Eur Radiol (2007) 17: 2262–2267DOI 10.1007/s00330-007-0640-z HEPATOBILIARY-PANCREAS

Utaroh MotosugiTomoaki IchikawaTsutomu ArakiFumiaki KitaharaTadashi SatoJun ItakuraHideki Fujii

Received: 27 December 2006Revised: 28 February 2007Accepted: 22 March 2007Published online: 20 April 2007# Springer-Verlag 2007

Secretin-stimulating MRCP in patientswith pancreatobiliary maljunction and occultpancreatobiliary reflux: direct demonstrationof pancreatobiliary reflux

Abstract We propose the hypothesisthat the enlargement of the commonbile duct (CBD) or gallbladder (GB)that is occasionally demonstrated onmagnetic resonance cholangiopancre-atography (MRCP) after secretinstimulation is caused by pancreato-biliary reflux. Recently, occult pan-creatobiliary reflux (OPR) has beendemonstrated in patients withoutmorphological pancreatobiliary mal-junction (MPBM). The aim of thisstudy was to evaluate the efficacy ofsecretin-stimulating MRCP (SMRCP)in the diagnosis of pancreatobiliaryreflux. The study included 14 patientswith MPBM and 32 patients with anormal pancreatobiliary junction.OPR was evaluated by bile collectionand diagnosed in seven of the 32patients. All the patients underwentSMRCP; the related findings were

considered positive when enlargementof the CBD or GB was observed.Positive findings on SMRCP wereobserved in all MPBM patients. In thepatients with normal pancreatobiliaryjunction, there was significant differ-ence between the mean amylase levelsin the patients with positive andnegative SMRCP findings (mean,4,755.7 and 29.7 IU/l). The sensitivityand specificity of SMRCP for diag-nosing OPR was 85.7% and 68.0%,respectively. SMRCP provides a noninvasive method for excluding PBRand can identify patients who couldbenefit from bile duct sampling todiagnose OPR.

Keywords Secretin .Pancreatobiliary maljunction .Pancreatobiliary reflux . Gallbladdercarcinoma . MRCP

Introduction

Pancreatobiliary reflux (flow of pancreatic juice into thebiliary tract) is usually recognized as one of the mostimportant risk factors for biliary malignancies and isnormally caused by a morphological pancreatobiliarymaljunction (MPBM). MPBM is defined as the union ofthe main pancreatic duct (MPD) and the common bile duct(CBD) outside the duodenal wall where the sphinctermechanism is absent [1]. Thus, the absence of the sphinctermechanism due to the union of the MPD and CBD outsidethe duodenal wall results in the free passage of pancreaticjuice into the CBD. Several investigators have mentionedthat pancreatobiliary reflux induces chronic inflammationand increased proliferation of the biliary tract epithelium,

thereby leading to hyperplasia and metaplasia andultimately resulting in carcinoma of the biliary tract [2–5].

Since it may be impossible to visualize pancreatobiliaryreflux by using an imaging modality, the diagnosis ofMPBM by endoscopic retrograde cholangiopancreatogra-phy (ERCP) is considered to be the “gold standard” fordetecting the presence of pancreatobiliary reflux. However,recent investigations have demonstrated the occurrence ofpancreatobiliary reflux in patients with a normal pancrea-tobiliary junction [this has been termed as “occultpancreatobiliary reflux” (OPR)]; this was indicated by ahigher amylase level in CBD bile than in serum [6, 7]. Ithas been proved that OPR might also be a risk factor forbiliary malignancies [8]. As a result of this better under-standing, the diagnosis of pancreatobiliary reflux by

U. Motosugi (*) . T. Ichikawa .T. ArakiDepartment of Radiology,University of Yamanashi,1110 Shimokato,Chuo-shi, Yamanashi, 409-3898, Japane-mail: utaroh-motosugi@nifty.comTel.: +81-55-2731111Fax: +81-55-2736744

F. Kitahara . T. SatoFirst Department of Internal Medicine,University of Yamanashi,Chuo-shi, Yamanashi, Japan

J. Itakura . H. FujiiFirst Department of Surgery,University of Yamanashi,Chuo-shi, Yamanashi, Japan

merely using morphology-based imaging techniquesrepresented by ERCP is inadequate. Currently, there is acompelling need to establish a new imaging technique thatenables direct demonstration of pancreatobiliary reflux.

