Sharlene A. Teefey, MD #{149}Richard L. Baron, MD #{149}Scott J. Schulte, MDRandall M. Patten, MD #{149}Margaret H. Molloy, MD

Patterns of Intrahepatic Bile Duct Dilatationat CT: Correlation with ObstructiveDisease Processes’

I From the Department of Radiology, University of Washington School of Medicine, Seattle(SAT., R.L.B., 5.1.5., R.M.P., M.H.M.); and the Department of Radiology, University of PittsburghSchool of Medicine, Pittsburgh (R.L.B.). Received May 20, 1991; revision requested June 21; revisionreceived July 26; accepted August 5. Address reprint requests to SAT., Seattle Veterans AffairsMedical Center, 1660 5 Columbian Way, Seattle, WA 98108.

, � RSNA, 1992Abbreviation: PSC = primary sclerosingchobangitis.

The authors performed a blinded,retrospective analysis of 100 com-puted tomographic (CT) scans of pa-tients with proved extrahepatic bileduct obstruction, including primarysclerosing chobangitis (PSC), to deter-mine whether certain patterns of in-trahepatic bile duct dilatation aresuggestive of specific disease pro-cesses. Among 30 patients with be-nign obstructive disease, CT showedpruning of the intrahepatic ducts infour patients (13%), beading in four(13%), and skip dilatations in one(3%). Among 54 patients with malig-nant obstructive disease, CT illus-trated pruning in eight (15%) pa-tients, beading in 11 (20%), and skip

dilatations in two (4%). Among 16patients with PSC, CT demonstratedpruning in four (25%), beading in two

(13%), and skip dilatations in five(31%). The majority of patients with

malignant or benign obstructive dis-ease or PSC had intrahepatic ductdilatation in both lobes of the liver. Itextended into the periphery in 46 of54 patients (85%) with malignant ob-structive disease, in 20 of 30 (67%)with benign obstructive disease, andin 10 of 16 (63%) with PSC. The CTfinding of skip dibatations is stronglysuggestive of PSC. The CT findingsof pruning and beading are nonspe-cific and may be observed at CT inpatients with bile duct obstructiondue to a wide variety of causes. Thedistribution and extent of intrahe-patic duct dilatation at CT do not dif-fer among bibiary disease processes.

Index terms: Bile ducts, CT, 765.1211 #{149}Bile

ducts, enlarged, 765.36 #{149}Bile ducts, stenosis or

obstruction, 765.288 #{149}Cholangitis, 765.288

Radiology 1992; 182:139-142

P AST reports have described certaincomputed tomographic (CT)

findings as suggestive of primary scle-

nosing cholangitis (PSC) (1,2). CT find-

ings observed in the intrahepatic bile

ducts in these patients included scat-

tered, focal areas of peripheral ductdilatation with no visible connectionto the central hilar ducts (termed skipdilatations) and segments of irregular

duct dilatation producing a beadedappearance (1,2). We have also de-scnibed CT findings of intrahepatic

duct pruning, beading, and skip diba-tations associated with PSC but havestated that the specificity of these CTfindings is unknown (3). We under-

took a blinded, retrospective review

of CT scans from 100 patients withproved bile duct obstruction of vary-ing causes (including PSC) to deter-mine whether certain patterns of in-

trahepatic bile duct dilatation aresuggestive of specific disease pro-

cesses.

