Anderson et Gastrointestinal Imaging • Clinical Obser vations

al.
MDCT
Diagnosis of
Choledochol
ithiasis

A C E N T U
R Y O F

Accuracy of MDCT in the

MEDICAL IMAGING
Diagnosis of Choledocholithiasis
Stephan W. Anderson1 OBJECTIVE. Our purpose was to evaluate the diagnostic performance of contrast-enhanced
Brian C. Lucey and unenhanced MDCT, performed for various indications, in detecting choledocholithiasis.
Jose C. Varghese CONCLUSION. Unenhanced and contrast-enhanced MDCT images, interpreted in
Jorge A. Soto PACS workstations with axial images, are moderately sensitive and specific for showing cho-
ledocholithiasis.
Anderson SW, Lucey BC, Varghese JC, Soto
JA holedocholithiasis is a common In the emergency setting, CT scans are be-

C complication of cholelithiasis.
The reported incidence of com-
mon bile duct stones in patients
ing acquired with increasing frequency in pa-
tients presenting with pain or other nonspe-
cific abdominal complaints. Thus, CT often
undergoing cholecystectomy varies between provides the first opportunity to detect bile
4% and 7% but may be as high as 33% in duct stones in patients undergoing imaging
high-risk groups, such as elderly patients over for various medical conditions. Not surpris-
80 years of age [1]. Common bile duct stones ingly, CT scans with different protocols are
are far more difficult to detect and result in acquired in many patients with common bile
greater morbidity and mortality than stones duct stones. With an almost universal use of
within the gallbladder. Multiple imaging tests MDCT in emergency radiology departments,
have been used to diagnose stones in the com- acquisition of images with high spatial reso-
mon bile duct. In patients with suspected bil- lution is now routine. Thus, it is important to
iary tract disease and jaundice, sonography is recognize the variable appearance of bile duct
typically the initial imaging technique used stones on CT images acquired with MDCT
[2–4]. The reported sensitivity of sonography technology. In this study, we assessed the
for detecting choledocholithiasis varies be- ability of MDCT performed with and without
tween approximately 20% [3] and 80% [4]. an IV contrast agent to detect common bile
Traditionally, ERCP was the gold standard for duct stones.
investigation for bile duct diseases, but its role
is now limited to therapeutic use [5]. Because Materials and Methods
of their improved safety profile, MR cholan- Subjects
giography [6, 7] and endoscopic sonography Approval from our institutional review board
[8] have replaced ERCP for the diagnosis of was obtained for this retrospective study. The study
Keywords: biliary system, CT imaging, gastrointestinal bile duct stones. population included all patients who underwent ab-
radiology
Although CT is not the imaging technique dominal MDCT and ERCP within 3 months during
DOI:10.2214/AJR.05.0459 of choice for patients with clinical suspicion of a 3-year period (May 2001 to June 2004). Mean
choledocholithiasis, it is commonly used in pa- time between CT and ERCP was 7 days (range,
Received March 15, 2005; accepted after revision tients with jaundice, abnormal liver function same day to 15 days). This resulted in a study pop-
July 10, 2005.
test results, or other symptoms possibly related ulation of 72 patients consisting of 27 males and 45
1All authors: Radiology Department, Boston Medical to the biliary tract. Previous studies have re- females with a mean age of 54 years (range, 16–91
Center, 88 East Newton Street, 2nd Floor, Boston, MA ported the performance of CT using different years). Thirty-three patients had studies performed
02215. Address correspondence to S. W. Anderson techniques for showing common bile duct with IV contrast agent only, 15 patients had scans
([Link]@[Link]). stones. These include incremental [9] and heli- performed without IV contrast agent only, and 24
AJR 2006; 187:174–180
cal CT [10–13] acquired both with [11] and patients had studies with and without an IV contrast
without the addition of an IV contrast agent agent. Thus, a total of 96 CT examinations were ac-
0361–803X/06/1871–174
[12, 13]. Reported sensitivities have ranged quired in these 72 patients: 57 with IV contrast ma-
© American Roentgen Ray Society from 20% [14] to as high as 88% [12]. terial (group 1) and 39 without IV contrast material

