Bernd

K Wailner,

MD

Karl

A. Schumacher,

#{149}

Dilated Billary Cholanglography Contrast-enhanced

terms:

Bile ducts,

Bile ducts, neoplasms, stenosis or obstruction,

Magnetic

resonance

Radiology

1991;

Werner

Weidenmaier,

#{149}

MD

Tract: Evaluation with a T2-welghted Fast Sequence’

A heavily T2-weighted gradient-echo sequence was used for magnetic resonance (MR) imaging of the biliary system in five healthy volunteers and 13 patients with obstructive jaundice. Images were obtained in the sagittal and coronal planes during sequential breath-hold intervals and were postprocessed by using a maximum-intensity projection algorithm. The extrahepatic and intrahepatic bile ducts were well visualized in 11 patients. The level of obstruction and the grade of dilatation were depicted with MR cholangiography. However, the cause of obstruction could be determined with MR cholangiography in only eight cases. The part of the biliary system below the obstruction could not be visualized with MR cholangiography. In the volunteers, MR cholangiography could demonstrate the anatomy of the biliary tract in only two subjects. Possible causes for this phenomenon are the limited spatial resolution of MR imaging, partial volume effects, or flow within the bile ducts. MR cholangiography may be a useful adjunctive tool for noninvasive evaluation of patients with obstructive jaundice. However, further technical advances are necessary to improve image quality. Index

MD

MR studies,

76.1214 76.32, 76.33 #{149}Bile ducts, 76.289 #{149} Jaundice (MR), rapid imaging

181:805-808

T

with

role of noninvasive techniques like ultrasound (US) and cornputed tomography (CT) for evaluation of the biliary tract has been previously described (1-4). US and CT are currently the modalities of choice for diagnosis of dilatation of intrahepatic and extrahepatic bile ducts. However, CT is limited to the axial plane, and US

HE

has

a limited

field

of view

Jean

and

is

operator-dependent. Frequently, invasive techniques such as pencutaneous transhepatic chobangiography (PTC) on endoscopic retrograde cholangiopancreatography (ERCP) are necessary to characterize the cause of obstruction. These techniques provide a good overview of the whole biiany system, and images can be obtained in different planes. Magnetic resonance (MR) imaging has come into general use for evabuation of patients with suspected liver disease. Its ability to demonstrate dilatation of the intrahepatic and extrahepatic biiary tract has previously been shown (5). In that study, however, the images were obtained with spin-echo sequences-most in the axial planeresulting in images comparable with those obtained with CT. We describe the application of a novel approach with use of a rapid sequential gradient-echo acquisition and a three-dimensional postpnocessing technique to produce direct coronal and sagittal images. This method, which we call “MR chobangiography,” was applied to healthy volunteers and patients with obstructive jaundice for noninvasive imaging of the biliary system.

AND

steady-state

which ordinary

From

the Department

of Radiology,

Univer-

sity of Ulm, Ulm, Germany. Received March 28, 1991; revision requested May 15; revision re-

ceived

June 27; accepted

requests University

to B.K.W.,

Hospital, Ulm, Germany. C

RSNA,

1991

July 1. Address

Department

reprint

of Radiology

Robert-Koch-Str,

D-7900

I,

magnetization,

T2-weighted

steady-state

quence

In an

(6,7).

gradient-echo

like fast imaging

with

se-

steady

pre-

cession

(FISP) or gradient recalled acquisition in the steady-state (GRASS; GE Medical Systems, Milwaukee), the two echoes are coincident, so that the signal is

a complex

function

of tissue

relaxation

times, proton density, and sequence parameters. With contrast-enhanced fast (CE-FAST) imaging, however, only the second echo is measured. For these Sequences, the effective echo time (TE) is longer than the TR and is equal to 2TR TE; this is achieved by refocusing the signal from one excitation during a subsequent excitation, resulting in a long effective TE. The long effective TE permits T2 contrast effects to develop (unlike FISP

and GRASS, which have TE). The signal is roughly exp (-2TR/T2) (8). All examinations were 1.5-T

whole-body

a much shorter proportional to performed

imaging

tom 63SP; Siemens, A circularly polarized

with

system

Erlangen,

a

(Magne-

Germany).

body coil was used as transmitter and receiver. For MR imaging of the biliary system, a CE-FAST se-

quence with first-order in the readout direction

motion rephasing was employed.

Imaging

parameters

msec/TE aquisitions. 256 matrix

msec), 70#{176} flip angle, and three A single section with a 192 x was obtained during a breath-

hold

period

was

imaged

and

were

of 12 seconds.

17/7

(TR

The whole

sagittal

5 mm, and the field of view ranged 32 x 32 to 45 x 45 cm, depending size of the patient. The images

processed projection

liver

with sequential contiguous acquisitions in the coronal planes. Section thickness was

breath-hold

by using algorithm

cholangiogram user-defined

rameters

Technical

transverse

is heavily

TI-

were

and

from on

were

the

post-

a maximum-intensity to create a projection

that angles.

with spin-echo

METHODS

MD

MR

derwent

SUBJECTS

M. Friedrich,

#{149}

from un-

could be viewed All subjects also

T2-weighted

sequences. 600/15 with

imaging

Imaging pafour excitations

Considerations

Gradient-echo nal with first, the

nab sity-

sequences produce sigtwo different types of contrast: echo produced by the longitudi-

magnetization,

which

or Ti-weighted,

ffip angle and repetition ond, the echo produced

is proton

depending time

by the

denon the

(TR); secpersistent

Abbreviations: CE-FAST fast, ERCP = endoscopic pancreatography, FISP = steady precession, VFC = hepatic cholangiography,

repetition

time.