MATERIALS AND METHODS

Medical, surgical, and radiobogic recordswere reviewed to identify CT studiesshowing abnormally dilated intrahepatic

ducts (visualization of only the central

night and left hepatic ducts was consid-

ered normal). One hundred CT studies

were selected to include a variety of dis-

ease processes with proved.extrahepaticbiliary obstruction. CT studies were per-

formed on one of three types of scanners:GE 9800 (GE Medical Systems, Milwaukee)(n = 34), GE 8800 (GE Medical Systems)(n = 50), or Picker 1200 (Picker Interna-

tionab, Highland Heights, Ohio) (ii = 16).Contrast material was administered intra-venously with a bolus rapid-infusion tech-nique. As this was a retrospective study,scan techniques varied with the clinical

indications. The collimation used was pre-dominantly 5 or 10 mm (in one case it was3 mm), and scan intervals were either 5 or

10 mm.CT scans of the intrahepatic biliary sys-

tem were evaluated for the following find-ings: (a) pruning, considered presentwhen on a single CT section a dilated in-trahepatic bile duct (excluding the central

right and left ducts) was observed over a4-cm or greater segment without the ex-pected, dilated side branches (Fig 1);(b) beading, considered present when atbeast three closely alternating regions ofdilatation and stenosis were observed inan intrahepatic duct on a single CT section(Fig 2); and (c) skip dilatation, considered

present when CT images showed isolated,dilated peripheral ducts with no visibleconnection to dilated central ducts or withno connection to other dilated ducts oncontiguous CT sections (Fig 3).

The distribution and extent of the intra-

hepatic duct dilatation were also evalu-ated. Dilatation was either limited to thecentral two-thirds of the liver (other than

the central right and left hepatic ducts) orextended into the peripheral one-third. AllCT findings were retrospectively evalu-ated by three of the investigators (R.L.B.,S.J.S., R.M.P.), who were blinded to the

diagnosis and who viewed masked imagesshowing only the liver and not the porta

hepatis and distal common bile duct. Aconsensus of opinion was achieved in all

but one case, in which two of three read-ens considered beading not present and

one considered it present. For purposes ofanalysis, beading was scored as not

present in this case.Medical records were reviewed to deter-

mine the clinical diagnosis at the time ofthe CT examination. Thirteen pathologicconditions causing extrahepatic bile ductobstruction anywhere from the level ofthe porta hepatis to the ampulla were

identified: pancreatic carcinoma (n = 25),

PSC (ii = 16), common bile duct stones(n = 15), chobangiocarcinoma (n = 14),metastatic disease (n = 9), pancreatitis(n = 7), chobedochocyst (n = 3), lym-phoma (n = 3), gallbladder carcinoma

(n = 2), papillary stenosis (n = 2), recur-

a. b.Figure 1. Pruning. (a) CT scan of a patient with PSC shows a left intrahepatic duct stricture(arrowheads) with dilatation proximally and distally. The expected dilated side branches are

absent, which correlates with the cholangiographic finding of pruning. Also noted are spleno-megaly and enlargement of the lateral segment of the left lobe of the liver secondary to cirnho-sis. (Reprinted, with permission, from reference 3.) (b) CT scan of a patient with pancreatic

carcinoma shows no pruning; expected dilated side branches are present (arrowheads).

.�k �

rent pyogenic cholangitis (n = 2), hepato-cellular carcinoma (iz = 1), and ascendingcholangitis (n = 1). The clinical diagnosiswas confirmed by means of tissue speci-men examination (percutaneous on exci-sionab biopsy) in 60 patients and cholan-

giography in 30. In these 90 patients in

whom histologic proof on cholangio-graphic evidence of the diagnosis was ob-

tamed, the time interval between the diag-nosis and the CT study was 1 week or bessin 46 patients, 1-2 weeks in 14, and greaterthan 2 weeks in 30. The clinical diagnosiswas confirmed by the clinical presentationand follow-up findings in four additionalpatients and by examination of a tissuespecimen from a primary neoplasm withclinical and imaging findings compatible

with metastatic disease in six. In these 10patients, clinical and/or imaging conreba-tion was made at the time of the CT study.The diagnosis of PSC was based on find-ings at chobangiography and a clinical

course excluding malignancy in all 16cases.

For purposes of analysis, cases were cat-egonized as follows: (a) benign obstructivedisease (excluding PSC diagnosed bymeans of cholangiography), includingcommon bile duct stones, pancreatitis,chobedochocyst, papillary stenosis, recur-rent pyogenic cholangitis, and ascendingcholangitis; (b) malignant obstructive dis-ease, including pancreatic carcinoma,cholangiocarcinoma, metastatic disease,bymphoma, gallbladder carcinoma, andhepatocellubar carcinoma; and (c) PSC.