174 AJR:187, July 2006

 MDCT Diagnosis of Choledocholithiasis

Fig. 1—82-year-old (group 2). Indications for CT in the patient popula-

woman with tion were pancreatitis (n = 26), nonspecific abdom-
periumbilical pain and
weight loss. Axial inal pain (n = 22), trauma (n = 5), jaundice (n = 5),
contrast-enhanced CT suspected pancreatic tumor (n = 2), biliary dilata-
image reveals common tion (n = 3), isolated abnormal liver function test
bile duct to be well
results (n = 2), and others (n = 7). Indications for
visualized without
evidence of ERCP in the patient population were as follows:
choledocholithiasis suspected choledocholithiasis from findings on
(arrows). Even in other imaging techniques including sonography,
retrospect, small stone
retrieved at ERCP could CT, and MR cholangiopancreatography (MRCP)
not be identified. (n = 18); very high clinical suspicion of choledo-
cholithiasis (n = 24); palliation/diagnosis of known
or suspected neoplasm including pancreatic, gall-
bladder, and ampullary tumors (n = 17); trauma
with suspected bile leak (n = 10); and treatment of
pancreatic pseudocyst (n = 3).

CT Technique
All CT studies were performed on a 4-MDCT
(MX 8000, Philips Medical Systems) scanner.
Fig. 2—47-year-old Scans were acquired in a craniocaudal direction
woman
postcholecystectomy with the following parameters: slice thickness, 3.2
with persistent right mm; reconstruction interval, 3 mm; pitch factor, 6;
upper quadrant pain. tube current, 120 kVp and 200–300 mAs. For those
Axial CT image reveals
patients who received IV contrast material en-
small calcified common
bile duct stone (arrow). hancement, 100 mL of iohexol (Optiray, 320
At ERCP, several small mg/mL; Mallinckrodt Imaging) was power injected
stones of varying sizes through a cannula located in an antecubital or hand
were retrieved from
common bile duct. vein at a rate of 3–4 mL/s. The delay between the
beginning of contrast material injection and image
acquisition was 40 seconds for patients in whom
the indication for the CT examination was sus-
pected pancreatic disease and 60 seconds for all
other indications. Eighty-five studies (89%) were
performed with an oral contrast agent (900 mL of
2.2% barium sulfate suspension [Medescan; Lafay-
ette Pharmaceuticals]). Of the 57 scans obtained us-
ing IV contrast material, 50 (88%) included admin-
istration of oral contrast material. Of the 39 scans
without IV contrast material, 35 (90%) included
oral contrast material. The radiologist in charge of
the abdominal CT scan service at the time of the
study made the decision to administer oral contrast
material, IV contrast material, or both.

Image Analysis
The CT scan data sets were transferred to PACS
Fig. 3—83-year-old workstations for analysis (Aurora, software 6.5;
woman with nonspecific
abdominal pain. Merge eFilm). CT images were interpreted inde-
Transverse contrast- pendently by two radiologists, observer 1 and ob-
enhanced CT image server 2, who had 9 and 3 years of experience, re-
reveals large soft-tissue
spectively, in cross-sectional abdominal imaging.
attenuating common bile
duct (CBD) stone (black Observers were free to use the window settings they
arrow). Layering sludge preferred, which included narrow settings if a com-
is also noted in mon duct stone was not initially identified on soft-
gallbladder (white
arrow). At ERCP, 9-mm tissue window settings. The radiologists had the
stone was retrieved from option to generate multiplanar reformations of the
distal CBD. axial images using a postprocessing software pack-

AJR:187, July 2006 175

 Anderson et al.

