| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Pediatric Imaging |
1 From the Departments of Radiology (T.S.J., Y.T.K., G.C.L., C.K.W.) and Pediatric Surgery (S.H.C.), Kaohsiung Medical College, 100 Shih-Chuan First Rd, Kaohsiung 807, Taiwan, Republic of China. From the 1997 RSNA scientific assembly. Received April 29, 1998; revision requested June 16; final revision received December 15; accepted January 19, 1999. Address reprint requests to T.S.J. (e-mail: m810101@cc.kmc.edu.tw).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: MR cholangiography was performed on 10 control and 16 jaundiced neonates and infants aged 3 days to 5 months. Diagnosis of biliary atresia (n = 6) was confirmed with surgery and liver biopsy, with or without surgical cholangiography. Diagnosis of neonatal hepatitis (n = 9) was confirmed with clinical follow-up until jaundice resolved. In one infant, paucity of intrahepatic ducts was diagnosed at liver biopsy. MR cholangiography was performed with respiratory-triggered, heavily T2-weighted turbo spin-echo and optional inversion-recovery turbo spin-echo sequences. Diagnosis of biliary atresia was based on nonvisualization of either the common bile duct or common hepatic duct. Cholescintigraphy with technetium 99m disofenin was performed in all 16 jaundiced patients.
RESULTS: In the 10 controls, the nine patients with neonatal hepatitis, and the one infant with paucity of intrahepatic ducts, MR cholangiography clearly depicted the gallbladder and common hepatic and common bile ducts. MR cholangiography was 100% accurate in excluding biliary atresia as the cause of neonatal cholestasis, while 99mTc disofenin cholescintigraphic findings were false-positive in four of 10 patients with nonobstructive cholestasis.
CONCLUSION: MR cholangiography can be used to depict the major biliary structures of neonates and small infants and to exclude biliary atresia as the cause of neonatal cholestasis by allowing visualization of the biliary tract.
Index terms: Bile ducts, abnormalities, 768.1434 • Bile ducts, MR, 768.121411, 768.121413, 768.121415, 768.121416 • Bile ducts, radionuclide studies, 768.12172 • Gallbladder, 762.1434 • Infants, 762.1434
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The control group included seven male and three female neonates and infants, aged 3 days to 5 months (mean age, 106 days). Sixteen consecutive neonates and infants who had neonatal cholestatic jaundice were included in this prospective study. They included 11 male and five female neonates and infants, aged 15–65 days (mean age, 46 days). Clinically, these patients had clay-colored stool and prolonged conjugated hyperbilirubinemia that occurred within the first 2 months of life (9). Prolonged conjugated hyperbilirubinemia was defined as a total serum bilirubin level of more than 3.0 mg/dL (51 µmol/L), with the direct form more than 40% of the total (normal values, 0.2–1.0 mg/dL [3.4–17 µmol/L] and 0–0.2 mg/dL [0–3.4 µmol/L], respectively). Most of the jaundiced patients also had an increased serum 
-glutamyltransferase level (normal values, 3–30 U/L) (Table).
|
|
The findings from liver biopsy revealed that four patients with biliary atresia showed typical histopathologic features, with bile duct proliferation, bile plugs, and periportal fibrosis. In the other two patients with biliary atresia, the findings from biopsy showed giant cell transformation and canalicular and cellular bile stasis, without definite bile duct proliferation or periportal fibrosis. In one patient with neonatal hepatitis, the findings from biopsy showed giant cell transformation, inflammatory cell infiltration, and lobular disorganization. The results of liver biopsy in the patient with paucity of intrahepatic ducts showed no interlobular bile duct in several portal areas.
In the nine patients with idiopathic neonatal hepatitis, the diagnosis was verified by the patient's recovery from jaundice and the normalization of laboratory values during the clinical follow-up period. The duration of follow-up was 3–12 months. In the patient with paucity of intrahepatic ducts, azotemia and imaging findings of infantile polycystic kidney disease were noted in addition to the cholestasis.
Imaging
All patients had fasted for at least 3 hours before the MR examination and were sedated with rectal administration of secobarbital (5–10 mg per kilogram of body weight) or with oral or rectal administration of chloral hydrate (50 mg/kg). MR examination was performed with a 1.5-T magnet (Gyroscan ACS-NT; Philips Medical Systems, Best, the Netherlands). A flexible surface coil (E1; rectangular and measuring 16 x 13.5 cm) was wrapped around the upper portion of the abdomen of the neonates and infants.
