Diagnostic pitfalls in the cholangiographic diagnosis ofcholedochoceles: cholangiographic quality and its effect onvisualization

K. B. Park,1* Y. H. Auh,1 J. H. Kim, 1 M. G. Lee,1 H. K. Ha,1 P. N. Kim,1 Y. M. Shin,1 M. H. Kim, 2

H. J. Kim, 2 Y. I. Min 2

1Department of Diagnostic Radiology, Asan Medical Center, University of Ulsan, College of Medicine, Seoul, Korea2Department of Internal Medicine, Asan Medical Center, University of Ulsan, College of Medicine, Seoul, Korea

Received: 20 January 2000/Revision accepted: 31 May 2000

AbstractBackground:We wanted to establish reasonable cholan-giographic diagnostic criteria by determining the sensi-tivity of cholangiography in detecting choledochocelesand those factors that could compromise visualization ofcholedochoceles.Methods:Over 4 years, 21 patients (seven male, 14 fe-male; mean age5 67 years) were confirmed as havingcholedochoceles on endoscopic retrograde cholangiopan-creatography (ERCP). Cholangiographic diagnosis wasmade by following three criteria: a radiolucent haloaround the distal common bile duct (CBD), bulbous di-latation of the distal CBD, and the presence of sequentialmorphologic changes on serial cholangiography. Any twoor more combinations of these three criteria were consid-ered enough to diagnose a choledochocele on cholangiog-raphy. We compared cholangiographic imaging findingswith the ERCP results.Results:Of 21 patients with choledochoceles, nine (43%)were correctly diagnosed on cholangiography. A radiolu-cent halo was present in six (28%) patients; four of thesecases showed optimal duodenal filling, one showed faintduodenal filling, and one showed poor duodenal filling.The shapes of the distal CBD were bulbous, conelike, andblunt. Morphologic changes such as collapsing and bulg-ing of the choledochocele could be seen in 12 (57%)patients on serial cholangiography. Waists were seen in11 (52%), pseudowebs in four (19%), and wrinkling ofthe distal CBD in seven (33%).

Conclusion: Cholangiography should be obtained withoptimal timing and adequate conditions to diagnose cho-ledochocele correctly.

Key words: Bile ducts, cysts—Bile duct radiography—Endoscopic retrograde cholangiopancreatography.

Choledochocele is a cystic or diverticulum-like dilatationof the intramural segment of the distal common bile duct(CBD) that protrudes into the duodenal lumen. Althoughthe exact etiology is unknown, investigators have sug-gested two different distinct types, congenital and ac-quired, of this abnormality. According to Alonso et al. [1]and Todani et al. [2], a choledochocele is generally clas-sified as a type III congenital choledochal cyst.

The choledochocele is a very rare abnormality. In1988, Sarris and Tsang extensively reviewed the English-language literature and found only 48 cases publishedbefore 1984 [3]. Schmidt et al. reviewed 9850 patientswho underwent endoscopic retrograde cholangiopancre-atography (ERCP), and only 10 (0.1%) were found tohave choledochocele [4]. The reported incidence rate withERCP in the recent English-language literature rangedfrom 0.1% to 2.0% [4, 5]. Despite the low incidence ratein previous reports, the progressively increasing rate ofdetection shows that the choledochocele is not an uncom-mon disease and that it can be detected more often by avery carefully designed scheme of ERCP and cholangiog-raphy taken with optimal radiographic techniques.

The diagnosis of choledochocele is important becauseit may cause idiopathic recurrent pancreatitis [6], biliarycolic due to associated biliary stone disease, cholestatic

*Present address:Department of Radiology, Samsung Medical Center,50, 135-710, Ilwon-Dong, Kangnam ku, Seoul, Korea

Correspondence to:K. B. Park

Abdom Imaging 26:48–54 (2001)DOI: 10.1007/s002610000114 Abdominal

Imaging© Springer-Verlag New York Inc. 2001

jaundice [7], and ascending cholangitis, and it may de-generate in a malignant manner. ERCP is the diagnosticmethod of choice for choledochocele [8]. The diagnosisof choledochocele and reported rate of incidence havebeen based mainly on criteria of endoscopic diagnosisduring ERCP. Cholangiographic findings have not beenthoroughly investigated in the radiologic literature untilrecently.

