Magnetic Resonance Cholangiography: Comparison With Endoscopic Retrograde Cholangiopancreatography

GASTROENTEROLOGY 1996;110:589 597 Magnetic Resonance Cholangiography: Comparison With Endoscopic Retrograde Cholangiopancreatography JORGE A. SOTO,* MATTHEW A. BARISH,* E. KENT YUCEL,* DAVID SIEGENBERG, JOSEPH T. FERRUCCI,* and RAM CHUTTANI Departments of *Radiology and Medicine, Boston University Medical Center, Boston, Massachusetts detection of biliary ductal dilatation, but results in the diagnosis of choledocholithiasis are conflicting, with sen- sitivities ranging from 71% 6 to 100%. 5 The capability of MRC to complement or replace diagnostic ERCP and PTC has not been explored. Three-dimensional fast spin-echo (3D FSE) MRC yields better quality projectional images than similar sequences obtained with two-dimensional fast spin-echo (2D FSE) used in other studies. 5,6 In a previous report, 7 we described our initial experience with 3D FSE MRC in the evaluation of both the biliary and pancreatic ducts. The purpose of this study was to determine the sensitivity and specificity of 3D FSE MRC for the evaluation of the biliary tree, as determined by direct cholangiography. To accomplish this, we performed a prospective, blinded comparison in 46 patients referred to Boston University Medical Center for elective ERCP. Background & Aims: Magnetic resonance cholangiography (MRC) is a noninvasive diagnostic modality capable of producing high-quality images of the biliary tree. The purpose of this study was to determine in a prospective, blinded fashion the sensitivity and specificity of three-dimensional fast spin-echo (3D FSE) MRC for the evaluation of biliary tract abnormalities. Methods: Forty-six patients referred for elective direct cholangi- ography (45 endoscopic retrograde cholangiopancreatography and 1 percutaneous transhepatic cholangiog- raphy) were studied prospectively with 3D FSE MRC during a 1-year period. All images were interpreted blindly by two radiologists. The presence of dilatation, strictures, and intraductal abnormalities was recorded. Sensitivity and specificity of 3D FSE MRC were determined using findings on direct cholangiography as the gold standard. Results: MRC images of diagnostic quality were obtained in 44 (95.7%) of the patients. Sensitivity for the detection of bile duct dilatation (n Å 27), biliary strictures (n Å 10), and intraductal abnormalities (n Å 7) was 96.3%, 90%, and 100%, respectively. In addition, the MRC showed 16 of 17 patients with normal bile ducts (specificity, 94.1%). Conclusions: MRC has a very high sensitivity and specificity in the evaluation of the biliary tract. Based on these data, we believe that the efficacy of MRC using 3D FSE is sufficient to warrant its use in the routine diagnosis of biliary tract disease. M agnetic resonance cholangiography (MRC) is a relatively new imaging technique for the evaluation of the biliary tract. 1 6 Projectional images, similar in appearance to direct cholangiograms produced through endoscopic retrograde cholangiopancreatography (ERCP) or percutaneous transhepatic cholangiography (PTC), are obtained without administration of any oral or intravenous contrast agent. Its completely noninvasive nature makes it an appealing modality for the visualization of all segments of the biliary tract. Initial reports using various magnetic resonance se- quences have described the findings on MRC in patients with obstructive jaundice 1 4,6 and choledocholithiasis. 5,6 These early studies have shown a high accuracy in the Materials and Methods Patient Population Forty-six patients (15 male and 31 female) were examined prospectively with 3D FSE MRC at our institution during a period of 1 year (May 1994 to April 1995). The patients were randomly recruited from referrals for elective ERCP. The mean age was 51.4 years (range, 17 97 years). All patients underwent direct cholangiography (45 ERCP and 1 PTC) within 24 hours after the completion of MRC. The final diagnoses in the study population are listed in Table 1. Fifteen patients (32.6%) had undergone previous cholecystectomy. Malignant