MR imaging sequences included a coronal T2-weighted, spectral presaturation with inversion recovery (IR), turbo spin-echo (SE) sequence (1,650/80 [repetition time msec/effective echo time msec]) with an echo train length of 14; an axial T2-weighted, turbo SE sequence (1,650/100 [effective]) with an echo train length of 18; and an axial T1-weighted turbo gradient-echo sequence (15/5; flip angle, 25°). MR cholangiography was performed with a two-dimensional (2D), non–breath-hold, respiratory-triggered, heavily T2-weighted, fat-suppressed, turbo SE sequence (3,000/700 [effective]) with an echo train length of 128. Coronal source images were obtained with a 2-mm section thickness and a 1-mm overlap. The number of signals acquired was six. We used a 195 x 256 matrix. In addition, a three-dimensional (3D) IR turbo SE sequence (7,000/1,000 [effective]/50 [repetition time msec/effective echo time msec/inversion time msec]) with an echo train length of 116 was also used in five control subjects, six neonates and infants with biliary atresia, and one infant with paucity of intrahepatic ducts. The number of signals acquired was two.
The 2D or 3D acquisition images were reformatted by using a standard maximum intensity projection (MIP) algorithm to create a rotating display. Imaging time for MR cholangiography varied from 6 minutes 11 seconds to 12 minutes 36 seconds. The total imaging time ranged from 20 to 30 minutes.
Cholescintigraphy with 99mTc disofenin was also performed with a gamma camera scanner (GCA 602A; Toshiba, Nasu, Japan) in all 16 patients within a week of MR examination. Phenobarbital (5 mg/kg/d) was given to each patient for at least 5 days before scintigraphic examination. The neonates and infants had fasted for 3 hours before the examination. Cholescintigraphy was performed with intravenous administration of 99mTc disofenin in a dose of 0.25 mCi/kg (9.25 MBq/kg). Sequential anterior scintigrams of the abdomen were obtained at 1, 5, 15, and 30 minutes after injection and at 1, 2, 4, and 6 hours. Delayed images were obtained at 24 hours if there was still no visualization of bowel activity by 6 hours (11). The hepatic uptake of radiotracer, the depiction of the gallbladder, and the presence of activity in the bowel were observed on the serial scans within 6 hours and on a 24-hour–delay scan.
Image Analysis
All MR images were interpreted independently by three radiologists (Y.T.K., C.K.W., G.C.L.) who were experienced in abdominal MR imaging and were blinded to the clinical history and the results of other tests of the patients. The conspicuity or visibility of the common bile duct, common hepatic duct, gallbladder, cystic duct, and intrahepatic ducts was evaluated on the basis of the source and MIP-reformatted images. Discrepancies among observers were resolved by consensus.
For the patients with neonatal cholestasis, biliary atresia was excluded if there was visualization of a normal extrahepatic biliary system at MR cholangiography. In the patients with biliary atresia, MR findings were correlated with those found at surgery. The studies of the 10 control patients, who underwent MR imaging only, were interpreted in the same manner as those for the 16 jaundiced patients. The cholescintigraphic studies were interpreted in the same fashion as the MR images. The interpretation of the MR images and cholescintigraphic studies was separated by 1–2 weeks to prevent recall of the findings, and the patient order in which those two types of studies were interpreted was different.
Statistical Analysis
The McNemar exact test was used to compare the visibility of intrahepatic ducts between 2D turbo SE and 3D IR turbo SE images in five control pairs. The Student t test was used to evaluate the statistically significant difference in the mean gallbladder size between patients with neonatal hepatitis and those with biliary atresia. The Fisher exact test was used to test the significance of a small gallbladder on the MR cholangiogram in the differentiation of biliary atresia from other causes of neonatal cholestasis. A P value of less than .05 was considered to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
The neonates and infants with neonatal hepatitis had gallbladders ranging from 1.4 to 4.5 cm in length (mean ± SD, 2.9 cm ± 0.9), while the gallbladders in the neonates and infants with biliary atresia were 0.8–2.1 cm in length (mean ± SD, 1.2 cm ± 0.4) (Student t test, t = 4.33, P < .001). MR cholangiography showed small gallbladders in all patients with biliary atresia except one and showed a small gallbladder in only one of the 10 patients without biliary atresia (Fisher exact probability test, P < .001). Thus, a small gallbladder could be considered highly suggestive but not diagnostic of biliary atresia. The diagnostic accuracy was 88% (14 of 16 patients), with a sensitivity of 83% (five of six patients), specificity of 90% (nine of 10 patients), and positive and negative predictive values of 83% (five of six patients) and 90% (nine of 10 patients), respectively.