RESULTS

Benign Obstructive Disease

Common bile duct stones-Amongthe 15 patients, CT showed pruningof the intrahepatic ducts in one andbeading in one. Skip dibatations werenot identified. Intrahepatic duct dila-tation was present in both lobes in 14patients and in one lobe in one. It wascentral in four patients and extendedinto the periphery in 10. In the casewith duct dilatation limited to onelobe, dilatation was central.

Pancreatitis.-Among the seven pa-tients, CT demonstrated pruning of

the intnahepatic ducts in one. Beadingand skip dibatations were not ob-served. Intrahepatic duct dilatation

was present in both lobes in all seven

patients. It was central in four pa-

tients and extended into the peniph-

ery in three.Others (choledochocyst, papillary steno-

sis, recurrent pyogenic cholangitis, and

ascending cholangitis).-Among the

eight patients, CT illustrated beading

of the intrahepatic ducts in three pa-

tients (recurrent pyogenic cholangitis,

chobedochocyst, ascending cholangi-

tis), pruning in two (recurrent pyo-

genic cholangitis, papillary stenosis),

and skip dilatations in one (recurrent

pyogenic cholangitis). Intrahepatic

duct dilatation was present in both

lobes in all eight patients. It was cen-

trab in one patient and extended into

the periphery in seven.

Summary.-In the 30 patients with

benign obstructive disease, CT

showed pruning of the intrahepatic

ducts in four, beading in four, and

skip dibatations in one. Intrahepatic

duct dilatation was present in both

lobes in 29 patients and in one lobe in

one. It was central in nine patients

and extended into the periphery in

20. In the case with duct dilatation

limited to one lobe, dilatation was

central.

Malignant Obstructive Disease

Pancreatic carcinoma-Among the

25 patients, CT demonstrated pruning

in four patients and beading in seven.

Skip dibatations were not observed.

Intrahepatic duct dilatation was

present in both lobes in 24 patients

and in one lobe in one. It was central

in two patients and extended into the

periphery in 22. In the one case with

duct dilatation limited to one lobe,

dilatation was central.

Cholangiocarcinoma.-Among the 14

patients, CT showed pruning in three

patients and beading in three. Skip

dilatations were not observed. Intra-

hepatic duct dilatation was present in

both lobes in all 14 patients. It was

central in one patient and extended

into the periphery in 13.

Others (metastatic disease, lymphoma,

gallbladder carcinoma, and hepatocellularcarcinoma).-Among the 15 patients,

CT illustrated pruning in one patient

Figure 2. Beading in a patient with PSC. CT

scan shows closely alternating regions of di-

latation and stenosis (arrowheads) in an in-

trahepatic duct. (Reprinted, with permission,

from reference 3.)

(metastatic disease), beading in one

(bymphoma), and skip dibatations in

two (metastatic disease). In these bat-

ten two cases of skip dilatations, he-

patic metastases were present and

visualized on CT scans between the

segments of interrupted ducts. Intra-

hepatic duct dilatation was present in

both lobes in 12 patients and in one

lobe in three. It was central in two

patients and extended into the pe-

riphery in 10. In the three patients

with duct dilatation limited to one

lobe, dilatation was central in two and

peripheral in one.

Summary-In the 54 patients withmalignant obstructive disease, CT

showed pruning of the intrahepatic

a. b.

Figure 3. Skip dilatations in a patient with PSC. (a) CT scan shows isolated, dilated peniph-eral ducts (arrowheads) with no visible connection to dilated central ducts. (b) Contiguous CTsection also does not show a connection to dilated central ducts.

