26 GASTROINTESTINAL ENDOSCOPY VOLUME 49, NO. 1, 1999

Choledocholithiasis is a common complication ofgallbladder stones, occurring among 15% to 20% ofpatients.1 Among patients who have undergonecholecystectomy, 1% to 5% have retained or recur-

rent bile duct stones.1 Clinical and biochemicalabnormalities associated with this condition are nei-ther accurate nor specific enough for diagnosis. Atpresent, two noninvasive procedures, abdominal USand CT, are the imaging methods of first choice fordiagnosis, but they are not sufficiently sensitive.2-5ERCP with or without sphincterotomy and intraop-erative cholangiography with or without choledo-choscopy are accurate procedures used as a secondchoice because of their invasive nature.6-9 There isneed for a less invasive but highly accurate preoper-ative imaging method.

EUS and magnetic resonance cholangiopancre-atography (MRCP) are nonaggressive imaging pro-cedures that are particularly useful for explorationof the biliopancreatic region.10-14 MRCP results in

Diagnosis of choledocholithiasis: EUS or magneticresonance cholangiography? A prospective controlledstudyVictor de Ldinghen, MD, Robin Lecesne, MD, Jean-Michel Raymond, MD, Vronique Gense, MD,Michel Amouretti, MD, Jacques Drouillard, MD, Patrice Couzigou, MD, Christine Silvain, MDPessac and Poitiers, France

Background: Endoscopic ultrasonography (EUS) appears to be the bestimaging method for the diagnosis of choledocholithiasis.The aim of this pre-liminary, prospective, controlled study was to assess the accuracy of EUSand magnetic resonance cholangiopancreatography (MRCP) in the diagno-sis of common bile duct stones.Methods: From December 1995 through April 1997, all patients referredbecause of suspicion of the presence of common bile duct stones wereincluded in the study. EUS and MRCP were performed. Each examinationwas performed by a different operator unaware of the result of the other pro-cedure. The definitive diagnosis was established by means of endoscopicretrograde cholangiography with sphincterotomy or a surgical procedure.Results: Forty-three patients (18 men, 25 women) with a mean age of 60.9 14.5 years (range 25 to 81 years) were included in the study. Eleven patientswere excluded because of unavailability of magnetic resonance imaging(n = 5) or EUS (n = 6). Ten patients (31.2%) had choledocholithiasis. For thisdiagnosis, the sensitivity of EUS was 100%, the specificity was 95.4%, thepositive predictive value was 90.9%, and the negative predictive value was100%. The corresponding values for MRCP were 100%, 72.7%, 62.5%, and100%, not significantly different from EUS results. The accuracy of EUS was96.9%, and that of MRCP was 82.2%.Conclusion: This preliminary study confirmed EUS as an accurate and non-invasive procedure for the diagnosis of common bile duct stones. MRCP,which had a high sensitivity and high negative predictive value, might be anaccurate technique for patients with a contraindication to EUS.(Gastrointest Endosc 1999;49:26-31.)

Received September 18, 1997. For revision January 30, 1998.Accepted June 16, 1998.From the Departments of Hepatogastroenterology and MedicalImaging, Hpital du Haut-Lvque, and Department ofHepatogastroenterology, Hpital Jean Bernard, Pessac, France.Reprint requests: Docteur Victor de Ldinghen, Service dHpato-Gastroentrologie, Hpital du Haut-Lvque, 33604 Pessac cedex,France.Copyright 1999 by the American Society for GastrointestinalEndoscopy0016-5107/99/$8.00 + 0 37/1/92453

Diagnosis of choledocholithiasis: EUS or MRCP? V de Ldinghen, R Lecesne, J-M Raymond, et al.

VOLUME 49, NO. 1, 1999 GASTROINTESTINAL ENDOSCOPY 27

the diagnosis of choledocholithiasis have not beencompared with those of EUS. The aims of this pre-liminary, prospective, controlled study were to eval-uate and compare the diagnostic accuracy of EUSand MRCP in the care of patients with suspectedcholedocholithiasis.