obstruction was confirmed in 5 patients by cytology (n Å 3) or surgical findings and pathology (n Å 2). Benign strictures were diagnosed based on imaging findings and clini- cal follow-up. Sphincter of Oddi dysfunction was diagnosed based on cholangiographic findings, namely, bile duct dilata- tion with delayed emptying and abnormal pressure of the Abbreviations used in this paper: CBD, common bile duct; ERCP, endoscopic retrograde cholangiopancreatography; MR, magnetic resonance; MRC, magnetic resonance cholangiography; PTC, percutaneous transhepatic cholangiography; 3D FSE, three-dimensional fast spin-echo; 2D FSE, two-dimensional fast spin-echo. 1996 by the American Gastroenterological Association 0016-5085/96/$3.00

590 SOTO ET AL. GASTROENTEROLOGY Vol. 110, No. 2 Table 1. Diagnoses in Patient Population Image Interpretation Final diagnosis No. of patients The MRC images were evaluated independently by two radiologists with experience in biliary tract imaging. No Normal 17 Ampullary stenosis 12 clinical information or findings of other imaging studies were Choledocholithiasis 6 provided. The maximum intensity pixel reconstructions alone Chronic pancreatitis with stricture 4 were given for interpretation; the source images were provided Pancreatic head adenocarcinoma with stricture 2 only on request to confirm the presence of intraductal abnor- Mid-CBD cholangiocarcinoma 1 malities. Differences in interpretation were settled by consen- Periportal adenopathy 1 Sclerosing cholangitis 1 sus between the radiologists. Common bile duct polyp 1 MRC images were initially assessed for overall quality and Intrahepatic cholangiocarcinoma 1 visualization of the intrahepatic and extrahepatic ductal sys- Total 46 tems. If no bile ducts were visualized, the study was considered nondiagnostic. If the biliary tree was visualized, the images were then evaluated for the presence of common bile duct (CBD) dilatation. The CBD was considered dilated if it measphincter of Oddi (ú40 mm Hg at baseline). A case of bile sured more than 6 mm, except in postcholecystectomy paduct polyp was diagnosed endoscopically. In addition to diag- tients, in whom the upper limit of normal was set at 8 mm. nostic cholangiography, 31 patients (67.4%) had one or more In cases of dilatation, the degree of dilatation was graded interventional procedures performed at the time of ERCP: subjectively as mild, moderate, or severe. The level of obstrucsphincterotomy (n Å 17), sphincterotomy and stone extraction tion was determined and divided into suprapancreatic, intra- (n Å 6), stricture stenting (n Å 4), brushings for cytology pancreatic, and periampullary. The presence and number of (n Å 5), and bile duct polypectomy (n Å 1). The study was stones in the CBD were recorded. Finally, the intrahepatic approved by our institutional review board, and informed con- biliary tree was evaluated for the visualization and presence of sent was obtained from each patient before enrollment. Seven dilatation, determined subjectively as being present or absent. patients were included in an early report which described the For each patient, all direct cholangiography radiographs technique used herein. 7 obtained were provided for interpretation. Duct caliber was MRC determined using the known diameter of the endoscope as a correction factor for magnification. The same criteria described The MR cholangiograms were acquired using commercially above for MRC were used for the evaluation and scoring of available software in a clinical MR scanner (Philips ACS direct cholangiograms. The interpretation of the direct cholan- 1.5T Gyroscan, Shelton, CT) using a body coil. The 3D FSE giography images was used as the standard of reference for the sequence was obtained in the coronal plane with a multislab determination of sensitivity and specificity of MRC. acquisition mode: 10 slabs (each 1.8-cm thick) with a 40% overlap between adjacent slabs and 50 partitions with a partition Results thickness of 2 mm. Other imaging parameters were repetition Image Quality time of 5000 milliseconds, echo time of 240 milliseconds, 1 signal