Volume 1�2 #{149}Number 1 Ra�1iolnav #{149}141

ducts in eight patients, beading in 11,

and skip dibatations in two. Intrahe-

patic duct dilatation was present in

both lobes in 50 patients and in one

lobe in four patients. It was central in

five and extended into the periphery

in 45. In the four cases with duct dila-

tation limited to one lobe, dilatation

was central in three and peripheral in

one.

PSC

Among the 16 patients, CT demon-

strated pruning in four patients,

beading in two, and skip dilatations

in five. Intnahepatic duct dilatation

was present in both lobes in 15 pa-

tients and in one lobe in one. It was

central in six patients and extended

into the periphery in nine. In the one

case with duct dilatation limited to

one lobe, dilatation was peripheral.

Summary of CT Findings

Beading was seen at CT in 17 pa-

tients, pruning in 16, and skip dibata-

tions in eight. Of the 17 patients with

beading, seven had pancreatic carci-

noma, three had chobangiocancinoma,

two had PSC, and one each had lym-

phoma, recurrent pyogenic cholangi-

tis, common bile duct stones, chole-

dochocyst, on ascending cholangitis.

Of the 16 patients with pruning, four

had pancreatic carcinoma, four had

PSC, three had cholangiocarcinoma,

and one each had metastatic disease,

recurrent pyogenic cholangitis, papib-

bary stenosis, pancreatitis, or common

bile duct stones. Of the eight patients

with skip dibatations, five had PSC,

two had liver metastases (separating

the proximal and distal duct dilata-

tion) clearly demonstrable at CT, and

the remaining patient had recurrent

pyogenic cholangitis.

Intrahepatic duct dilatation was

present in both lobes in 94 patients

and in one lobe in six. Of the 94 pa-

tients with intnahepatic duct dibata-

tion in both lobes, 50 had malignant

obstructive disease, 29 had benign

obstructive disease, and 15 had PSC.

Of the six patients with intrahepatic

duct dilatation limited to one lobe,

four had malignant obstructive dis-

ease, one had benign obstructive dis-

ease, and one had PSC.

There were six patients with intra-

hepatic duct dilatation limited to one

lobe. Four patients had malignant ob-

structive disease, one had benign ob-

structive disease, and one had PSC.

Intrahepatic duct dilatation was

central in 24 patients and extended

into the periphery in 76. Of the 24

patients with central dilatation, 10

had benign obstructive disease, six

had PSC, and eight had malignant

obstructive disease. Of the 76 patients

with extension of duct dilatation into

the periphery, 46 had malignant ob-

stmuctive disease, 20 had benign ob-

structive disease, and 10 had PSC.

DISCUSSION

CT has become an accepted method

for evaluating intra- and extrahepatic

biliary obstruction. Several studieshave reported that CT is accurate in

determining the presence and distni-

bution of bile duct dilatation as well

as the bevel and cause of the obstruct-

ing lesion (4-10). Recently, authors

have reported an association between

PSC and the CT findings of beading,

pruning, and skip dibatations in the

intrahepatic ducts (1-3). However,

these studies analyzed only patients

with known PSC, and to date no

blinded or controlled study has been

performed-to our knowledge-to

determine whether these CT findings

are specific for PSC.

Our blinded, retrospective review

shows that pruning is not specific for

PSC. Pruning was present in only

25% of cases of PSC; it was also

present in 13% of cases of benign and

15% of cases of malignant obstructive

disease. Pruning as seen at cholan-

giography is the visualization of a di-

bated intrahepatic bile duct without

the expected smaller side branches. In

PSC, these side branches are not seen

at cholangiogmaphy due to narrowing

and sclerosis, which either obstruct or

restrict the flow of contrast material.