age incorporated into the PACS workstations Statistical Analysis two observers. Between the two observers, a
(Voxar 3D; Barco). The radiologists received no The independent interpretations by the two ra- total of seven false-negative interpretations
clinical information or results of earlier or subse- diologists were used to determine the following were made in six patients. In one of these pa-
quent imaging studies. However, they were aware diagnostic performance parameters for MDCT tients who had a dual study, both observers
of the aim of the study. images: sensitivity, specificity, positive predic- missed a stone, a solitary 8-mm soft-tissue
Each radiologist was shown the group 1 (unen- tive value (PPV), negative predictive value density stone situated within the distal com-
hanced) CT images separately in a randomized or- (NPV), and diagnostic accuracy. We calculated mon bile duct. Even on retrospective review,
der with all patient details removed for analysis. the 95% confidence intervals (CIs) for each of this stone was not apparent on the unenhanced
After this, the observers were shown the group 2 these parameters. This analysis was done sepa- CT, but both observers detected it prospec-
(IV contrast-enhanced) CT images independently rately for studies performed with IV contrast ma- tively on the IV contrast-enhanced study
with the patient details removed for analysis. A di- terial and for the studies performed without IV (Figs. 4A and 4B). On consensus review, the
agnosis of choledocholithiasis was made when contrast material. Finally, we evaluated the agree- observers considered the stone impossible to
there was direct visualization of a stone as a high- ment between the two observers using the kappa detect because of its isodensity with surround-
attenuation focus or filling defect with surrounding statistics. Agreement was classified as follows: ing soft tissue and the lack of a target or cres-
bile (target sign, crescent sign) seen within the bile 0.00–0.20, poor; 0.21–0.40, fair; 0.41–0.60, mod- cent sign. The remaining five studies were mis-
duct. Common bile duct caliber was not measured, erate; 0.61–0.80, good; 0.81–1.00, very good. interpreted by only one of any of the two
and bile duct features such as dilatation, sharp am- Diagnostic performance parameters were also observers. At CT, the stones were classified as
putation, or both were not used as ancillary signs of calculated for the original interpretations of the soft-tissue attenuation in two, as homoge-
choledocholithiasis. When stones were identified, CT scans at the time of their completion. Sensi- neously calcified in two, and as having partial
an analysis of the CT attenuation characteristics tivity, specificity, PPV and NPV, and diagnostic calcification (calcified rim) in one. On retro-
was made to classify them as diffusely calcified, accuracy were calculated including 95% CIs. spective review, the consensus of the two ob-
partially calcified, or of soft-tissue attenuation. Again, this was done separately for those studies servers was that all these stones were visible,
All ERCP images were also interpreted on the performed with IV contrast material and those but the major causes for not detecting them
PACS workstation by a gastrointestinal radiologist performed without IV contrast material. were related to their small size, similar density
with more than 8 years of experience in cross-sec- to surrounding tissue, confusion of dense
tional abdominal imaging. Studies were classified Results lower bile duct stone for oral contrast material
as positive or negative for the presence of stones in At ERCP, stones were found in 21 (29.2%) in the duodenum, or all three (Fig. 5).
the biliary tract. Stones in the gallbladder or cystic of the 72 patients. The stones ranged in num- Between the two observers, a total of four
duct were not assessed in this study. If stones were ber from one to more than 10, with 16 false-positive interpretations were made in
found, the number of stones (up to a maximum of (76.2%) of the 21 patients having only a sin- three patients using unenhanced CT for de-
10) was recorded, and the size of the largest stone gle stone in their bile duct. The stones ranged tection of choledocholithiasis. Interpreta-
was measured using electronic calipers on the in maximum size from 2.5 to 28 mm, with a tions of both observers were falsely positive
workstation. The size of a stone was determined by mean of 8.8 mm. Using CT evaluation, a in one patient with a dual study. In this same
its size relative to that of the endoscope taken to stone was detected by at least one of the ob- patient, both observers also falsely identified
measure 11 mm. The results from the ERCP were servers in 20 (95.2%) of the 21 patients with a common bile duct stone on the correspond-
used as the gold standard in determining the accu- choledocholithiasis. Both reviewers failed to ing IV contrast-enhanced study. In the re-
racy of CT findings. identify an ERCP-confirmed 3-mm calculus maining two patients, false-positive misin-
Finally, after the results of the ERCP were dis- in one patient who had a dual CT examination terpretations were made by only one of any
closed to the radiologists, they were asked to per- performed (Fig. 1). In the 20 patients with of the two observers. On consensus review,
form a consensus review of the CT data sets to deter- choledocholithiasis identified at CT, the all these misinterpretations were thought to
mine a possible explanation for any false-positive or stones were classified as diffusely calcified in result from volume averaging from sur-
false-negative interpretations made. three (15%) (Fig. 2), as partially calcified in rounding soft tissue such as the pancreas or
The originally dictated reports by the attending nine (45%), and of soft-tissue density in eight common bile duct wall.
radiologists interpreting the CT scans at the time of (40%) (Fig. 3).
completion of the studies were also evaluated. It was Contrast-Enhanced CT (Group 2)
noted, based on the dictated reports, whether com- Unenhanced CT (Group 1) Fifteen (26.3%) of 57 patients in this group
mon bile duct stones were directly visualized or not Thirteen (33%) of 39 patients who had un- were found to have bile duct stones at ERCP.
at that time. A total of eight attending radiologists enhanced CT had choledocholithiasis at Table 2 gives the results for the diagnosis of
with variable experience (2–25 years of experience) ERCP. Table 1 gives the results for the diagno- stones per CT study for the two observers. Be-
were responsible for the original dictations. sis of choledocholithiasis per CT study for the tween the two observers, a total of four false-
negative interpretations were made in three
patients. Of these, both observers made a
false-negative interpretation in one patient.
TABLE 1: Results of Unenhanced CT (Group I) for Both Observers This was a patient with a 4.5-mm stone re-
Observer True-Positive True-Negative False-Positive False-Negative moved at ERCP that was not apparent by CT
1 10 24 2 3 even on retrospective review. The remaining
two false-negative interpretations occurred in
2 9 24 2 4
two different patients, and on retrospective re-