average, echo train length of 31, echo spacing of 15 milliin 44 of the 46 subjects (95.7%). In the remaining 2 MRC studies of diagnostic quality were obtained seconds field of view of 240 mm, and matrix 186 1 256. Chemipatients, the motion artifact significantly degraded the cal selective fat suppression and respiratory triggering were used. Respiratory triggering decreases the respiratory motion artifact quality of the images. In 1 patient ERCP revealed a and improves the quality of the images, the main drawback being normal biliary tree and in the other patient a short stric- an increase in scanning time, from a nominal scan time of 10:30 ture involving the mid-cbd with moderate dilatation (minutes:seconds) to an actual scan time that varies between 11:30 of the biliary tree. Final diagnosis in this patient was and 16:00, depending on the respiratory rate. cholangiocarcinoma. To obtain the projectional images, the raw data (or source images) acquired with this sequence were later postprocessed at a Normal Bile Ducts separate work station using a maximum intensity pixel projection MRC correctly identified normal-caliber CBD and algorithm. The complete set of raw data was first carefully recommon hepatic and intrahepatic bile ducts in 16 patients. viewed, and the volume used for the three-dimensional recon- An example of a typical MRC in a patient with normal structions was limited to the region and images that contained pertinent information regarding the biliary tree (targeted reconhad undergone previous cholecystectomy. In 2 of them, bile ducts is shown in Figure 1. Seventeen patients (37%) struction). This eliminated many overlapping fluid-containing organs (e.g., bowel, kidneys, etc.) that would have otherwise surgical clips caused imaging artifacts, seen as periductal obscured the bile duct on the different projections. The final data foci of signal dropout. Review of the source images by the set was reconstructed at 13 15 of obliquity, producing 12 14 radiologists confirmed the extraductal source of the signal projectional images, including a frontal and a lateral view. void and avoided misinterpretation for intraductal abnor-

February 1996 MR CHOLANGIOGRAPHY: COMPARISON WITH ERCP 591 Figure 1. Normal biliary tree. MRC in a 36-year-old patient with nonspecific abdominal pain clearly shows the normal-caliber intrahepatic ducts (short arrow) and common bile duct (long arrow). The normal gallbladder and cystic duct (open arrow), pancreatic duct (arrowhead), and duodenum (curved arrow) are also well shown. malities. In summary, MRC showed 16 of the 17 patients with normal CBD (specificity, 94.1%). Biliary Ductal Dilatation MRC correctly showed the presence of bile duct dilatation and the site of the obstruction in 26 of the 27 patients (96.3%) shown to have dilatation of the CBD and intrahepatic ducts on direct cholangiography. Dilatation was properly scored as mild in 11 patients, moderate in 10, and severe in 6. However, in 2 cases the degree of dilatation was underestimated; both had sphincter of Oddi dysfunction and moderate dilatation on ERCP, considered mild on MRC. The site of the obstruction was periampullary in 19 patients (see Table 1): sphincter of Oddi dysfunction (n Å 12) (Figure 2), choledocholithiasis with impacted stone (n Å 6) (Figure 3), and distal CBD polyp (n Å 1). Intrapancreatic obstruction was present in 6 patients: 4 strictures secondary to chronic pancreatitis (Figure 4) and 2 malignant strictures (Figure 5) due to ductal Figure 2. Sphincter of Oddi dysfunction. A 44-year-old woman with adenocarcinoma. There were 2 cases of suprapancreatic obthe common bile duct (short arrow) and pancreatic duct (long arrow). mild abdominal discomfort. (A) MRC showing moderate dilatation of struction of the CBD, one due to periportal lymphadenopathy Both ducts taper distally to the level of the ampulla. (B) ERCP confirms from endometrial carcinoma (Figure 6) and the other one the presence of biliary ductal dilatation; sphincter pressure was 65 caused by a cholangiocarcinoma. One patient had focal dilata- mm Hg, confirmatory of sphincter dysfunction.