The purpose of this study was to establish reasonablecholangiographic criteria of choledochocele by determin-ing the sensitivity of cholangiography in detecting chole-dochoceles and by determining those factors that couldcompromise visualization of choledochoceles.

Materials and methods

Between January 1993 and November 1996, we retro-spectively evaluated 3858 consecutive patients who un-derwent ERCP. Patients whose ERCP had failed to can-nulate the papilla were excluded from the study. With thereview of ERCP and medical records, 21 patients wereconfirmed as having choledochocele (0.6%). Seven weremale and 14 were female, with an age range of 49–85years (mean age5 67 years).

ERCPs were performed by an endoscopist who used aside-view duodenoscope (Olympus, Tokyo, Japan). Oro-pharyngeal anesthesia was provided with lidocaine spray(xylocaine 10% spray; Astra, Linz, Austria) before theprocedure. The endoscope was then introduced into theduodenum. After the identification of papilla, cannulationof the CBD was attempted, followed by an injection of a30% solution of contrast medium (Telebrix, 30 Meglu-mine, Laboratoire Guerbet, France) into the CBD lumen.Immediate and delayed spot film radiographs were ob-tained with digital fluoroscopy (MEDIX DR-2000 MCSystem, Hitachi, Japan). The diagnosis of choledochocelewas made on ERCP by inspecting the ampulla of Vaterand on cholangiography during the procedure; the finalreview of the video tapes was performed by two endos-copists to confirm the diagnosis. Conclusive diagnosiswas established by reviewing the video tapes of the en-doscopic procedures. The following criteria were used toestablish the diagnosis of the choledochocele during en-doscopy: (a) spherical or pear-shaped, enlarged bulgingof the papilla, (b) ballooning of the choledochocele dur-ing injection of contrast medium, (c) morphologicchanges of spherical, cystlike, contrast-filled dilatationsalong the duodenal peristalsis, (d) soft overlying mucosawith a smooth appearance, and (e) no impacted stone inthe papilla.

An experienced gastrointestinal radiologist retrospec-tively reviewed cholangiographies obtained during ERCPof 21 patients. We established three major criteria in thecholangiographic diagnosis of choledochocele to deter-mine how many patients fulfilled these criteria: (a) a

radiolucent halo around the distal end of the CBD, (b)bulbous end of the distal CBD, and (c) dynamic sequen-tial morphologic changes of the distal CBD. Patients whomeet two or more combinations of these criteria werediagnosed as having choledochocele on cholangiography.In addition to these three diagnostic criteria, we observedthe presence of a waist, a pseudoweb, and wrinkling ofthe distal CBD.

The cases were reviewed to determine whether spe-cific technical factors affected visualization of chole-dochocele. These factors included degree of distention of

Fig. 1. Different manifestations of the choledochocele according to theshape of the distal CBD. Types A to C basically have a bulbous sac; typeA also has a collapsed sac, type B has cone-shaped ends, and type C hasa blunt end. Type D basically has a blunt end combined with cone-shaped ends. Type E has a constant bulbous end. Type F has a blunt end.Type G has a cone-shaped end.

Fig. 2. A 59-year-old woman with a typical choledochocele. Four spotserial cholangiographs clearly show the changes in shape of the chole-dochocele sac, from dilated to collapsed (arrows). This choledochoceleis morphologic type A.

K. B. Park et al.: Diagnostic pitfalls in the cholangiographic diagnosis of choledochoceles 49

the distal CBD, adequate sequential images showingopening and closure of the sphincter complex, filling ofduodenal lumen, patient position and image plane, andtime of exposure. We graded the degrees of duodenallumen filling to four scales: overfilling, optimal, faint, andnonfilling. Contrast filling of the distal CBD was classi-fied as complete or incomplete. It was difficult to definethe shape of the distal CBD as a single type because itchanged during the filming sequence. Therefore, theshapes of the distal CBD were classified into seven typesaccording to the presence of sequential morphologicchanges (Fig. 1).