PATIENTS AND METHODSInclusion criteria

From December 1995 through April 1997, patientswere included if they had clinical or biochemical signs ofcholedocholithiasis according to the following criteria:combination of epigastric or right upper quadrant painwith fever or jaundice; one or two of the previous signstogether with an elevation of serum alkaline phosphataselevel or an elevation of serum g -glutamyl transpeptidaseor transaminase level more than the upper limit of nor-mal; acute pancreatitis, defined as acute epigastric painassociated with an elevation of serum amylase, lipase, orurinary amylase level more than two times the upperlimit of normal; and unexplained cholestasis defined by anelevation of serum alkaline phosphatase level and an ele-vation of serum g -glutamyl transpeptidase level to morethan two times the upper limit of normal.

Exclusion criteriaPatients were excluded if long-term daily alcohol

intake exceeded 80 g, they were taking a hepatotoxic drug,or if serum hepatitis B or C antibodies were present.

Radiologic methodsAll patients underwent EUS and MRCP. All examina-

tions were performed by two different operators unawareof the results of the other investigation. EUS was per-formed with an Olympus GF EUM20 endoscope system(Scop, Rungis, France). The transducer was inserted to thedistal portion of the second duodenum and graduallydrawn back to the stomach. Acoustic coupling of the trans-ducer to the digestive wall was achieved with a balloonfilled with 5 to 20 mL of water. The procedures were per-formed with general anesthesia. Examination time was 15to 30 minutes. The EUS diagnostic criterion for choledo-cholithiasis was a hyperechoic structure within the com-mon bile duct sometimes associated with an acousticshadow. The common hepatic duct was consideredenlarged if the diameter was more than 7 mm (more than10 mm for patients who had undergone cholecystectomy).

All magnetic resonance examinations were performedafter fasting for a minimum of 4 hours to promote fillingof the gallbladder and gastric emptying. Magnetic reso-nance imaging was performed with a 1 T system(Siemens, Erlangen, Germany). The patients were exam-ined in the supine position. A phase-array receiving coilstrapped around the abdomen was used. Antiperistalticdrugs were injected intravenously. Before MRCP, a T1-weighted ultrafast low-angle shot sequence was per-formed to localize the biliary tree. MRCP was performedwith a half-fourier acquisition single shot Turbo spin echo

(HASTE) sequence with an effective echo time of 87 ms,one excitation, and a 240 256 matrix.15 Because theHASTE sequence is a single-shot sequence, there is norepetition time. The fat-suppression technique was usedwhen peritoneal fat resulted in artifact.

Two acquisition techniques were used. First, sequen-tial multisection imaging was performed to obtain eight 5-mm sections in a single breath hold of 14 seconds. Theentire biliary tree was explored in 4 to 6 breath-hold peri-ods in a coronal plane and in 6 to 8 periods in an axialplane. Maximum intensity projection reconstruction wasnot performed. Second, projection imaging was performedwith a 20-mm thick section acquired in a coronal planewithin a single breath hold of 2 seconds. The section posi-tion of projection imaging corresponded to the position ofthe 5-mm section that included the common bile duct. TheMRCP results were focused on the common bile duct (sizeand number of common bile duct stones), and the entirebiliary tract was explored. At the end of the examination,MRCP images were reviewed at the console jointly by twoexperienced radiologists to establish a real-time report.Common bile duct stones were diagnosed with MRCPwhen a round, oval, or multifaceted area of signal void(hypointensity) was present within the lumen of thehyperintense bile duct. The common hepatic duct was con-sidered enlarged if the diameter was more than 7 mm(more than 10 mm for patients who had undergone chole-cystectomy).