While CT can demonstrate pruned

intrahepatic ducts, it does not allow

one to determine whether the nonvi-

sualized side branches are sclemosed

due to PSC or are of normal caliber

and not seen as confluent structures

(Fig 4). According to Laplace’s law,

the widest portion of the biliary tree

will dilate first. We hypothesize that

in cases of early or mild intrahepatic

duct dilatation due to a central ob-

structing process of any cause, pmun-

ing of intrahepatic ducts may be ob-

served at CT, as the smaller-caliber

side branches will not be dilated and

thus not visualized. This hypothesis

correlates with our findings, as prun-

ing was observed on CT scans of pa-

tients with extrahepatic bile duct ob-

struction of varying benign and

malignant causes, including pancre-

atic carcinoma, pancreatitis, common

bile duct stones, and cholangiocarci-

noma.

Beading is also nonspecific for PSC

and was present in only 13% of cases

of PSC. It was also present in 20% of

cases of malignant and 13% of cases

of benign obstructive disease. Beading

as seen at chobangiography is due to

closely alternating regions of dilata-

tion and stenosis in an intrahepatic

duct. This cholangiographic pattern is

unique to PSC. The pattern may be

observed at CT but may also be simu-

bated by dilated, tortuous intrahepatic

ducts (Fig 5). Beading was observed

most frequently on CT scans of pa-

tients with malignant obstructing be-

sions; this finding may be due to the

greater degree of intrahepatic duct

dilatation in these cases. Although the

CT finding of gross intrahepatic duct

dilatation is not specific for malignant

disease, it has been reported to be

more frequent in this patient group

(4) and would be expected to occur

Figure 5. Apparent intrahepatic bibiarybeading in a patient with metastatic disease.

CT scan illustrates a dilated intrahepatic ductthat has a beaded appearance (arrowheads)

in the medial segment of the left lobe of the

liver. At cholangiography, there was no evi-

dence of PSC.

142 #{149}Radiology January 1992

less often in cases of PSC owing to the

scberosing nature of that disease pro-

cess.

Skip dilatations are highly sugges-

tive of PSC and were present in 31%

of such cases, compared with only 3%

and 4% of cases of benign on mabig-

nant obstructive disease, respectively.

In two patients with liven metastases,

skip dilatations at CT could have been

simulated by the strategic location of

the metastases, which obstructed the

adjacent peripheral ducts, but the

clear visualization of the masses ac-

counted for the loss of duct visualiza-

tion and thus the metastases were not

confused with PSC. Recurrent pyo-

genic chobangitis can also produce

skip dilatations at CT, as the diffuse,

scattered bibiary strictures seen in this

disease process are similar to those in

PSC. CT features characteristic of me-

current pyogenic cholangitis such as

marked intra- and extrahepatic bile

duct dilatation, sludge, and intra- and

extrahepatic duct calculi (11) can aid

in its differentiation from PSC.Skip dilatations may be incorrectly

interpreted at CT due to technical fac-

tons. Subtle variations in patient

breathing at the time of CT may result

in noncontiguous sections and simu-

bate skip dilatations. Skip dibatations

as seen at cholangiogmaphy are due to

segments of intrahepatic duct nan-

rowing with proximal and distal duct

dilatation. Whereas chobangiography

allows contiguous evaluation of the

bibiany tree, CT demonstration of skip

dibatations requires mentally recon-

structing the course of a dilated intra-hepatic duct from contiguous CT sec-

tions. Thus, in diagnosing skip

dibatations there is greaten potential

for interpretive and technical errors

with CT than with cholangiognaphy.

Technical factors may also affect visu-

alization of beading and pruning.

Collimation differences may allow for

volume averaging and affect detec-

tion of the fine detail occasionally me-

quined to detect these abnormalities.

While PSC is focal and sporadic, it

typically diffusely involves the intra-

hepatic bile ducts (12,13), and there-

fore in our study the CT finding of

intrahepatic duct dilatation in both

lobes of the liver in all but one case

was not unexpected. However, a cen-

tral obstructing lesion of any cause

may produce a similar picture. In the

majority of our non-PSC cases, CT

showed diffuse (but not focal) duct

Figure 4. Apparent intrahepatic bibiany ductpruning in a patient with pancreatic carci-

noma. CT scan shows a dilated left intrahe-patic duct (arrowheads) without the ex-pected dilated side branches.

dilatation in both lobes of the liver.