176 AJR:187, July 2006

 MDCT Diagnosis of Choledocholithiasis

A B
Fig. 4—74-year-old woman with abdominal pain.
A, Axial unenhanced CT image from study in which there was no evidence of choledocholithiasis.
B, Axial contrast-enhanced CT image reveals soft-tissue attenuating common bile duct (CBD) stone (black arrow). Also noted are dependently layering gallstones (white
arrow). ERCP revealed 8-mm distal CBD stone.

view, both observers agreed the stones were two stones measuring 3.5 mm and 6 mm were The observers made a total of 12 false-
present: one as a minimally hyperattenuating found in the second patient, respectively. The positive interpretations in 10 patients using
stone and the other as a soft-tissue attenuation reasons for these misses were thought to be IV contrast-enhanced CT for detection of
focus. At ERCP, a single 4-mm common bile related to their small size and isodensity with choledocholithiasis. Both observers made
duct stone was found in the first patient, and surrounding tissue. false-positive interpretations in two patients.
Even at retrospective review, a calcified fo-
cus was thought to be clearly present within
Fig. 5—72-year-old man the common bile duct in both patients by
presenting with diarrhea. both observers (Fig. 6). At ERCP performed
Axial unenhanced CT
image reveals large between 1 and 6 days after the CT, no stones
calcified focus within were found. On consensus review, the only
common bile duct (CBD) explanation offered was that the stones
(arrow). This is false-
negative for one likely passed while awaiting therapeutic
observer that was ERCP. Of the remaining eight patients with
misinterpretation of large false-positive CT studies, four were de-
CBD stone as oral
contrast material within
scribed as being of soft-tissue density; three
duodenal diverticulum. were described as being partially calcified;
At ERCP, single large CBD and, in one patient, as being uniformly calci-
stone was retrieved. fied. At consensus review, these false-posi-
tive findings were thought to have likely re-
sulted from partial volume averaging from
enhancing the wall of the common bile duct,
interference from oral contrast material ad-
ministered, and partial volume averaging
from IV contrast-enhanced blood vessels
coursing close to the bile duct (Fig. 7).

Statistical Analysis
TABLE 2: Results of Contrast-Enhanced CT (Group 2) for Both Observers Table 3 gives the sensitivity, specificity,
Observer True-Positive True-Negative False-Positive False-Negative PPV, NPV, and diagnostic accuracy for the
detection of choledocholithiasis using unen-
1 13 37 5 2
hanced and IV contrast-enhanced CT for
2 13 35 7 2 both observers. Overall, the sensitivity for

AJR:187, July 2006 177

 Anderson et al.