Final diagnosisAfter EUS and MRCP were performed, patients under-

went surgical treatment or endoscopic investigation. Thechoice between endoscopic investigation and surgicaltreatment depended on surgical risk and whether thepatient had undergone cholecystectomy. All patientswhose gallbladders had been removed underwent ERCPwith general anesthesia. When ERCP was performed,endoscopic sphincterotomy was systematically conductedafter opacification of the common bile duct. Bile duct stone clearance was attempted with stone baskets, balloon

Figure 1. Magnetic resonance cholangiogram shows chole-docholithiasis. CBD, common bile duct; PD, pancreatic duct;S, stones.

catheters, or both. In the surgical group, all patientsunderwent intraoperative cholangiography performedthrough the cystic duct or after choledochotomy. If intra-operative cholangiography did not demonstrate stones, abasket was passed through the common bile duct. WhenEUS, MRCP, or intraoperative cholangiography led to asuspicion of choledocholithiasis, intraoperative choledo-choscopy was performed systematically. Choledochoscopywas performed after choledochotomy or through the cysticduct if the diameter was large enough. No complicationswere encountered.

Statistical analysisThe sensitivity, specificity, and accuracy of EUS and

MRCP in the diagnosis of choledocholithiasis were calcu-lated with a 95% confidence interval (CI).

RESULTSFrom December 1995 through April 1997, 43

patients (18 men, 25 women) with a mean age of 60.9 14.5 years (range 25 to 81 years) enrolled in thestudy. Eleven patients were excluded because mag-netic resonance imaging (n = 5) or EUS (n = 6) wasnot available. Therefore 32 of 43 patients (74.4%)underwent EUS and MRCP. Median time betweenEUS and MRCP was 1 day (range 0 to 7 days). EUSwas performed before MRCP in 12 instances andafter MRCP in 20 instances. EUS and MRCP weresuccessful in all instances. No inhospital morbidityor mortality was observed. Twenty-five patientsunderwent surgical treatment (including 5 withprior cholecystectomy). No clinical or biochemical dif-ference was observed between patients who under-went operations and those who underwent endo-scopic investigation. Twenty-three patients had gall-bladder stones at the time of presentation. Mediantime between EUS or MRCP and operation or ERCPwas 4.5 days (range 1 to 50 days).

Patients were separated into three groups, as fol-lows: those with stones at EUS and MRCP (group A,n = 11; Fig 1), those without stones at EUS andMRCP (group B, n = 16), and patients with no stonesat EUS but with stones at MRCP (group C, n = 5).Results are shown in Tables 1 and 2.

In group A, 10 of 11 patients (90.9%) were foundto have stones at surgical intervention (n = 7) orERCP (n = 3). Two of these patients had a stone 10mm or more in diameter and 8 a stone less than 10mm in diameter. No discrepancy was observedbetween EUS and MRCP findings concerning stonesize. The single false-positive finding for EUS andMRCP in this group was that of a patient who hadgallbladder and cystic duct stones found during anoperation performed 24 hours after EUS. The stonediameter was 4 mm by EUS and MRCP in a normal-diameter common bile duct. This patient also had

surgically proved swelling of the papilla caused byrecent stone migration. All 16 patients (100%) ingroup B were found to be free of common bile ductstones at operation (n = 12) or ERCP (n = 4). In groupC, all 5 patients (100%) were found to be free of com-mon bile duct stones at operation (n = 5). For all ofthem, stone diameter was less than 10 mm at MRCP.

Choledocholithiasis was found in 10 (31.2%) ofthe 32 consecutively treated patients. The mean ageof patients with stones was 66.3 8.4 years. Twopatients had previously undergone cholecystectomy,and 6 had gallbladder stones. In 2 patients the com-mon bile duct was not enlarged. Twenty-twopatients did not have choledocholithiasis. Theirmean age was 54.3 17.1 years, and 2 of them hadpreviously undergone cholecystectomy.

Sensitivity, specificity, positive predictive value,and negative predictive value of EUS and MRCP areshown in Table 3. In the diagnosis of common bileduct stones, the accuracy of EUS was 96.87% (95%CI [83.78, 99.92]) and that of MRCP 81.25% (95% CI [63.56, 92.79]).

DISCUSSIONDiagnostic imaging of the biliary ductal system

typically begins with noninvasive modalities such asultrasonography or CT. However, informationobtained with these techniques often is insufficientfor diagnosis despite successive examinations thatmarkedly increase the cost of diagnostic evaluation.In our study, 22 patients did not have choledo-cholithiasis. This ratio is high because patients wereselected on few predictive features of choledo-cholithiasis.