Cholangiographic films showing at least two phasesof distal CBD shape, such as collapsed and dilated, wereconsidered as having adequate exposure time. The posi-tions of patients were regarded as optimal when the

endoscope did not overlap with the contrast-filled distalCBD and when the image was taken in profile view.

Results

Morphologic characteristics and cholangiographicconditions

A radiolucent halo around the distal CBD was noted in six(28%) patients, and the degree of duodenal filling forthese patients was optimal in four, faint in one, and nonein one. In the case where the duodenum was not filled, ahalo could be seen as a result of air contrast, i.e., gaseousdistention of the second duodenal loop helped to visualizethe even thickness of the radiolucent band around the

Fig. 3. Atypical features of a chole-dochocele as confirmed on ERCP by typi-cal ballooning of the papilla in threedifferent patients.A A 43-year-old womanwith type F choledochocele. There is bluntending of the distal CBD (arrow) oncholangiography, which is inadequate fordiagnosis.B A 38-year-old woman withtype G choledochocele. A pseudoweb isvisible just above the choledochocele (ar-rowhead). C A 28-year-old man with typeG choledochocele. There are kinking (ar-row) and fine inner wall irregularities(wrinkling) of the distal CBD (arrowheads)with cone-shaped ends.

50 K. B. Park et al.: Diagnostic pitfalls in the cholangiographic diagnosis of choledochoceles

bulbous end of the distal CBD. In the evaluation of theradiolucent halo, the most important cause of detectionerror was inadequate filling of the duodenum.

Sequential changes of the choledochocele such ascollapsing and bulging were present in the cholangiogra-phies of 12 (57%) patients (Fig. 2). The typical bulbousshape of the distal CBD was seen in 13 (62%) patients.The different appearances of the distal CBDs are shownin Figures 1 and 3. For the evaluation of the shape of thedistal CBD and the sequential changes in morphology,exposure time was the most critical factor for successfuldemonstration of these findings (Fig. 4). Causes for non-visualization of these three major diagnostic findings aresummarized in Table 1.

Of these 21 patients with choledochocele, only five(24%) patients fulfilled the three major cholangiographiccriteria completely. Four patients fulfilled two of thesecriteria. Thus, with these combinations, we correctly di-agnosed nine (43%) patients as having choledochocelebased on cholangiography alone. Principal causes of di-agnostic failure in each patient were attributed to incom-plete filling of the distal CBD in five, incomplete filling ofthe duodenum in 10, inadequate timing of cholangiogra-phy in six, and suboptimal positioning of patients in two.

CBD waists were seen in 11 (52%) patients. Waistswere defined as very short segmental luminal narrowingsof the CBD that mimicked the shape of an hourglass, andthese waists could be seen just above the dilated bulboussac in most cases. Pseudowebs were seen in four (19%)patients, and these appeared as a narrow radiolucent bandwith even thickness just above the sac of choledochocele(Fig. 3B). Pseudowebs did not cause obstruction of theCBD. Wrinkling of the distal CBD was seen in seven(33%) patients, and fine irregularities of the inner wall ofthe distal CBD were visible (Fig. 3C). Wrinkling of the

CBD was better visualized when the choledochocele wascollapsed.

Exposure time was adequate in 11 but inadequate in10 patients. Adequate exposure time coincided with thepresence of sequential shape changes in the distal CBD.Positioning was optimal in 12 patients and suboptimal innine. Filling of the CBD was complete in 15 patients andincomplete in six. The degree of duodenal filling wasoverfilling in two, optimal in four, faint in three, andnonfilling in 12 patients (Fig. 5A–C).