Several studies have reported that secretin-stimulatingmagnetic resonance cholangiopancreatography (SMRCP)may allow direct visualization of pancreatobiliary reflux[9]. Therefore, the purpose of this study was to evaluate theability of SMRCP to function as a comprehensivediagnostic tool for pancreatobiliary reflux and a tool thatcan detect this condition in OPR cases as well.

Materials and methods

The SMRCP study protocol was prospectively included inthe routine MR protocol for consecutive patients withsuspected pancreatobiliary diseases who visited ourinstitute between September 2002 and March 2005. Theexclusion criteria for SMRCP included a lack of informedconsent from the patient, the probable presence ofobstruction or severe stenosis of the CBD or MPD(which might render the observation of pancreatobiliaryreflux impossible), or an increased risk of developing acutepancreatitis. Finally, 191 patients successfully underwentSMRCP. This study was approved by the institutionalreview board of our hospital.

Diagnosis of MPBM and OPR

Based on the imaging analysis results of the morphologicalabnormality of the pancreatobiliary junction, an experi-enced radiologist and gastroenterologist attempted toretrospectively determine the number of MPBM patients.They were blinded to the results of the analysis of amylaselevels in bile. When the common canal was longer than15 mm, MPBM was merely morphologically diagnosedusing the ERCP images. The term “common canal”referred to the duct formed from the union of the MPDand CBD to the papilla of Vater. Finally, based on theERCP findings, 14 of the 191 patients, including one with a

congenital choledochal cyst, were diagnosed with MPBM.Of the 14MPBM patients, seven consented to bile collectionthat was performed to confirm the high amylase level.

Of the remaining 177 patients without MPBM, 60underwent ERCP for the evaluation of pancreatobiliarydisease. Of these, 32 agreed to participate in this study.

Bile was collected via a canula that was deeply insertedin the CBD during ERCP, followed by measurement of theamylase level in the collected bile (mean, 3.5 ml; range, 2–5 ml). In non-MPBM patients, amylase level measurementin the bile was employed as the gold standard fordiagnosing OPR; OPR was diagnosed when the amylaselevel in the collected bile was higher than 115 IU/l, whichis the upper limit of the serum amylase value.

Demonstration of pancreatobiliary reflux in MPBMand OPR patients by using secretin stimulation

The present study proposed the hypothesis that theobservation of pancreatobiliary reflux after secretin stim-ulation reveals the presence of MPBM or OPR. However,whether the enlargement of the gallbladder (GB) or CBDreally represents the reflux of pancreatic juice remained inquestion. To confirm this hypothesis, bile was collectedand the amylase level was subsequently measured twice(prior to and 3 min after secretin stimulation) in twopatients after obtaining their informed consent. Thesepatients had been diagnosed with MPBM (n=1) or withpossible OPR (n=1; based on the negative morphologicalfindings in ERCP and positive findings in SMRCPimaging). Bile was intraoperatively collected from onepatient and endoscopically collected from the other.

SMRCP imaging

SMRCP was performed in all patients by using acommercially available 1.5-Tesla superconducting MRunit (Sigma Lx; GE Healthcare, Milwaukee, Wis.). TheSMRCP images included two types of MRCP images:those obtained by the thin-slab, multislice MRCP tech-

Table 1 The results of SMRCP

aOPR was diagnosed when am-ylase level of the bile was higherthan 115 IU/l

Positive SMRCP NegativeSMRCPDynamic enlargement Slight GB

enlargementDynamic CBDenlargement

Dynamic GBenlargement

MPBM 10 4 0 0

OPRa 3 1 2 1

Non-OPR

0 7 1 17

2263

nique (hereafter referred to as multislice MRCP) and thoseobtained with the thick-slab, double-slice sequentialMRCP technique (hereafter referred to as dynamicMRCP). The SMRCP parameters were as follows: se-quence, single-shot fast spin-echo; echo time and slicethickness, 80–95 ms and 5 mm for multislice MRCP and800–900 ms and 30 mm for dynamic MRCP, respectively;field of view, 20 cm; and acquisition time/slice, 2 s.