592 SOTO ET AL. GASTROENTEROLOGY Vol. 110, No. 2 tion of the left intrahepatic biliary ductal system, caused by a cholangiocarcinoma encasing the distal left hepatic duct. MRC correctly depicted the dilated intrahepatic duct segment and the normal extrahepatic biliary tree in this patient. The 1 patient in whom MRC did not show biliary ductal dilatation had a mid-cbd cholangiocarcinoma with dilatation of the biliary ductal system proximal to the obstruction shown by ERCP; the MRC images in this case were of poor quality and no interpretation was rendered. Overall, MRC showed dilatation of the bile ducts, along with the site of the obstruction in 27 of 28 patients (sensitivity, 96.4%). Biliary Strictures MRC correctly showed the presence, location, and length of the strictures in 9 of 10 cases (sensitivity, 90%). The 10 cases of strictures involving the biliary tree shown by direct cholangiography are included in Table 1. Six were intrapancreatic: 4 secondary to chronic pancreatitis (Figure 4) and 2 due to pancreatic head carcinoma (Figure 5). Three were suprapancreatic and involved the CBD: 2 were mid- CBD strictures, 1 caused by metastatic periportal lymphadenopathy from endometrial carcinoma (Figure 6) and the other caused by cholangiocarcinoma, and 1 was a long stricture sparing only the more distal CBD, secondary to sclerosing cholangitis. The other patient with a biliary stricture shown on MRC had a cholangiocarcinoma encasing the left hepatic duct; the extrahepatic biliary tree was normal. In the remaining patient with biliary stricture (CBD cholangiocarcinoma), the quality of the MRC was suboptimal and the biliary tree was not visualized, as described previously. The strictures were seen on MRC as focal stenotic segments with proximal dilatation in 8 patients. In 1 patient, a diffusely narrowed CBD was seen as a long segment of signal dropout, with a normal-caliber duct proximal and distal to it. There was no intrahepatic ductal dilatation, and the correct diagnosis (sclerosing cholangitis) was made based on these findings. However, complete characterization of the narrowed segment itself was not possible with MRC. Furthermore, the intrahepatic ductal changes of sclerosing cholangitis (evident on ERCP) were not appreciated on MRC. Intraductal Abnormalities All 7 cases of intraductal abnormalities were correctly diagnosed on MRC: 6 patients with choledocholithiasis (Figure 3) and 1 patient with a bile duct polyp. The number of stones was 1 in 3 patients, 2 in 1 patient, Figure 3. Choledocholithiasis occurring after cholecystectomy. (A) and 3 in 2 patients. The size and number of stones were MRC showing moderate biliary ductal dilatation and two distal intraaccurately predicted by MRC in all patients. The size of ductal foci of low-signal intensity representing stones (arrows); the gallbladder is absent. (B) ERCP at the time of the removal of stones stones varied from 3 mm to 1.5 cm. Review of the MRC (arrowheads). source images by the radiologists was very important in

February 1996 MR CHOLANGIOGRAPHY: COMPARISON WITH ERCP 593 Figure 4. Stricture secondary to chronic pancreatitis in a 56-year-old patient with increasing bilirubin levels. (A) Original MRC. (B ) Original ERCP. (C ) MRC postendoscopic placement of biliary stent. (D) ERCP poststenting. Original MRC and ERCP both show biliary ductal dilatation and a smooth stricture at the head of the pancreas (arrows). (C ) Poststent MRC and (D) ERCP; the stent itself is well depicted (arrowhead). confirming the presence and number of stones because they allowed a better demonstration of the typical meniscus sign created by the high signal intensity of bile that was partially (or completely) surrounding the hypointense stone. Imaging findings for the stones and the polyp were identical: intraductal foci of low signal on MRC and intraluminal filling defects on direct cholangiography. In fact, the polyp was misinterpreted as a stone by both MRC and ERCP and is therefore not considered to be a false-positive finding of choledocholithiasis (direct cholangiography was the standard of reference). The true diagnosis was made with direct endoscopic visualization

