to familiarity with the technique, a priori knowledge that the patients studied were not suspected of having biliary disease, or the perceived lack of added value to administering gadolinium-EOB in the patients examined, although direct measurement of these domains remain to be performed.
As with any technologic development, implementation of gadolinium-EOB into clinical practice comes at an economic cost, either directly because of the increased cost of the new contrast agent or indirectly because of the cost of foregoing other opportunities. We, in essence, asked the reviewers to internally trade the cost of gadolinium-EOB against the perceived improvement in visualization conferred by the agent and asked them to assign a dollar value to this trade-off. Under standard contingent valuation, the reviewers selected the maximal amount they were willing to pay for the use value of gadolinium-EOB in improving ductal visualization in patients examined [17]. The dollar value selected is the "willingness-to-pay" for the contrast agent. Willingness-to-pay measures the reviewer's valuation of intrabiliary contrast and is theoretically capable of accounting for valuation of other characteristics of health care beyond health outcome [17]. This value includes the intangible attributes of the techniques under comparison (EOB-MRC and T2-MRCP), not otherwise measured in the quantitative analysis of structure visualization, for example, items of process utility such as familiarity with and ease of interpretation of conventional T2-MRCP or the relative desire to use the new contrast agent. The study design did not permit the evaluation of the intangible attributes of EOB-MRC and conventional T2-MRCP; however, exploration of these non—health-outcome-related characteristics of imaging techniques will be an interesting extension of this preliminary work.
The median value of willingness-to-pay is $25, an amount that is below the cost of current clinically available non—organ selective contrast agents such as gadopentetate dimeglumine or gadodiamide and presumably below the market cost of gadolinium-EOB, when made available. The most commonly selected values, however, were $0 and $50, suggesting the examinations appear to be segregated into two groups: the first, in which gadolinium-EOB had no added value and the second, in which gadolinium-EOB had some added value for which the reviewers were most commonly willing to pay $50. Nevertheless, the dollar amount may still be below the actual cost of the product and supports the supposition that the reviewers did not perceive significant incremental value to adding intrabiliary contrast media.
In our study, the results, that the study reviewers continued to prefer unpaired conventional T2-MRCP despite better biliary visualization with EOB-MRC and that the reviewers were willing to pay a median of $25, are complementary pieces of information. That the reviewers were willing to pay such a low sum for intrabiliary contrast media when forced to assign a dollar value to obtain paired examinations suggests that the quantitatively measured improved visualization is insufficient added value to justify higher expenditure for adding contrast media. The continued reviewer preference for unpaired conventional T2-MRCP despite statistically significant improvement in visualization further supports the hypothesis that the added visualization is insufficient to justify adding intrabiliary contrast media.
Although the foremost limitation of this study is the small number of patients who received gadolinium-EOB, review of the literature shows that our preliminary study represents the largest clinical population in whom the value of gadolinium-EOB as an intrabiliary contrast agent has been assessed. As with many small studies examining technologic advances, the expected bias is toward the new technique, in this case, EOB-MRC. Despite this expected bias, the reviewer preferences favor conventional T2-MRCP.
Regarding patient selection, we performed the study in a population with undilated ducts; the study cannot address the efficacy of gadolinium-EOB in individuals with dilated ducts. We did not include individuals with smaller-than-normal ducts, such as in patients with primary sclerosing cholangitis or in children with suspected biliary disease, and we cannot directly assess the effect of improved visualization on reviewer preference and willingness-to-pay in these other populations, although it has been argued that the true utility of gadolinium-EOB is in individuals with normal-sized ducts, the population we considered for the study. Further, animals with induced cholestasis after biliary ligation experienced prolonged hepatic enhancement and delayed biliary excretion of contrast material until development of biliary collaterals [16], suggesting a more limited usefulness of biliary contrast material in patients with obstructive biliary dilatation, particularly in the acute setting, in which a potential application of gadolinium-EOB imaging would yield functional information regarding the capability of the hepatic segments to excrete bile. Similarly, gadolinium-EOB may be used in patients with suspected abnormal hepatocyte function to assess biliary excretion.
The reviewers were not blinded to the two techniques that are compared. This circumstance may lead to reviewer bias. Blinding the reviewers to MR cholangiographic technique was not possible because of the characteristic appearance of the images derived from the EOB-MRC and T2-MRCP techniques. However, reviewer bias is expected toward incorporation of the newer technique rather than toward conventional T2-MRCP alone. In fact, we expected the reviewers to be internally consistent between their quantitative assessment of visualization and their assignment of test preference. That the expected direction of bias is away from the actual results suggests that we may be underestimating the reviewer preference for conventional T2-MRCP alone.
The study focused on the value of gadolinium-EOB strictly as a biliary agent. For that specific purpose, the added value is small. We likely underestimated the overall clinical value of gadolinium-EOB, a large part of which will be derived from its efficacy in gadolinium-enhanced dynamic imaging of the liver. Although Kim et al. [18] reported that the addition of gadolinium-enhanced dynamic imaging did not significantly increase accuracy in differentiation of benign and malignant causes of biliary dilatation, it is theoretically possible that individuals without known disease, such as those considering liver donation, may be evaluated with gadolinium-EOB in a comprehensive examination to delineate the liver parenchyma, vasculature, and biliary tree [19].
Although the concept is well established, willingness-to-pay as a measure of value has its limitations. Although the reviewers are estimating their own willingness-to-pay, they do not bear the risk of the choice and bear only some risk of the expense, either directly or indirectly. That is, the patient bears both risks and may value the use of the new contrast agent differently. Nevertheless, the perspective of the reviewer is an important one because she or he is ultimately responsible for the choice of contrast agent and therefore represents the target market. Additionally, the willingness-to-pay generated in this preliminary study may provide information for future pricing and purchasing policies once the contrast agent becomes available. We measured the hypothetic willingness-to-pay, which in general, significantly overestimates the real willingness-to-pay [20]. This hypothetic bias further reduces the as yet unmeasured actual willingness-to-pay for gadolinium-EOB.
In conclusion, the combination of images obtained from standard T2-weighted MRC and from EOB-MRC conferred significant improved visualization of the biliary tree over each of the individual tests alone. Nevertheless, the perceived added value of gadolinium-EOB as an intrabiliary agent was modest. However statistically significant, the degree of improvement is not adequate to change reviewer preference away from the standard of care for biliary imaging, namely conventional T2-weighted MRCP. Our results suggest that the standard of care appears to be sufficient for diagnostic biliary ductal visualization and may reassure practices that cannot or will not undertake the expense of intrabiliary contrast media. Our study represents the preliminary assessment of the value radiologists place on gadolinium-EOB as an intrabiliary contrast agent and provides the setting for further debate.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  D. A. Bluemke, D. Sahani, M. Amendola, T. Balzer, J. Breuer, J. J. Brown, D. D. Casalino, P. L. Davis, I. R. Francis, G. Krinsky, et al. Efficacy and Safety of MR Imaging with Liver-specific Contrast Agent: U.S. Multicenter Phase III Study Radiology, October 1, 2005; 237(1): 89 - 98. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. M. Yeh, R. S. Breiman, B. Taouli, A. Qayyum, J. P. Roberts, and F. V. Coakley Biliary Tract Depiction in Living Potential Liver Donors: Comparison of Conventional MR, Mangafodipir Trisodium-enhanced Excretory MR, and Multi-Detector Row CT Cholangiography--Initial Experience Radiology, March 1, 2004; 230(3): 645 - 651. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