Ultrasonography is the easiest, fastest, and leastexpensive imaging procedure used for the diagnosisof choledocholithiasis. The diagnostic accuracy ofthis technique for extrahepatic cholestasis is high,the sensitivity reaching 94% and the specificity100%.3,16,17 The level of biliary obstruction is identi-fied more than 90% of the time, but the cause can bedetermined for only 71% of patients. Diagnostic fail-ures are caused mainly by the location of stoneswithin the intrapancreatic portion of the bile ductand the frequent absence of bile duct dilatation.With the use of real-time and high-definition ultra-sonography, diagnostic sensitivity for choledo-cholithiasis is 55%.2-4

The overall diagnostic accuracy of CT for extra-hepatic cholestasis is 87% to 98%.18-20 CT is moreaccurate than ultrasonography, especially in exami-nations of overweight patients and those with inter-posed digestive gas. CT has a sensitivity of 76% anda specificity of 98% for the diagnosis of choledo-cholithiasis.5

V de Ldinghen, R Lecesne, J-M Raymond, et al. Diagnosis of choledocholithiasis: EUS or MRCP?

28 GASTROINTESTINAL ENDOSCOPY VOLUME 49, NO. 1, 1999

EUS is a valuable, noninvasive technique in thediagnosis of choledocholithiasis. In our study, nomorbidity was encountered after EUS. Results areexcellent in the diagnosis of extrahepatic cholesta-sis.21 For choledocholithiasis, the sensitivity ofEUS is 93% to 97%, and the specificity is 97% to100%, better than for CT and ultrasonogra-phy.10,12,14 The results of EUS do depend on stonesize and the diameter of the bile duct. EUS is par-ticularly accurate for the diagnosis of small stonesin the distal common bile duct, especially if theduct is not enlarged. Our study confirmed all pre-vious reports concerning the accuracy of EUS inthe diagnosis of choledocholithiasis (sensitivity100%, specificity 95.4% regardless of stone or com-mon bile duct diameter). The negative predictivevalue was 100%, so use of EUS may prevent unnec-essary interventions.

Use of MRCP has increased rapidly in recentyears. Projectional images similar in appearance todirect cholangiograms produced with ERCP or per-cutaneous transhepatic cholangiography areobtained without oral or intravenous administrationof a contrast agent. The noninvasive nature ofMRCP makes it an appealing modality for the visu-alization of all segments of the biliary tract. MRCPis not a mature procedure, but in the detection ofcholedocholithiasis the sensitivity of HASTE MRCPis identical to that of other MRCP methods.22 In ourstudy, no morbidity was observed. Several recentstudies involving small numbers of patients havedemonstrated the success of this technique in thediagnosis of choledocholithiasis.23-27 MRCP has asensitivity of 71% to 100%, a specificity of 85%, andan accuracy of 89% to 94%,11,13,15,26 superior tothose of ultrasonography and CT. Our study con-firmed all previous results; MRCP in the diagnosisof choledocholithiasis had a sensitivity of 100% anda specificity of 72.7% regardless of the diameter ofthe stone or common bile duct. The negative predic-tive value was 100%, a factor that may preventunnecessary intervention.

ERCP and intraoperative cholangiography areconsidered to be the best imaging procedures forthe diagnosis of choledocholithiasis. However,

ERCP has a morbidity rate of 3% to 5%,6 8% to10% if sphincterotomy is performed,28 and a pro-cedure-related mortality rate of 1%.9 ERCP andintraoperative cholangiography may not helpdetect small stones (

exploration. Indeed, the intervals between MRCPand surgical intervention were 1, 3, and 9 days inthe three cases. In all cases of conflicting resultsbetween EUS and MRCP, MRCP was performedbefore EUS (range 0 to 4 days). In our study, themedian time between EUS or MRCP and ERCP orintraoperative cholangiography was 1 day. Thistime may not have precluded the possibility ofspontaneous stone fragmentation or movementfrom the bile duct into the duodenum.