594 SOTO ET AL. GASTROENTEROLOGY Vol. 110, No. 2 not available at the time of the blind interpretation. Overall, the sensitivity for detection of intraductal abnormalities was 100%. Figure 5. Pancreatic head carcinoma. A 65-year-old man with increasing abdominal pain and abnormal liver function tests. (A) MRC show- ing severe biliary ductal dilatation and a stricture in the head of the pancreas (short arrow). The pancreatic duct (long arrow) is of normal caliber but is not seen in the region of the biliary stricture (open arrow), suggesting involvement by the tumor mass. (B) ERCP confirms the presence of strictures involving the common bile duct (curved arrow) and pancreatic duct (arrowhead); this is the typical doubleduct sign of pancreatic head carcinoma. Discussion Diagnostic imaging of the biliary ductal system typically begins with noninvasive modalities, such as ultrasonography, computed tomography, and occasionally biliary scintigraphy. However, the information provided by these studies is often insufficient for diagnosis despite multiple successive examinations that add a significant cost to the diagnostic workup. As a result, direct cholangiography obtained through ERCP or PTC is often required as a definitive diagnostic test. These tests are associated with complications 8,9 such as bleeding (in PTC) and pancreatitis (in ERCP). MR cholangiography is a newly introduced imaging modality for evaluating the biliary and pancreatic ductal systems. Advantages over ultrasonography, computed tomography, and biliary scintigraphy include high-resolution projectional images, not achievable with these other noninvasive techniques. Furthermore, MR offers the opportunity of providing both high-quality cross-sectional images of extraductal structures and projectional images of the biliary tree and pancreatic duct without administration of any oral or intravenous contrast material. Computed tomography cholangiography was recently introduced as another method for indirect cholangiography. 10,11 It takes advantage of the high contrast and spatial resolution of spiral computed tomography to produce three-dimensional reconstructions of the opacified biliary tree. However, the major drawback (compared with MRC) is the requirement for contrast materials such as iodipamide meglumine or its derivatives. These agents have fallen into disfavor because of high frequency and severity of adverse reactions. 9,12 Moreover, computed tomography cholangiography requires the use of ionizing radiation and depends on sufficient hepatic function for adequate uptake and biliary excretion of contrast agents. In addition, opacification of the pancreatic duct is not obtained with this technique. Several studies utilizing a variety of pulse sequences for MRC have been reported to date. Breathhold 1 4 and nonbreathhold techniques 5,6,13,14 have been used. The principle is the same, regardless of the specific sequence; on heavily T2-weighted images, stationary fluid (such as bile) exhibits a high signal intensity, whereas the signal of surrounding tissues (background) is markedly sup- pressed. Published data indicate that MRC is highly ac- curate in the detection of the dilated bile ducts, 1 4,13,14 but is less clear about its potential in the diagnosis of choledocholithiasis, because reported sensitivities have

February 1996 MR CHOLANGIOGRAPHY: COMPARISON WITH ERCP 595 Figure 6. Malignant biliary obstruction due to metastatic peri-portal lymphadenopathy from endometrial carcinoma. (A) MRC. (B) ERCP. Both studies show marked biliary ductal dilatation above a midcommon bile duct stricture (short arrows). The normal distal common bile duct (long arrows) and pancreatic duct (arrowheads) are also seen in both cholangiographic studies. However, only MRC visualized the distended cystic duct (broad arrow) and gallbladder. varied from 71% 6 to 100%. 5 The technique we use for clot, migrating parasite, or intraductal tumor. However, MRC differs from those described by other investigators. similar difficulties may be encountered when interpreting This 3D multislab FSE sequence yields better-quality ERCP images, because air bubbles introduced during reconstruction images than the 2D FSE sequences used in bile duct cannulation can be mistaken for stones. other studies. 5,6 Furthermore, the addition of respiratory The presence, site, and length of biliary strictures were triggering decreases the respiratory motion artifact. accurately shown in 90% of the patients in our study. Our results confirm that the sensitivity for detec- However, a depiction of the morphological characteristics tion of biliary ductal dilatation and choledocholithiasis of the stricture may be impossible, as in 1 of the cases (including the size, number, and location of stones) ap- in our study. Severe narrowing may be seen only as a proaches 100%. Apparent underestimation of the severity segment of signal dropout, generally associated with dila- of dilatation when compared with direct cholangio- tation of the proximal ductal system. Lack of dilatation grams is more likely related to overdistention of the ducts may be a clue indicating the presence of diffuse or in the latter, because MRC depicts the ducts in their multifocal strictures involving the bile ducts; although resting state. The sensitivity for a diagnosis of choledocholithiasis this was true in 1 of our patients who had sclerosing is superior to that of other noninvasive tech- cholangitis, it remains to be proven with a larger popula- niques, such as ultrasonography (20% 65%) 15 18 and tion of this subgroup of patients. computed tomography (45% 85%). 