Table 1. Major diagnostic findings of choledochocele and causes ofnonvisualization of each finding

Findings Present Absent Causes ofnonvisualizationof findings

npatients

Radiolucenthalo

6 15 Inadequate fillingof duodenum

Overfilling 2Nonfilling 8

Incomplete fillingof CBD

2

Suboptimal positioningof patients

2

Inadequate timing 1Bulbous end 13 8 Incomplete filling

of CBD2

Suboptimal positioningof patients

2

Inadequate timing 4Change ofCBD shape

12 9 Incomplete fillingof CBD

2

Inadequate timing 7

Fig. 4. A 46-year-old man withtype B choledochocele demon-strates inadequate and adequatecholangiographic visualization.A Exposure time was inadequatewhen the choledochocele wascollapsed. The duodenum wasnot filled, and there is incom-plete expansion of the chole-dochocele.B Adequatecholangiography clearly showsthe choledochocele sac (arrow)and surrounding radiolucent halo(arrowhead). Exposure time wasadequate, duodenal filling wasoptimal, and filling of the CBDalso was optimal.

K. B. Park et al.: Diagnostic pitfalls in the cholangiographic diagnosis of choledochoceles 51

Associated diseases

Biliary stone diseases were present in 12 (57%) patients.Two patients had gallbladder (GB) stones, eight had CBDstones, and two had both GB and CBD stones. A Klatskintumor and a type I choledochal cyst were incidentlynoted. Two patients had acute recurrent pancreatitis. Inthese cases, the pancreatic duct communicated directlywith the choledochocele, and spontaneous retrograde fill-ing of the pancreatic duct was noted on ERCP. Thesebelonged to type A3 of the Sarris classification (Table 2).All patients were treated with endoscopic sphincterot-omy.

Discussion

The incidence of choledochoceles has been increasingcontinuously since 1990, and this increase is mainly dueto the technical advance of ERCP. ERCP is presentlyaccepted as the diagnostic procedure of choice for preop-erative diagnosis, and it has the highest sensitivity andspecificity when compared with other diagnostic tools [4,8]. However, we wondered how many cases would bediagnosed as choledochocele if the radiologist read onlythe cholangiography films. Although many institutionsevaluate both the endoscopic and cholangiographic as-pects equally, cholangiography is usually not obtainedunder the best exposure conditions and is sometimesignored because of clear duodenscopic evidence.

Duodenoscopic features of choledochocele are char-acteristic. According to Kim et al., the choledochoceleappears hemispherical or as a pear-shaped bulge protrud-ing into the duodenal lumen; the overlying mucosa is softand smooth, is readily compressible with a cannulatingcatheter, and inflates during the injection of contrast me-dium [9]. However, 53% of choledochoceles did not showabnormality of the papilla on initial inspection of ERCPbefore the injection of contrast materials. In these pa-tients, the papilla was normal or even flat. However,during the injection of contrast materials, the papilla

Fig. 5. The importance of duodenal filling in the diagnosis of choledochocele in three different patients.A A 63-year-old woman with type G choledochocele. Inadequate cholangiography was obtained withoutcontrast filling of the duodenal lumen. The distal CBD shows spherical dilatation with a cone-shaped end(arrowheads) and proximal waist formation (arrow). However, it was impossible to diagnose this chole-dochocele on cholangiography because we could not be sure that the dilated sac was in the intramuralportion.B A 57-year-old man with type E choledochocele. Faint filling of duodenum shows a radiolucenthalo (white arrow) around the choledochocele. Note the combined GB stones (black arrow). C A 26-year-old man with a choledochocele. Overfilling of the duodenum shows a halo (arrowheads), but theradiopaque barium pool slightly masks this halo. However, we can be sure that the dilated sac is in theintramural portion. Note the combined CBD stone (arrow).

Table 2. Classification of shape of distal CBD

Types npatients

A. Bulbous and collapsed sac 4B. Bulbous and smooth tapered or cone-shaped end 5C. Bulbous and blunt end 2D. Blunt and cone-shaped end 1E. Bulbous end 3F. Blunt 2G. Cone shaped 4

52 K. B. Park et al.: Diagnostic pitfalls in the cholangiographic diagnosis of choledochoceles

bulged. Ballooning of the papilla during contrast injectioncould not be seen in cases without a choledochocele [9].