The standard imaging methods for the diagnosisof choledocholithiasis (ERCP and intraoperativecholangiography) are not sensitive enough to com-pletely reassure the surgeon or endoscopist thatinterventional manipulation of the common bileduct is not necessary. We have shown that EUS orMRCP may prevent some of these unnecessaryinterventions. These preliminary results show thatEUS and MRCP appear to be the most reliablepretherapeutic diagnostic modalities for choledo-cholithiasis, but larger prospective studies are need-ed to confirm this conclusion. For patients withsymptomatic choledocholithiasis, the strategy fordiagnosis should include ultrasonography as thefirst choice of imaging followed by EUS or MRCP.MRCP might be more useful than EUS in the careof elderly patients at high surgical risk and patientsfor whom endoscopy has been unsuccessful, such aspatients who have undergone surgical bypass proce-dures. ERCP with sphincterotomy and choledo-choscopy can be reserved for therapeutic use.

REFERENCES1. Hermann RE. The spectrum of biliary stone disease. Am J

Surg 1989;158:171-3.2. Cronan JJ. US diagnosis of choledocholithiasis: a reappraisal.

Radiology 1986;161:133-4.3. Laing FC, Jeffrey RB, Wing WW, Nyberg DA. Biliary dilata-

tion: defining the level and cause by real time US. Radiology1986;160:39-42.

4. Mitchell SE, Clark RA. A comparison of computed tomogra-phy and sonography in choledocholithiasis. Am J Roentgenol1984;142:729-33.

5. Baron RL. Common bile duct stones: reassessment of criteriafor CT diagnosis. Radiology 1987;162:419-24.

6. Cotton PB. Progress report: ERCP. Gut 1977;18:316-41.

V de Ldinghen, R Lecesne, J-M Raymond, et al. Diagnosis of choledocholithiasis: EUS or MRCP?

30 GASTROINTESTINAL ENDOSCOPY VOLUME 49, NO. 1, 1999

7. Frey CF, Burbige EJ, Meinke WB, Pullos TG, Nang Wong H,Hickman DM, Belber J. Endoscopic retrograde cholangiopan-creatography. Am J Surg 1982;144:109-14.

8. Shimizu S, Tada M, Kawai K. Diagnostic ERCP. Endoscopy1994;26:88-92.

9. Vaira D, DAnna L, Ainley C, Dowsett J, Williams S, Baillie J,et al. Endoscopic sphincterotomy in 1000 consecutive patients.Lancet 1989;2:431-3.

10. Amouyal P, Amouyal G, Lvy P, Tuzet S, Palazzo L, Vilgrain V,et al. Diagnosis of choledocholithiasis by endoscopic ultra-sonography. Gastroenterology 1994;106:1062-7.

11. Guibaud L, Bret PM, Reinhold C, Atri M, Barkun AN. Bileduct obstruction and choledocholithiasis: diagnosis with MRcholangiography. Radiology 1995;197:109-15.

12. Prat F, Amouyal G, Amouyal P, Pelletier G, Fritsch J, ChouryAD, et al. Prospective controlled study of endoscopic ultra-sonography and endoscopic retrograde cholangiography inpatients with suspected common-bile duct lithiasis. Lancet1996;347:75-9.

13. Soto JA, Barish MA, Yucel EK, Siegenberg D, Ferrucci JT,Chuttani R. Magnetic resonance cholangiography: compari-son with endoscopic retrograde cholangiopancreatography.Gastroenterology 1996;110:589-97.

14. Palazzo L, Girollet PP, Salmeron M, Silvain C, Roseau G,Canard JM, et al. Value of endoscopic ultrasonography in thediagnosis of common bile duct stones: comparison with surgi-cal exploration and ERCP. Gastrointest Endosc 1995;42:225-31.

15. Chan YL, Chan ACW, Lam WWM, Lee DWH, Chung SSC,Sung JJY, et al. Choledocholithiasis: comparison of MRcholangiography and endoscopic retrograde cholangiography.Radiology 1996;200:85-9.

16. Ferrucci JT, Adson M, Mueller PR, Stanley R, Stewart E.Advances in the radiology of jaundice: a symposium andreview. Am J Roentgenol 1983;141:1-20.

17. Taylor TV. Diagnostic accuracy of gray scale ultrasonographyfor jaundiced patients. A report of 275 cases. Arch Intern Med1979;139:60-3.