19 21 The impor- In our patient population, the only case of abnormal tance of a careful review of the source images in confirming bile ducts and stricture not shown with MRC was related the presence of intraductal lesions cannot be overem- to a suboptimal study due to a motion artifact. Hence, phasized. The only caveat is the unlikely event of when image quality is not ideal, further evaluation with confusing a stone for another type of intraluminal filling direct cholangiography is generally warranted. Pitfalls of defect, which occurred in 1 of our patients with a bile the technique include the presence of metallic clips in the duct polyp; this could also conceivably happen with a upper abdomen, which can occasionally create artifacts,

596 SOTO ET AL. GASTROENTEROLOGY Vol. 110, No. 2 usually seen as periductal foci of low signal. Additional References images obtained in the axial plane may help distinguish 1. Wallner BK, Schumacher KA, Weidenmaier W, and Friedrich, JM. intraluminal from extraluminal abnormalities. A practi- Dilated biliary tract: evaluation with MR cholangiography with a cal approach toward making this distinction is to corre- heavily T2-weighted contrast-enhanced fast-sequence. Radiology late the MRC images with a plain radiograph of the 1991;181:805 808. 2. Morimoto K, Shimoi M, Shirakawa T, Aoki T, Choi S, Miyata Y, upper abdomen, looking for metallic objects that could Hara K. Biliary obstruction: evaluation with three-dimensional MR be responsible for the abnormal focus of signal intensity. cholangiography. Radiology 1992;183:578 580. Other artifacts, such as the overestimation of duct narresonance 3. Ishizaki Y, Wakayama T, Okada Y, and Kobayashi T. Magnetic cholangiography for evaluation of obstructive jaun- rowing or obscuration of intraductal defects, may result dice. Am J Gastroenterol 1993;12:2072 2077. from the reconstruction algorithm itself. 22 4. Hall-Craggs M, Allen C, Owens C, Theis BA, Donald JJ, Paley M, Despite the excellent diagnostic capabilities of MRC, Wilkinson ID, Chong WK, Hatfield ARW, Lees WR, Russell RCG. its major limitation compared with ERCP and PTC is MR cholangiography: Clinical evaluation in 40 cases. Radiology 1993;189:423 427. its inability to offer therapeutic intervention. In fact, 5. Guibaud L, Bret PM, Reinhold C, Atri M, Barkum AN. Diagnosis 31 of our 46 patients (67.4%) required some type of of choledocholithiasis: value of MR cholangiography. AJR endoscopic intervention: sphincterotomy, stone removal, 1994;163:847 850. and stricture stenting. Although MRC cannot replace 6. Macaulay SE, Schulte SJ, Sekijima JH, Obregon RG, Simon HE, Rohrmann CA, Freeny PC, Schmiedl UP. Evaluation of a non- ERCP or PTC as a therapeutic maneuver, 23 it may sigbreath hold MR cholangiography technique. Radiology 1995; nificantly decrease the need for purely diagnostic ERCP. 196:227 232. Furthermore, MRC is an appropriate triage study before 7. Barish MA, Yucel EK, Soto JA, Chuttani R, Ferrucci JT. MR cholangiopancreatography: efficacy of three-dimensional turbo spinendoscopic or radiological intervention; for example, it echo technique. AJR 1995;165:295 300. may be more suitable for a patient with high-grade 8. Bilbao MK, Dotter CT, Lee TG, Katon RM. Complications of endoobstruction at the porta hepatis to undergo antegrade scopic retrograde cholangiopancreatography (ERCP). A study of cholangiography (PTC) and stenting than ERCP, 10,000 cases. Gastroenterology 1976;70:314 320. 