Fig. 6—77-year-old woman with nonspecific abdominal pain. Axial unenhanced CT Fig. 7—74-year-old man presenting with pain and pancreatic mass on sonogram.
image clearly shows calcified focus within common bile duct (arrow). Radiologists Axial contrast-enhanced CT image shows questionable hyperdense focus (arrow)
hypothesized stone may have passed in interim between CT and ERCP, when no within common bile duct (CBD). This is a false-positive for one observer, and at
stone was found. consensus review, hyperdensity was concluded to be partial volume averaging of
enhanced blood vessel in immediate vicinity of CBD.

diagnosis of choledocholithiasis between toms. Thus, it is important for the practicing ra- Heavily calcified stones are relatively easily
the two observers ranged from 69% to 87%, diologist to recognize the variable appearance identified, whereas soft-tissue density stones
specificity from 83% to 92%, and accuracy of bile duct stones on CT and to be aware of the can be isoattenuating to surrounding tissue,
from 84% to 88%. The kappa values for the limitations of diagnosis. The published sensi- making them difficult to identify. The attenu-
interobserver agreement were 0.61 (good) tivities for detection of choledocholithiasis ation of biliary stones varies with their com-
for the unenhanced studies and 0.56 (mod- range from 20% to 90% [2, 3, 10, 12–16], with position. They can be made up of varying
erate) for the contrast-enhanced studies. a mean sensitivity of approximately 80%. amounts of bile pigment, cholesterol, fatty ac-
Table 4 gives the sensitivity, specificity, The ability to detect bile duct stones at CT ids, and calcium. Pure cholesterol stones are
PPV, NPV, and diagnostic accuracy for the depends on a number of factors related to the iso- or slightly hypoattenuating relative to
detection of choledocholithiasis using the stone (size, shape, position, density), bile duct bile, making them difficult, if not impossible,
original interpretations at the time of the CT (dilated vs nondilated), technology used (con- to detect. This imposes a theoretic upper limit
scans. Overall, the sensitivity is much de- ventional vs helical CT), technique used (slice for the CT detectability of choledocholithia-
creased for both IV contrast-enhanced and thickness, reconstruction interval, pitch, kVp, sis of approximately 80% that cannot be im-
unenhanced studies compared with the re- administration of contrast material), patient se- proved on irrespective of any future advances
sults described earlier by the two observers lection (screening population vs all comers), in CT technology [17–19].
during this study. Specificity, however, is and interpreter variability (experience, antici- Even in light of the fact that Jeffrey et al.
somewhat increased, with 100% and 98% pation of bile duct stones). The detection of [16] achieved a sensitivity of 90% in diagnos-
for the unenhanced and IV contrast-en- nearly isoattenuating stones may be improved ing choledocholithiasis using conventional
hanced groups, respectively. by narrow window settings, a technique that (incremental) CT, it is generally accepted that
has been used in previous studies [12]. As helical CT is superior in the diagnosis of cho-
Discussion noted, our radiologists used narrow settings if ledocholithiasis. The potential for helical CT,
Although CT is not the primary technique a common duct stone was not identified on especially using MDCT technology, to image
for diagnosis of choledocholithiasis, bile duct standard soft-tissue window settings. It is well the anatomy using thin slices in a single
stones are often found in patients undergoing recognized that small stones situated within breath-hold and to reconstruct those slices re-
abdominal CT for various indications in rou- the intrahepatic ducts or impacted at the am- trospectively using variable overlap should
tine clinical practice. CT is being increasingly pulla are difficult to identify, particularly in reduce much of the image degradation previ-
used in patients presenting to the emergency nondilated biliary ducts. ously experienced from motion artifacts and
department with abdominal pain, nausea, vom- The attenuation of bile duct stones strongly volume averaging. For our study, we used 4-
iting, or other nonspecific abdominal symp- influences the ability to detect them using CT. MDCT and acquired images with 3.2-mm