First, axial multislice MRCP images were obtained priorto secretin stimulation for use as a reference to evaluate theSMRCP images. Immediately after the intravenous admin-istration of 50 IU of secretin (Secrepan; Eisai, Tokyo,Japan), dynamic MRCP imaging was repeated every 15 sfor the first 5 min and every 30 s for another 5 min. Thus, atotal of 30 frames were included in one set of dynamicMRCP images obtained during the 10 min followingsecretin administration. Each frame of the dynamic MRCPimages included two MRCP images that were simulta-neously obtained when different slice directions were used.Two slice directions—the coronal plane for visualizing thepancreatobiliary junction and the oblique plane forobserving the GB—were employed. Following dynamicMRCP imaging, post-secretin stimulation multisliceMRCP images were obtained by following the sameprocedure as that used to obtain the pre-secretin images.

Assessment of SMRCP

All SMRCP images were independently interpreted by twoexperienced radiologists (U.M. and T.I.). In case ofdisagreement, a third radiologist (T.A.) was consulted.Pancreatobiliary reflux was considered positive when thefollowing findings were obtained from the SMRCPimages: (1) The CBD was gradually enlarged in thesequential dynamic MRCP (dynamic CBD enlargement),(2) the CBD was not enlarged, but the GB was graduallyenlarged in the sequential dynamic MRCP (dynamic GBenlargement), or (3) neither the GB nor CBD was enlargedin the sequential dynamic MRCP, but the GB was enlargedin the post-secretin stimulation multislice axial MRCPimages compared with the pre-secretin stimulation multi-slice axial MRCP images (slight GB enlargement). The GBand CBD were judged as enlarged when the increase intheir diameter was over 2 mm.

By using the Mann-Whitney U-test, the amylase levelsin the bile were compared between patients with positiveSMRCP findings and those with negative SMRCPfindings. The diagnostic ability (sensitivity and specificity)of SMRCP for detecting OPR was calculated.

Results

Amylase levels in the bile and OPR diagnosis

The mean amylase level in the bile in the MPBM and non-MPBM patients were 28,028 IU/l (185–328,500 IU/l) and2,097 IU/l (0–48,100 IU/l), respectively.

Among the 32 non-MPBM patients, OPR was diagnosedin seven based on the high amylase levels in the bile (mean,9530 IU/L; individual levels, 135, 232, 280, 411, 6,790,

Fig. 1 a Schema of normal pancreatobiliary junction and b MPBM.Note that the union of common bile duct and pancreatic duct is outside the sphincter muscle in the case of the MPBM. c RepresentativeERCP of the MPBM. Anomalous union of the common bile ductand pancreatic duct leads to pancreatobiliary reflux and thedevelopment of biliary malignancies. Schema a and b wereoriginally created by K. Suda

2264

10,765, and 48,100 IU/l); the remaining 25 patients wereconsidered as non-OPR (mean, 16 IU/l; range, 0–77 IU/l).

Efficacy of SMRCP imaging for MPBM and OPRdiagnoses

All 14 MPBM patients had positive SMRCP findings ondynamic MRCP (100% sensitivity). Dynamic CBD en-largement was observed in ten patients and dynamic GBenlargement, in four patients (Table 1).

Among the 32 non-MPBM patients, 14 were judged aspositive, and the remaining 18 were judged as negativebased on the SMRCP findings. The mean amylase level inthe positive and negative groups was 4,755.7 (range, 1–48,100) and 29.7 (range, 0–270) IU/l, respectively. Theamylase levels in patients with positive SMRCP findingswere significantly higher than those in patients withnegative findings (P<0.002).

Of the seven patients who were suspected of havingOPR based on the increased amylase levels in the bile, sixwere found to be positive by SMRCP (85.7% sensitivity).Dynamic CBD enlargement was observed in three of thesix patients. Of the remaining 25 non-OPR patients, eightwere judged as positive by SMRCP imaging (68.0%specificity). None of the eight patients showed CBDenlargement on dynamic MRCP (Table 1).