Choledochocele is a disease with dynamic features;therefore, it is crucial to use a diagnostic modality that isable to capture these dynamic changes, e.g., fluoroscopy.In the cholangiographic diagnosis of choledochocele, themost important causes of diagnostic failure were incom-plete filling of the duodenal lumen and inadequate expo-sure time. In our study, only five of 21 patients com-pletely fulfilled the three diagnostic criteria ofcholangiography. However, most of these cholangio-graphs are not obtained with optimal exposure becauseendoscopists are often satisfied with the endoscopic datasuch as video tapes or even inspection. In addition, it isimportant to remember that failure to find indications ofcholedochoceles is not always the result of technicalfactors; some choledochoceles are so small or show verymild degrees of prolapse that they cannot be readilydetected.

If an endoscopist observes bulging of the papilladuring ERCP, an adequate cholangiogram is useful asobjective evidence of choledochocele. Despite its lowdiagnostic efficacy, cholangiography obtained duringERCP can play an important role in the management ofcholedochoceles. First, cholangiography shows the entiresac of the choledochocele when it is filled with contrastmaterial. Second, it provides exact information about thelength of the involved segment, thus determining appro-priate therapy. Third, cholangiography shows the detailedmorphology of the distal CBD, e.g., the presence ofirregularity in the inner wall of the CBD, the presence ofcongenital anomalous conditions in the pancreaticobiliaryduct system, and the presence of a web or an intraductalfilling defect. Fourth, it clearly shows the presence ofcombined stone disease or malignant stricture. Therefore,it is essential to obtain high-quality cholangiography forthe patient with a choledochocele.

A radiolucent halo can be seen between the contrast-filled dilated sac and the contrast-filled duodenal lumen.Filling of the duodenal lumen is very important to identifythe dilated end of the distal CBD in the intramural portionand to demonstrate the radiolucent halo. For this purpose,the duodenal lumen should be filled with an optimalamount of contrast material, and the patient should bepositioned so that the dilated sac does not overlap with thesecond duodenal portion or the endoscope. Cholangiog-raphy should be done in profile. When adequate cholan-giographic images are combined with a careful ERCPexamination, it is possible to diagnose small chole-dochoceles that may be overlooked by inspection duringERCP alone.

Most reported choledochoceles have a clubbed endrather than a tapered end, and, in general, the CBD abovethe choledochocele is not dilated. However, cholangiog-raphy obtained when the choledochocele is collapsedshows the end of the distal CBD as a cone shape or blunt.

In addition, optimal filling of the CBD lumen is essentialto demonstrate the exact shape of the distal CBD. Venu etal. stated that optimal radiographic technique duringERCP is necessary and proposed that a cannulating cath-eter should be withdrawn from the duodenum after con-trast material has been instilled into the bile duct [10].This is necessary to avoid masking the dilated sac by theballoon and the catheter. Both incomplete contrast fillingof the CBD lumen and exposure during collapse of theCBD make depicting the bulbous dilatation of the distalCBD difficult. Therefore, one cannot confirm the diagno-sis of choledochocele by the shape of the distal CBD oncholangiography obtained with inadequate exposure timeor with a single phase of serial cholangiographic films.For an endoscopist, it is sometimes difficult to obtain anexact view when the choledochocele is maintained inoptimal dilatation. However, a radiologist should encour-age and inform the endoscopist to obtain a set of twodifferent phases of the choledochocele or should activelyparticipate in ERCP procedure.