Thus, the distribution of intrahepatic

duct dilatation is not helpful in distin-

guishing PSC from other causes of

bile duct obstruction. We observed six

cases of intrahepatic duct dilatationlimited to one lobe; in four of these

cases (with obstructions of varying

causes), dilatation was limited to the

right lobe, which may have been due

to a small left lobe and difficulty in

detecting mild duct dilatation.No one pattern (central vs peniph-

eral) of bile duct dilatation predomi-

nated on the CT scans of patients

with PSC. Of interest is the CT find-

ing of peripheral bile duct dilatation

in 85% of patients with malignant

obstructive disease and 67% with be-

nign obstructive disease. While gross

intrahepatic duct dilatation is not spe-

cific for malignant disease, it has been

reported to occur more frequently

with malignant lesions than with be-

nign lesions (4).

In conclusion, the CT finding of

skip dibatations is strongly suggestive

of PSC in the absence of a mass lesion

in the liver or recurrent pyogenic

chobangitis. The CT findings of prun-ing and beading are nonspecific and

may be observed on CT scans of pa-

tients with bile duct obstructions of a

wide variety of causes. The distribu-

tion and extent of mtrahepatic duct

dilatation on CT are also nonspecific

forPSC. U

References1. Rahn NH, Koehber RE, Weyman PJ, Truss

CD, Sagel 55, Stanley RJ. CT appearanceof sclerosing chobangitis. AIR 1983; 141:549-552.

2. Ament AE, Haaga JR. Wiedenmann SD,BarkmeirJD, Morrison SC. Primary scbe-nosing chobangitis: CT findings. J Comput

Assist Tomogr 1983; 7:795-800.3. Teefey SA, Baron RL, Rohrmann CA, Shu-

man WP, Freeny PC. Sclerosing cholangi-tis: CT findings. Radiology 1988; 169:635-639.

4. Baron RL, Stanley RJ, Lee JKT, Koehler RE,Levitt RG. Computed tomographic fea-tunes of biliary obstruction. AJR 1983; 140:

1173-1178.

5. Baron RL, Stanley RJK, Lee JKT, et al. A

prospective comparison of the evaluation

of bibiary obstruction using CT and ultra-sonography. Radiology 1982; 145:91-98.

6. Goldberg HI, Filly RA, Korobkin M, MossAA, Kressel HY, Callen PW. Capability of

CT body scanning and ultrasonognaphy todemonstrate the status of the biliary ductalsystem in patients with jaundice. Radiol-

ogy 1978; 129:731-737.7. Havribla TR, Haaga JR, Alfidi RJ, Reich NE.

Computed tomography and obstructivebiliary disease. AJR 1977; 128:765-768.

8. Levitt RG, Sagel 55, Stanley RJ, Jost RG.Accuracy of computed tomography of theliver and biliary tract. Radiology 1977; 124:123-128.

9. Pedrosa CS, Casanova R, Rodrigues R.

Computed tomography in obstructivejaundice. I. The level of obstruction. Radi-obogy 1981; 139:627-634.

10. Pedrosa CS, Casanova R, Lezana AH,

Fernandez MC. Computed tomographyin obstructive jaundice. II. The cause of ob-struction. Radiology 1981; 139:635-645.

11. Chan FL, Man SW, Leong LLY, Fan ST.Evaluation of recurrent pyogenic cholangi-tis with CT: analysis of 50 patients. Radiol-ogy 1989; 170:165-169.

12. LaRusso NF, Wiesner RH, Ludwig J, Mac-Carty RL. Current concepts: primary scle-nosing cholangitis. N Engl J Med 1984; 310:899-903.

13. MacCarty RL, LaRusso NF, Wiesner RH,Ludwig J. Primary sclerosing cholangitis:

findings on cholangiography and pancre-atography. Radiology 1983; 149:39-44.