Demonstration of pancreatobiliary reflux after secretinstimulation in two cases of MPBM and OPR

In the MPBM patient, the amylase levels in the bile prior toand 3 min after secretin stimulation were 157 IU/l and12,050 IU/l, respectively, while they were 135 IU/l and8,220 IU/l in the patient with possible OPR. Thus,pancreatobiliary reflux under secretin stimulation wasconfirmed in both the patients based on the remarkable

Fig. 2 a - c A case of GB car-cinoma with MPBM. a MRCPimaging before secretin injec-tion, b MRCP imaging 3 minafter secretin injection; SMRCPdemonstrates dynamic enlarge-ment of the CBD. The GB didnot enlarge after the secretinstimulation because a tumor waslocated at the neck of the GB,and the cystic duct was oc-cluded. c ERCP showed ananomalous union of the CBDand pancreatic duct (arrow).This patient’s bile was intra-operatively collected. The bili-ary amylase level after secretinstimulation (12,050 IU/l) wasclearly higher than that beforestimulation (157 IU/l)

2265

increase in the amylase levels after secretin stimulation(Figs. 1, 2, 3).

Discussion

Secretin is sometimes used to let the pancreatic duct clearon MRCP. Secretin injection leads to pancreatic exocrinesecretion and increases signals of the pancreatic duct andduodenum on MRCP. Recent study also revealed thatSMRCP can measure pancreatic exocrine function and beused as a therapy evaluation in patients with chronicpancreatitis [10]. However, an efficacy of SMRCP forMPBM or OPR has not been well studied.

Numerous studies have demonstrated a relationshipbetween pancreatobiliary reflux and biliary cancer inpatients with MPBM [11, 12]. The frequency of biliarycancer in adult patients with MPBM ranges from 15% to24%, and associated GB cancer is observed in 18–67% ofpatients without choledochal cyst [12]. Pancreatobiliary

reflux can induce genetic changes, such as K-ras pointmutations, that may result in carcinogenesis in the biliarytract epithelium in MPBM patients [13]. Even if the CBD isnot dilated, preventive cholecystectomy is recommended atleast for patient with MPBM.

It was recently reported that the reflux of pancreatic juiceinto the CBD was observed in patients with a normalpancreatobiliary junction (OPR); in these patients, thebiliary amylase level was higher than the upper limit of theserum amylase level. In a previous report, OPR as well asMPBM was suggested to be an important risk factor for thedevelopment of GB carcinoma [8]. In this study, the resultsof endoscopic bile collection support the hypothesis thatSMRCP can be used to diagnose OPR. SMRCP isconsiderably sensitive in detecting MPBM and OPR;hence, a negative result in SMRCP imaging indicates theabsence of pancreatobiliary reflux. Although bile collec-tion is the only available method to diagnose OPR, it ispossible to use SMRCP as a screening tool [6, 7].

Fig. 3 a - c A case of normalpancreatobiliary junction (non-MPBM). a MRCP imaging be-fore secretin injection, b MRCPimaging 5 min after the secretininjection; SMRCP demonstratesdynamic enlargement of the GB.c Both the bile duct and pan-creatic duct were demonstratedas separate which indicate anormal pancreatobiliary junction(arrow). Bile was endoscopi-cally collected during ERCPbefore and after secretin stimu-lation. Before secretin injection,the amylase level of the col-lected bile (135 IU/l) was higherthan the upper limit of the serumamylase level. The amylaselevel increased remarkably(8,220 IU/l) after secretininjection

2266

Among the seven OPR patients in this study, onjeunderwent cholecystectomy for GB stones, one changedhospitals and thus follow-up was not possible, and theremaining five have been regularly followed up usingabdominal ultrasound and MRCP. To date, no biliarymalignancy has been detected in these OPR patients.

Positive results in SMRCP imaging, however, do notalways indicate the reflux of pancreatic juice; in particular,enlargement of the GB was considered to be a non-specificfinding [14]. Considering the result of 68% specificity,32% of the normal patients examined by SMRCP will haveto undergo unnecessary bile sampling to exclude OPR. Oneof the reasons for the high false-positive rate is that bile issecreted after secretin stimulation, as reported [14, 15].Therefore, CBD or GB enlargement implies pancreatobili-ary reflux, bile secretion, or both. The two causes ofenlargement cannot be differentiated unless the pancreaticjuice that is continuously ascending toward the CBD can bedirectly visualized by SMRCP. However, in this study,dynamic CBD enlargement was observed only in patientswith high biliary amylase levels. This may imply that CBDenlargement observed in SMRCP can indicate the presenceof OPR.