178 AJR:187, July 2006

 MDCT Diagnosis of Choledocholithiasis

TABLE 3: Performance Results for the Two Observers in the Analysis of Group 1 formed. As CT technology continues to im-
and Group 2 CT Examinations prove with the increasing use of 16-, 32-, and
Observer 1 Observer 2 64-MDCT, it is likely that the accuracy for de-
Contrast-Enhanced Contrast-Enhanced tection of choledocholithiasis in patients un-
Performance Unenhanced (%) (%) Unenhanced (%) (%) dergoing routine abdominal CT will improve
Measures (95% CI) (95% CI) (95% CI) (95% CI) compared with our results. The increasing res-
Sensitivity 87 (58–98) 87 (46–94) 69 (58–98) 87 (39–90) olution that will be afforded and decreasing ac-
Specificity 92 (74–96) 88 (73–93) 92 (68–92) 83 (73–92) quisition time limiting motion artifacts may
eliminate some of the false-positive and nega-
PPV 83 (46–100) 72 (51–97) 82 (41–84) 65 (48–97)
tive interpretations described in our study.
NPV 89 (81–100) 95 (69–100) 86 (80–99) 95 (66–95) However, for the theoretic limitations just
Accuracy 87 (77–98) 88 (79–96) 85 (73–96) 84 (74–94) mentioned, it cannot be expected to rival the
Note—Group 1 = unenhanced CT, Group 2 = contrast-enhanced CT, CI = confidence interval, PPV = positive excellent results obtained with MR cholan-
predictive value, NPV = negative predictive value. giography or endoscopic sonography.
Because of its retrospective nature, our
study has a number of limitations. First, the
TABLE 4: Performance Results Based trast-enhanced bile duct mucosa. However, mean delay of 7 days for ERCP after CT prob-
on Original Interpretations we also found at least one instance where ad- ably led to passage of some stones, resulting in
at the Time of CT Scan ministration of IV contrast material helped; an increased false-positive rate at CT. Second,
Contrast- this was in a patient with a soft-tissue density the fact that the observers in this study were
Performance Unenhanced (%) Enhanced (%) stone situated at the ampulla was only made primed for detection of choledocholithiasis
Measures (95% CI) (95% CI)
obvious after IV contrast enhancement of the likely led to an improved sensitivity for detec-
Sensitivity 46 (20–74) 33 (13–61) surrounding bile duct wall. In clinical prac- tion for bile duct stones compared with real-
Specificity 100 (84–100) 98 (85–100) tice, both oral and IV contrast material are time interpretations at the time of the examina-
PPV 100 (52–100) 83 (36–99) frequently used because they make the biliary tion. This likely explains the large disparity
NPV 79 (61–90) 80 (66–90) anatomy clearer and they help with diagnoses between sensitivities achieved by the two ob-
of mass lesions such as neoplasms. At least servers in this study compared with that
Accuracy 82 (70–94) 81 (70–91)
from our limited data, it appears that the ad- achieved at the original interpretations at the
Note—CI = confidence interval, PPV = positive ministration of IV contrast material does not time of the CT scans. The fact that the radiolo-
predictive value, NPV = negative predictive value.
limit the accuracy of MDCT for diagnosis of gists involved in this study were primed for the
choledocholithiasis in patients undergoing detection of common bile duct stones likely
routine abdominal imaging. also resulted in overdiagnosing choledoch-
thickness through the entire abdomen and In our study, we achieved a sensitivity of olithiasis, which explains the disparity be-
pelvis in a single breath-hold (averaging 12 69–87%, specificity of 83–92%, and accuracy tween specificity achieved in this study and
seconds) with retrospective image recon- of 84–88% in the CT diagnosis of choledo- that achieved at the time of the original inter-
struction at 3-mm intervals. Our accuracy of cholithiasis. Twenty of the 21 patients with pretation.
84–88% obtained using our technique may stones were identified by at least one of the two In conclusion, unenhanced and contrast-
have suffered compared with that of other in- observers, and stones in only one patient were enhanced CT performed with a 4-MDCT scan-
vestigators who performed CT using thinner missed by both observers. The stones missed ner is moderately accurate in the diagnosis of
slices focused around the biliary tract only. were mostly of soft-tissue or partially calcified choledocholithiasis in patients undergoing im-
We also performed our CT examinations density, measuring less than 6 mm, and situ- aging for various indications. The impact of
using various protocols for oral and IV con- ated in the lower common bile duct. The effect further improvements in CT technology, such
trast administration. Patients such as those of bile duct size on sensitivity was not evalu- as acquisition of images with isotropic voxels,
with right flank pain and suspected renal ated in this study. On retrospective review, the should be investigated.
stones had CT performed without any con- major reasons for missing stones were thought
trast material administration, patients such as to be related to their small size, similar density
those with suspected bowel disease had to surrounding tissue, confusion of dense References
administration of both oral and IV contrast lower bile duct stones for oral contrast material 1. DenBesten L, Doty JE. Pathogenesis and manage-
material, and patients being evaluated after in the duodenum, or all three. False-positive re- ment of choledocholithiasis. Surg Clin North Am
major trauma had IV contrast material admin- sults occurred in our study related to partial 1981; 61:893–907
istration only. It is generally believed that ad- volume averaging from surrounding tissues in- 2. Baron RL, Stanley RJ, Lee JK, et al. A prospective
ministration of oral contrast material, IV con- cluding opacified blood vessels coursing close comparison of the evaluation of biliary obstruction
trast material, or both reduces the accuracy to the bile duct, IV contrast enhancement of using computed tomography and ultrasonography.
for diagnosis of choledocholithiasis. In our bile duct mucosa, and reflux of oral contrast Radiology 1982; 145:91–98
study, we encountered false-positive findings material into the lower common bile duct. It is 3. Mitchell S, Clark R. A comparison of computed to-
for stones resulting from reflux of oral con- also highly likely that a few of our false-posi- mography and sonography in choledocholithiasis.
trast material into the lower bile duct and tive results were related to real bile duct stones AJR 1984; 142:729–733
from partial volume averaging from IV con- that had passed by the time ERCP was per- 4. Wermke W, Schulz HJ. Sonographic diagnosis of