Periportal fibrosis, as defined by the presence of increased T2 signal intensity on the T2-weighted conventional MR images, was seen in four of the six patients with biliary atresia (Fig 7) and in none of the nonjaundiced cohort and none of the patients with nonobstructive cholestasis. However, periportal thickening was hardly appreciable on MIP-reformatted MR cholangiographic images.
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
With these images, biliary atresia can be definitely excluded if the entire extrahepatic bile duct is visualized on the MR cholangiograms. If no complete extrahepatic bile duct can be visualized, especially when accompanied by an atrophic gallbladder, biliary atresia is suggested.
Biliary atresia and neonatal hepatitis are the two major causes of cholestatic jaundice in neonates and young infants (13). Biliary atresia was theorized to result from destructive inflammatory processes, leading to progressive fibrosis and obliteration of the extrahepatic bile duct (1). Patients with biliary atresia require an early surgical intervention, at best before 2 months of age (14). Many tests have been used to evaluate cholestasis in infants; however, none has proved infallible (9). A percutaneous liver biopsy has an extremely variable diagnostic accuracy of 60%–96.8% (9,13,15,16). Imaging studies, including ultrasonography, hepatobiliary scintigraphy, and endoscopic retrograde cholangiopancreatography, have been used to exclude biliary atresia as the cause of neonatal cholestasis (3–6,17–20). Use of endoscopic retrograde cholangiopancreatography to view the biliary tree presents a challenge in infants because of the lumen-to-tube ratio (21). Although laparotomy with surgical cholangiography can provide a direct and definite diagnosis, it is undesirable if alternative techniques are available.
MR cholangiography has developed rapidly over the past few years. Promising results in adult biliary disorders have been well documented in the literature (22–24). The feasibility and preliminary applications of MR cholangiography in neonates and infants have been mentioned in two recent articles (7,8). In our study, with the use of a non–breath-hold, respiratory-triggered, heavily T2-weighted, fat-suppressed, turbo SE technique, the main biliary structures were depicted in all control neonates and infants younger than 5 months of age. The second-order intrahepatic ducts could be visualized in half of the control group. This reflected that the resolution and quality of MR cholangiograms could ensure a clinical application for neonates and small infants.
The technique we used had several advantages: (a) higher signal-to-noise and contrast-to-noise ratios, which allow the use of thin sections, even for 2D imaging, and (b) diminished sensitivity to both motion artifact and slow flow (24). These advantages are mandatory for imaging neonates or small infants who have small-caliber biliary ducts and unavoidable motion. In addition, the respiratory-triggering technique allowed a non–breath-hold examination, which made imaging a baby possible.
These non–breath-hold fast SE techniques rely on an increase in signal acquisition (up to six signals acquired) to increase the signal-to-noise ratio and to compensate for signal losses caused by motion (25,26). The increase in the number of signals acquired, however, results in a long acquisition time, which should be minimized by choosing an adequate echo train length. Applying a surface coil instead of a body coil increases the signal-to-noise ratio and, therefore, improves visualization of small nondilated ducts (27). The addition of a fat-saturation technique has also improved the conspicuity of bile ducts from the surrounding intraabdominal fat while decreasing the motion artifacts associated with hyperintense subcutaneous fat (24).