A pathophysiology of choledochocele formation hasbeen proposed by many investigators, but it is still un-clear. Choledochoceles can be lined by either duodenalmucosa or biliary epithelium [3]. Neuromuscular incoor-dination of the sphincter of Oddi, inflammatory stenosisof the papilla, abnormal angulation and insertion of theampulla, or an anomalous union of the pancreatobiliarysystem may cause progressive dilatation of the ampullaby way of an inflammatory process induced by stasis ofbile and pancreatic juice. In this study, inner wall irreg-ularity or wrinkling of the distal CBD was noted in sevenpatients. These were probably due to the sequelae ofprevious recurrent inflammations involving the distalCBD or simply represent the collapsed state of the cho-ledochocele. Chronic inflammation leads to epithelialmetaplasia, and the incidence of carcinoma has beenreported to be as high as 20% in cases of choledochocelewhen there is biliary or undifferentiated epithelia [2, 11,12]. Ladas et al. reported an 2.5% incidence rate ofcarcinoma in association with choledochocele [13]. Thepseudoweb could not be proved surgically. However, weobserved a radiolucent band of even thickness in four of21 patients (19%). It is unclear whether this band indi-cated a pseudolesion between the dilated sac and thenormal duct. This lesion does not induce obstruction ofbile duct.

In conclusion, an optimally obtained cholangiographywith adequate duodenal filling, CBD filling, proper pa-tient positioning, and adequate exposure time is essentialfor correctly diagnosing a choledochocele. More investi-gations are needed with regard to the pathologic correla-tion of several additional observations such aspseudowebs or waists and inner wall wrinkling or irreg-ularities.

K. B. Park et al.: Diagnostic pitfalls in the cholangiographic diagnosis of choledochoceles 53

References

1. Alonso LF, Rever WB, Pessagno DJ. Congenital choledochal cyst,with a report of two and an analysis of 94 cases.Int Abstr Surg1959;108:1–30

2. Todani T, Watanabe Y, Narusue M, et al. Congenital bile ductcysts: cysts: classification, operative procedures and review of thirtyseven cases including cancer arising from choledochal cyst.Am JSurg1977;134:263–269

3. Sarris GE, Tsang D. Choledochocele: case report, literature reviewand a proposed classification.Surgery1989;105:408–414

4. Schmidt HG, Bauer J, Wiessner V, et al. Endoscopic aspects ofcholedochoceles.Hepato-Gastroenterology1996;43:143–146.

5. Hanson BL, Bosco JJ, Muggia RA, et al. Choledochoecles(CDC) inadults:an underdiagnosed cause of pancreatobiliary symptoms [ab-stract].Gastrointest Endosc1994;40(suppl):10

6. Greene FL, Brown JJ, Rubinstein P, et al. Choledochocele andrecurrent pancreatitis.Am J Surg1985;149:306–309

7. Yamaoka K, Tazawa J, Koizumi K, et al. Choledochocele with

obstructive jaundice: a case report and a review of the Japaneseliterature.J Gastroenterol1994;29:661–664

8. Masetti R, Antinori A, Coppola R, et al. Choledochocele: chang-ing trends in diagnosis and management.Surg Today1996;26:281–285

9. Kim MH, Myung SJ, Lee SK, et al. Ballooning of the papilla duringcontrast injection: the semaphore of a choledochocele.GastrointestEndosc1998;48:258–262

10. Venu RP, Geenen JE, Hogan WJ, et al. Role of endoscopic retor-grade cholangiopancreatography in the diagnosis and treatment ofcholedochocele.Gastroenterology1984;87:1144–1149

11. Rossi RL, Silverman ML, Braasch JW, et al. Cancer arising in acystic conditions of the bile ducts.Ann Surg1987;205:377–384

12. Iwai N, Deguchi E, Yanagihara J, et al. Cancer arising in a chole-dochal cyst in a 12-year-old girl.J Pediatr Surg1990;25:1261–1263

13. Ladas SD, Katsogridakis I, Tassios P, et al. Choledochocele, anoverlooked diagnosis: report of 15 cases and review of 56 publishedreports from 1984 to 1992.Endoscopy1995;27:233–239

54 K. B. Park et al.: Diagnostic pitfalls in the cholangiographic diagnosis of choledochoceles