AJR:187, July 2006 179

 Anderson et al.

bile duct calculi: results of a prospective study of Endosc 1998; 47:439–448 ERCP in the diagnosis of choledochal stones. Acta
222 cases of choledocholithiasis [in German]. Ul- 9. Baron RL. Common bile duct stones: reassessment Radiol 1992; 33:53–56
traschall Med 1987; 8:116–120 of criteria for CT diagnosis. Radiology 1987; 15. Sugiyama M, Atomi Y. Endoscopic ultrasonogra-
5. NIH state-of-the-science statement on endoscopic 162:419–424 phy for diagnosing choledocholithiasis: a prospec-
retrograde cholangiopancreatography (ERCP) for 10. Soto JA, Velez SM, Guzman J. Choledocholithiasis: tive comparative study with ultrasonography and
diagnosis and therapy. NIH Consens State Sci State- diagnosis with oral contrast enhanced CT cholan- computed tomography. Gastrointest Endosc 1997;
ment 2002; 19:1–26 giography. AJR 1997; 172:943–948 45:143–146
6. Soto JA, Alvarez O, Munera F, Velez SM, Valencia 11. Miller FH, Hwang CM, Gabriel H, Goodhartz 16. Jeffrey RB, Federle MP, Laing FC, Wall S, Rego J,
J, Ramirez N. Diagnosing bile duct stones: compar- LA, Omar AJ, Parsons WG III. Contrast-en- Moss AA. Computed tomography of choledo-
ison of unenhanced helical CT, oral contrast-en- hanced helical CT of choledocholithiasis. AJR cholithiasis. AJR 1983; 140:1179–1183
hanced CT cholangiography and MR cholangiog- 2003; 181:125–130 17. Baron RL, Rohrmann CA Jr, Lee SP, Shuman WP,
raphy. AJR 2000; 175:1127–1134 12. Neitlich J, Topazian M, Smith RC, Gupta A, Burrell Teefey SA. CT evaluation of gallstones in vitro: cor-
7. Griffin N, Wastle ML, Dunn WK, Ryder SD, MI, Rosenfield AT. Detection of choledocholithia- relation with chemical analysis. AJR 1988;
Beckingham IJ. Magnetic resonance cholangio- sis: comparison of unenhanced helical CT and en- 151:1123–1128
pancreatography versus endoscopic retrograde doscopic retrograde cholangiopancreatography. 18. Brakel K, Lameris JS, Nijs HG, Terpstra OT, Steen
cholangiopancreatography in the diagnosis of Radiology 1997; 203:753–757 G, Blijenberg BC. Predicting gallstone composition
choledocholithiasis. Eur J Gastroenterol Hepatol 13. Jimenez CI, delOlmo LM, Perez HM. Helical CT with CT: in vivo and in vitro analysis. Radiology
2003; 15:809–813 without contrast in choledocholithiasis. Eur Radiol 1990; 174:337–341
8. Canto MI, Chak A, Stellato T, Sivak MV Jr. Endo- 2001; 11:197–201 19. Baron RL. Diagnosing choledocholithiasis: how
scopic ultrasonography versus cholangiography for 14. Pasanen P, Partanen K, Pikkarainen P, Alhava E, far can we push helical CT? Radiology 1997;
the diagnosis of choledocholithiasis. Gastrointest Pirinen A, Janatuinen E. Ultrasonography, CT and 203:601–603

180 AJR:187, July 2006