In our control group, the 3D IR turbo SE images were slightly inferior to the 2D turbo SE images, especially in depicting the intrahepatic ducts. In the jaundiced cohort, 3D IR turbo SE imaging was performed only in the six patients with biliary atresia and the infant with paucity of intrahepatic ducts. Because none or only a few bile ducts were depicted in these patients, it is hard to compare these two sequences in the jaundiced group. In general, 3D IR turbo SE images proved to be slightly inferior in quality compared with the 2D turbo SE images. The increased motion artifacts caused by thin sections, the limitation of the number of signals acquired, and the long acquisition time degrade the visualization of the small-caliber intrahepatic ducts in small infants imaged with 3D IR turbo SE sequences. Because of the relatively higher quality and the shorter imaging time, we recommend 2D turbo SE sequences for MR cholangiographic study in small infants.
In this study, the extrahepatic bile ducts were visualized well in the nine patients with neonatal hepatitis and the infant with paucity of intrahepatic ducts. Biliary atresia was thereby excluded, and an unnecessary laparotomy was avoided in the four patients who had no evidence of hepatobiliary excretion at cholescintigraphy.
MR cholangiograms of all of the six patients with biliary atresia did not show the extrahepatic bile duct, except for one patient with cystic dilatation of the common hepatic duct. In the two patients with type B biliary atresia, a patent but very thin common bile duct with an atretic common hepatic duct was disclosed with surgical cholangiography. The lack of adequate water content in the hypoplastic common bile duct and its extremely small caliber (<0.5 mm) made it unable to be visualized on MR cholangiograms. Nevertheless, this did not affect the diagnosis of biliary atresia from a clinical point of view, nor did it have an influence on the choice of surgical procedures or the predicted outcome. In four patients with biliary atresia, MR cholangiograms also did not show the intrahepatic ducts. This was consistent with the surgical and histopathologic findings on the sections through the porta hepatis, which showed only a fibrous cord of the common hepatic duct and prominent fibrotic tissues with small bile ductules without normal structures of hepatic ducts.
Guibaud et al (7) have reported a false-positive diagnosis of biliary atresia in an infant with sclerosing cholangitis, in whom MR cholangiography did not depict the extrahepatic bile duct because of the small caliber of the ducts. Someday, with further improvements in the imaging resolution of MR cholangiography, visualization might be possible in cases such as that of the previous patient with sclerosing cholangitis and hypoplastic common bile ducts in type B biliary atresia. At such time, a reference to the normal caliber of the common bile duct (mean, 2.6 mm ± 0.5; range, 2.0–3.5 mm) will be necessary.
Guibaud et al (7) described periportal thickening on MR cholangiograms in all three of their patients with biliary atresia. Hyperintense periportal thickening was also found on T2-weighted MR images in our four patients with biliary atresia. However, periportal thickening was hardly appreciable on heavily T2-weighted MIP MR cholangiograms. The use of too heavily T2-weighted sequences and the MIP algorithm may suppress the signal intensity of periportal fibrosis.
Periportal thickening corresponded to the histologic findings of expansion of the portal tract by varying amounts of fibrosis, bile duct proliferation, and mixed inflammatory infiltrates (28). In the two patients with biliary atresia without MR findings of periportal thickening, no periportal fibrosis was noted at histopathologic examination of the specimen from liver biopsy. No periportal thickening was seen in our patients with nonobstructive cholestasis. This could be caused by the fact that progression of periportal fibrosis in intrahepatic cholestasis is not as prominent as that encountered in biliary atresia (29). Therefore, periportal thickening could also be a useful MR finding of biliary atresia. Larger prospective studies are required to determine the accuracy of periportal thickening in the diagnosis of biliary atresia.
In this prospective study of consecutive neonates and infants with neonatal cholestasis, the diagnostic accuracy of MR cholangiography was 100% (16 of 16 patients), while the accuracy of cholescintigraphy with 99mTc disofenin for the diagnosis of biliary atresia was only 75%, with a specificity of only 60%. The results of our scintigraphic studies are consistent with those from previous reports, which have revealed excellent sensitivity but low specificity in depicting biliary atresia (3–5). The liver uptake and excretion of the 99mTc-labeled radiopharmaceuticals are affected by the serum bilirubin levels and hepatic function of the patients (30). Therefore, several other causes of neonatal cholestasis may produce findings on cholescintigrams similar to those of biliary atresia (ie, lack of visualization of the gallbladder and of bowel activity, with retention of tracer in the liver by 24 hours), decreasing the specificity of the findings.