18. Baron RL, Stanley RJ, Lee JKT, Koehler RE, Levitt RG.Computed tomography features of biliary obstruction. Am JRoentgenol 1983;140:1173-8.

19. Shimizu H, Ida M, Takayama S, Seiki T, Yoneda M. The diag-nosis accuracy of computed tomography in obstructive biliarydisease: a comparative evaluation with direct cholangiogra-phy. Radiology 1981;138:411-6.

20. Pedrosa CS, Casanova R, Lezana AH, Fernandez MC.Computed tomography in obstructive jaundice: the cause ofobstruction. Radiology 1981;139:635-45.

21. Amouyal P, Palazzo L, Amouyal G, Ponsot P, Mompoint D,Vilgrain V, et al. Endosonography: promising method for diag-nosis of extrahepatic cholestasis. Lancet 1989;2:1195-8.

22. Miyazaki T, Yamashita T, Tsuchigame T, Yamamoto H, Urata

Table 3.Overall results of endoscopic ultrasonography (EUS) and magnetic resonance cholangiography (MRCP)for the diagnosis of choledocholithiasisValue EUS (%) 95% CI MRCP (%) 95% CI

Sensitivity 100 69.1100 100 69.1100Specificity 95.4 77.299.9 72.7 49.789.3Positive predictive value 90.9 58.799.8 62.5 35.484.8Negative predictive value 100 83.9100 100 79.4100

CI, Confidence interval.

Diagnosis of choledocholithiasis: EUS or MRCP? V de Ldinghen, R Lecesne, J-M Raymond, et al.

VOLUME 49, NO. 1, 1999 GASTROINTESTINAL ENDOSCOPY 31

J, Takahashi J . MR cholangiopancreatography using HASTE(Half-Fourier Acquisition Single-shot Turbo Spin-Echo)sequences. AJR 1996;166:1297-303.

23. Macaulay SE, Schulte SJ, Sekijima JH, Obregon RG, SimonHE, Rohrmann CA, Freeny PC, Schmiedl UP. Evaluation of anon-breath-hold MR cholangiography technique. Radiology1995;196:227-32.

24. Barish MA, Yucel EK, Soto JA, Chuttani R, Ferrucci JT. MRcholangiopancreatography: efficacy of three-dimensional turbospin-echo technique. Am J Roentgenol 1995;165:295-300.

25. Guibaud L, Bret PM, Reinhold C, Atri M, Barkun ANG.Diagnosis of choledocholithiasis: value of MR cholangiogra-phy. Am J Roentgenol 1994;163:847-50.

26. Morimoto K, Shimoi M, Shirakawa T, Aoki Y, Choi S, MiyataY, Hara K. Biliary obstruction: evaluation with three-dimensional MR cholangiography. Radiology 1992;183:578-80.

27. Hall-Craggs MA, Allen CM, Owens CM, Theis BA, Donald JJ,Paley M, et al. MR cholangiography: clinical evaluation in 40cases. Radiology 1993;189:423-7.

28. Davidson BR, Neoptolemos JP, Carr-Locke DL. Endoscopicsphincterotomy for common bile duct calculi in patients withgallbladder in situ considered unfit for surgery. Gut 1988;29:114-20.

29. Glenn F. Retained calculi within the biliary ductal system.Ann Surg 1974;179:528-39.

Availability of Journal back issues

As a service to our subscribers, copies of back issues of Gastrointestinal Endoscopy forthe preceding 5 years are maintained and are available for purchase from Mosby untilinventory is depleted at a cost of $15.00 per issue. The following quantity discounts areavailable: 25% off on quantities of 12 to 23, and one third off on quantities of 24 or more.Please write to Mosby, Inc., Subscription Services, 11830 Westline Industrial Dr., St.Louis, MO 63146-3318, or call 800-453-4351 or 314-453-4351 for information on avail-ability of particular issues. If unavailable from the publisher, photocopies of complete issues may be purchased from UMI, 300 N. Zeeb Rd., Ann Arbor, MI 48106;313-761-4700.