9. Ott D, Gelfand D. Complications of gastrointestinal radiologic thereby obviating the possibility of a failed endoscopic procedures. II. Complications related to biliary tract studies. Gasintervention. The study is also useful in cases in which trointest Radiol 1981;6:47 56. ERCP is incomplete or when cannulation of the bile 10. Van Beers BE, Lacrosse M, Triqaux JP, de Canniere L, DeRonde duct is not possible because of technical difficulties or T, Pringot J. Noninvasive imaging of the biliary tree before or after laparoscopic cholecystectomy: use of three-dimensional spiral anatomic factors. MRC may be used as the method for CT cholangiography. AJR 1994;162:1331 1335. obtaining cholangiographic images in patients with ab- 11. Stockberger SM, Wass JL, Sherman S, Lehman GA, Kopecky KK. normal liver function tests or upper abdominal sympendoscopic Intravenous cholangiography with helical CT: comparison with retrograde cholangiography. Radiology 1994; toms of possible biliary origin. The demonstration of a 192:675 680. normal bile tract allows avoidance of unnecessary inva- 12. Maglinte D, Dorenbusch M. Intravenous infusion cholangiograsive diagnostic tests. Finally, the high sensitivity for phy: an assessment of its role relevant to laparoscopic cholecystectomy. diagnosis of choledocholithiasis may make MRC the Radiol Diagn 1993;34:91 96. 13. Meakem TJ, Holland GA, McDermott VG, White LD, Korn RL, ideal method for evaluating patients before laparoscopic Schnall MD, Kowalski TE, Outwater EK, Kressel HY. Fast spin cholecystectomy. echo multicoil magnetic resonance cholangiography: Initial experience An additional value of MRC that ERCP and PTC do (abstr). SMRM 1993, 1:41. not have is the fact that excellent quality cross-sectional 14. Outwater EK. MR cholangiography with a fast spin-echo sequence (abstr). J Magn Reson Imaging 1993, 3:131. images of the liver and pancreas are obtained in the same 15. Gross BH, Harter LP, Gore RM, Callen PW, Filly RA, Shapiro HA, sitting, allowing direct visualization of both intraductal Goldberg, HI. Ultrasonic evaluation of common duct stones: prospective and extraductal lesions. In addition to its excellent diagcreatography. comparison with endoscopic retrograde cholangiopanand Radiology 1983; 146:471 477. nostic capabilities in biliary tract disease evaluation, the 16. Cronan JJ, Mueller PR, Simeone JF, O Connell RS, van Sonnennormal and abnormal 25,26 pancreatic duct can also be berg E, Wittenburg J, Ferrucci JT. Prospective diagnosis of choledconsistently shown with the same technique. MR chol- ocholithiasis. Radiology 1983;146:467 469. angiopancreatography, therefore, provides a comprehensonography 17. Einstein DM, Laysen SA, Ralls PW, Halls JM. The insensitivity of in the detection of choledocholithiasis. AJR sive (and single) noninvasive examination for suspected 1984;142:725 728. biliary and pancreatic diseases. It can decrease the total 18. Laing FC, Jeffrey RB, Jr. Choledocholithiasis and cystic duct obcost of the diagnostic workup by eliminating the need for struction: Difficult ultrasonographic diagnosis. Radiology 1983; multiple, consecutive, noninvasive, and invasive studies. 146:475 479. 19. Jeffrey RB, Federle MP, Laing FC, Wall S, Rego J, Moss AA. MR cholangiopancreatography has the potential to be- Computed tomography of choledocholithiasis. AJR 1983; come the diagnostic modality of choice in many patients 140:1179 1183. with suspected pancreatic or biliary abnormalities. 20. Mitchell SE, and Clark RA. A comparison of computed tomography

February 1996 MR CHOLANGIOGRAPHY: COMPARISON WITH ERCP 597 and sonography in choledocholithiasis. AJR 1984;142:729 dimensional fast spin-echo technique. Radiology 1995;196: 733. 459 464. 21. Baron RL, Stanley RJ, Lee JKT, Koehler RE, Levitt RG. Computed 26. Takehara Y, Ichijo K, Tooyama N, Kodaira N, Yamamoto H, Tatami tomographic features of biliary obstruction. AJR 1983;140: M, Saito M, Watahiki H, Takahashi M. Breath-hold MR cholangio- 1173 1178. pancreatography with a long-echo train fast spin-echo sequence 22. Anderson CM, Saloner D, Tsuruda JS, Shapeero LG, Lee RE. and a surface coil in chronic pancreatitis. Radiology 1994; Artifacts in maximum-intensity projection display of MR angio- 192:73 78. grams. AJR 1990;154:623 629. 23. Outwater EK, Gordon SJ. Imaging the pancreatic and biliary ducts with MR. Radiology 1994;192:19 21. Received July 20, 1995. Accepted October 19, 1995. 24. Laubenberger J, Buchert M, Schneider B, Blum U, Hennig J, Address requests for reprints to: Jorge A. Soto, M.D., Department Langer M. Breath-hold projection magnetic resonance-cholangio- of Radiology, Boston University Medical Center, 88 East Newton pancreaticography (MRCP): a new method for the examination of Street, Boston, Massachusetts 02118. Fax: (617) 638-6616. the bile and pancreatic ducts. Magn Reson Med 1995;33:18 Supported in part by a grant from Philips Medical Systems, Shelton, 23. Connecticut. 25. Soto JA, Barish MA, Yucel EK, Siegenberg D, Chuttani R, Ferrucci The authors thank Dr. J. Thomas LaMont for his valuable suggestions JT. Pancreatic duct: MR cholangiopancreatography with a three- regarding this manuscript.