Guibaud et al stated, "The conditions that mimic biliary atresia on scintigraphy will also mimic biliary atresia on MR cholangiography" (7, p 30). However, in the present study, MR cholangiography depicted the extrahepatic bile ducts in three patients with neonatal hepatitis and in the infant with paucity of intrahepatic ducts, all of whom showed neither the gallbladder nor bowel activity at cholescintigraphy. Unlike cholescintigraphy, MR cholangiography is not limited by serum bilirubin levels or the hepatic function of the patients.
MR cholangiography depicts fluid in the bile duct, rather than only bile. In healthy individuals, mucus-producing glands located in the deep layers of the extrahepatic bile ducts connect with the lumen by long ducts and secrete white viscous secretions (31). Although bile secretion is retarded in severe intrahepatic cholestasis, this white viscous secretion, together with that of the neck glands of the gallbladder, might contribute to the visualization of the extrahepatic bile ducts with MR cholangiography.
Several major drawbacks still need to be addressed. The spatial resolution and signal-to-noise ratio need to be improved. There are problems of motion induced by respiration, by bowel peristalsis, or especially by inadequate sedation, which is also the major cause of a failed examination. Appropriate sedation is mandatory. The presence of fluid within the regional gastrointestinal tract may obscure the bile ducts, although some fluid in the duodenum can serve as a good anatomic landmark. Adequate fasting or administration of negative contrast agents may reduce this bowel interference.
In conclusion, MR cholangiography can be used to depict the major biliary structures of neonates and small infants, and it can be used to accurately exclude biliary atresia as the cause of neonatal cholestasis by allowing visualization of the biliary tract. Further efforts are needed in improving the image resolution and in obtaining more clinical experience to investigate the accuracy of MR cholangiography in excluding biliary atresia.
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, T.S.J.; study concepts, T.S.J.; study design, T.S.J., Y.T.K.; definition of intellectual content, T.S.J.; literature research, T.S.J., Y.T.K.; clinical studies, T.S.J., S.H.C.; data acquisition, T.S.J., Y.T.K.; data analysis, T.S.J., Y.T.K., C.K.W., G.C.L.; statistical analysis, T.S.J.; manuscript preparation, T.S.J.; manuscript editing, T.S.J., G.C.L.; manuscript review, G.C.L.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  W. S. Kim, J.-E. Cheon, B. J. Youn, S.-Y. Yoo, W. Y. Kim, I.-O. Kim, K. M. Yeon, J. K. Seo, and K.-W. Park Hepatic Arterial Diameter Measured with US: Adjunct for US Diagnosis of Biliary Atresia Radiology, November 1, 2007; 245(2): 549 - 555. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. K. Ryeom, B. H. Choe, J. Y. Kim, S. Kwon, C. W. Ko, H. M. Kim, S. B. Lee, and D. S. Kang Biliary Atresia: Feasibility of Mangafodipir Trisodium-enhanced MR Cholangiography for Evaluation Radiology, April 1, 2005; 235(1): 250 - 258. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H E Woodley Imaging of the jaundiced child Imaging, September 1, 2004; 16(4): 301 - 313. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.-J. Lee, S.-M. Lee, W.-H. Park, and S.-O. Choi Objective Criteria of Triangular Cord Sign in Biliary Atresia on US Scans Radiology, November 1, 2003; 229(2): 395 - 400. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Can Computed Tomography Predict Functional Benefit from Lung Volume Reduction Surgery for Emphysema? Am. J. Respir. Crit. Care Med., December 15, 2001; 164(12): 2137 - 2138. [Full Text] [PDF]
|
|
|
|
|
|
M.-J. Kim, Y. N. Park, S. J. Han, C. S. Yoon, H. S. Yoo, E. H. Hwang, and K. S. Chung Biliary Atresia in Neonates and Infants: Triangular Area of High Signal Intensity in the Porta Hepatis at T2-weighted MR Cholangiography with US and Histopathologic Correlation Radiology, May 1, 2000; 215(2): 395 - 401. [Abstract] [Full Text]
|
|
|
|
|
|
K. I. Norton, R. B. Glass, D. Kogan, J. S. Lee, S. Emre, and B. L. Shneider MR Cholangiography in the Evaluation of Neonatal Cholestasis: Initial Results Radiology, March 1, 2002; 222(3): 687 - 691. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