In addition, intravenous secretin injection may stronglystimulate patients compared with a stimulation of physi-ological dietary intake; therefore, in some patients, refluxmay occur only if secretin is administered. Reflux underthis unnatural condition does not always indicate consis-tently high biliary amylase levels. Additional research maybe required to determine whether patients that exhibitpositive SMRCP findings can be classified as being “at ahigh risk for biliary malignancies.”

Conclusion

SMRCP provides a noninvasive method for excludingpancreatobiliary reflux and can identify patients who couldbenefit from bile duct sampling to diagnose OPR.

Acknowledgements We thank our departmental radiologists, T.Araki, A. Nambu, H. Nakajima, A. Ito, M. Hori, K. Ishigame, H.Sou, M. Nagata, K. Sano, and E. Jin, for collecting the MR data. Wegratefully acknowledge the continuing assistance of the doctors fromthe First Department of Internal Medicine and the First Departmentof Surgery for their help with regard to bile collection.

References

1. Matsumoto Y, Fujii H, Itakura J et al(2002) Recent advances in pancreati-cobiliary maljunction. J HepatobiliaryPancreat Surg 9:45–54

2. Tashiro S, Imaizumi T, Ohkawa H et al(2003) Pancreaticobiliary maljunction:retrospective and nationwide survey inJapan. J Hepatobiliary Pancreat Surg10:345–351

3. Matsubara T, Sakurai Y, Sasayama Y etal (1996) K-ras point mutations incancerous and noncancerous biliaryepithelium in patients with pancreati-cobiliary maljunction. Cancer 77(S8):1752–1757

4. Kaneko K, Ando H, Ito T et al (1996)Increased cell proliferation and trans-forming growth factor-alpha (TGFalpha) in the gall-bladder epithelium ofpatients with pancreaticobiliary mal-junction. Pathol Int 46:253–260

5. Matsubara T, Funabiki T, Jinno O et al(1999) p53 gene mutations and over-expression of p53 product in cancerousand noncancerous biliary epithelium inpatients with pancreaticobiliary mal-junction. J Hepatobiliary Pancreat Surg6:286–293

6. Sai JK, Ariyama J, Suyama M et al(2002) Occult regurgitation of pancre-atic juice into the biliary tract: diagno-sis with secretin injection magneticresonance cholangiopancreatography.Gastrointest Endosc 56:929–932

7. Sai JK, Suyama M, Kubokawa Y et al(2003) Occult pancreatobiliary reflux inpatients with a normal pancreaticobili-ary junction. Gastrointest Endosc57:364–368

8. Itoi T, Tsuchida A, Itokawa F et al(2005) Histologic and genetic analysisof the gallbladder in patients withoccult pancreatobiliary reflux. Int J MolMed 15:425–430

9. Hosoki T, Hasuike Y, Takeda Y et al(2004) Visualization of pancreatico-biliary reflux in anomalouspancreaticobiliary junction by secretin-stimulated dynamic magneticresonance cholangiopancreatography.Acta Radiol 45:375–382

10. Bali MA, Sztantics A, Metens T et al(2005) Quantification of pancreaticexocrine function with secretin-en-hanced magnetic resonance cholangio-pancreatography: normal values andshort-term effects of pancreatic ductdrainage procedures in chronic pancre-atitis. Initial results. Eur Radiol15:2110–2121

11. Kimura K, Ohto M, Saisho H et al(1985) Association of gallbladder car-cinoma and anomalous pancreatobiliaryductal union. Gastroenterol 89:1258–1265

12. Komi N, Kunitomo K, Tamura T et al(1984) Nationwide survey of cases ofcholedochal cyst: analysis of coexistantanomalies, complications and surgicaltreatment in 645 cases. Surg Gastro-enterol 3:69–73

13. Matsubara T, Sakurai Y, Sasayama Y etal (1996) K-ras point mutations incancerous and noncancerous biliaryepithelium in patients with pancreati-cobiliary maljunction. Cancer 77:1752–1757

14. Song HK, Kim MH, Lee SK et al(2003) Progressive bile duct and gall-bladder dilation on MRCP after secretinstimulation: physiologic or pathologicfinding? Gastrointest Endosc 58:165,author reply 165–166

15. Kanno N, LeSage G, Glaser S et al(2001) Regulation of cholangiocytebicarbonate secretion. Am J PhysiolGastrointest Liver Physiol 281:612–625

2267