= contrast-enhanced retrograde cholangio-

fast imaging percutaneous TE = echo

with trans-

time, TR

=

Level of Obstruction,

Diameter of the Common Bile Duct, Cholangiography in 13 Patients

and Cause

o f Obstruction: Maximum

:

Asses sment Diameter

of CBD (mm) Was

Patient

Cause

of Obstruction

Malignancy of the pancreatic Malignancy of the pancreatic Malignancy of the pancreatic Cholangiocarcinoma

5

Concrements

6

Chronic

7

Sclerosing

head head head

pancreatitis cholangitis

8 9 10 11

Malignancy of the pancreatic Sclerosing cholangitis Hepatocellular carcinoma Cholangiocarcinoma

12

Carcinoma

13

Cystic

of the

papilla

head

of Vater

adenocarcinoma

of the pancreas

and 2,500/20, 60, 120 with two excitations for Ti- and T2-weighted imaging, respectively.

Subjects Five

healthy

volunteers

obstruction

hepatic

bile

and

13 patients

of extrahepatic

ducts

underwent

or intraMR

examiof the

nation. Four patients had tumors pancreas with obstruction of the common bile duct, two had cholangiocarcinomas (Klatskin struction bile

tumors) of the

ducts,

two

and consecutive right and left main had

undergone

chemotherapy of liver sclerosing cholangitis, in the

lower

biliary

pancreatitis,

one

regional

metastases and had one had gallstones

tract, had

obhepatic

one

a cystic

had

chronic

adenocarci-

noma of the pancreas that infiltrated the left lobe of the liver, one had a carcinoma of the papilla of Vater, and one had a hepatoma. All patients were evaluated with contrast-enhanced CT and US and were selected for the study if dilated bile ducts were seen with these imaging modalities. Five patients with tumors underwent FTC; three underwent consecutive percutaneous transhepatic drainage, and two underwent structing

surgery. gallstones

The patient underwent

with ERCP

oband

extracorporeal shock wave lithotripsy. The other eight patients were evaluated with CT, US, and ERCP. CT was performed as a dynamic, contrast-enhanced study with injection

of a bolus

of iodinated

contrast

medium (Iopamidol; BYK-Gulden, Monstanz, Germany) at 2 mL/sec with a section thickness of 5 mm. The volunteers were evaluated with US for comparison with

MR

Image

cholangiography.

Evaluation

An experienced

radiologist performed the US examination in all subjects and evaluated the CT scans and conventional

#{149} Rdinlnv

CBD, CBD, CBD, Main

MR Cholangiography

CT

pancreatic head pancreatic head below the bifurcation hepatic bile duct

Level

Was

of Obstruction Seen with MR Cholangiography?

Cause

11 22 23 4

12 20 25 4

Yes Yes Yes Yes

Yes Yes Yes Yes

CBD

16

19

Yes

Yes

CBD, pancreatic head Bifurcation of the CBD Proximal CBD Bifurcation of the CBD Left main hepatic duct Proximal CBD

12 5 6 4 5 4

12 6 7 4 NA* NA*

Yes Yes Yes Yes No* No*

No Yes No

17

15

Yes

Yes

21

23

Yes

No

Distal

CBD

CBD, pancreatic

head

of

Obstruction Determined with MR Cholangiography?

Yes

No* No*

= common bile duct, NA = not available. quality of the MR cholangiograms was poor.

*Diagnostic

with

Level of Obstruction at CT, US

at CT, US

1 2 3 4

Note.-CBD

got;

,

US, CT and MR

with

radiographs

ERCP. tion,

obtained

during

He assessed the

cause

the

FTC

level

of obstruction,

or absence tion of the

of involvement common bile

maximum

diameter

the

common

obstruction the diameter hepatic bile

measured.

The

evaluated

independently

MR

presence

of the bifurcaduct, and the

of the

duct. In cases with of the bifurcation, and right common

or

of obstruc-

bile

at the level of the left ducts was

cholangiograms

were

by another

radi-

ologist who did not know the findings of the other imaging modalities. The maximum diameter of the bile ducts was measured sections cessing. third

in the coronal before they The results radiologist,

plane on the single underwent postprowere evaluated by

who

eter of the bile ducts MR cholangiography,

compared

assessed as well

and cause of obstruction either imaging modality.

the

depicted with MR cholangiography (Fig 1). In one patient, the common bile duct and the right hepatic bibiary tree were well visualized, whereas only the central part of the left cornmon hepatic bile duct was seen. In this case, there were motion artifacts

on the MR images a

diam-

with CT and as the level

determined

grams; the intrahepatic part of the biliary tree could not be visualized. In the ii other patients, there was good contrast between the bile ducts and surrounding liver tissue, and the anatomy of the bile tract was readily

with

RESULTS The tween patients

results and the comparison beinformation obtained in the with MR cholangiognaphy and the other imaging modalities are summarized in the Table. MR images of diagnostic quality were obtained in two volunteers and 11 patients. MR imaging of the bile ducts in the coronab plane provided a good view of the biliary system, and no additional information was found by imaging in the sagittal plane. One of the two patients with lowquality images was clinically decompensated and unable to cooperate for the study. The other patient had pneumobilia after having previously undergone surgery. In this patient, only the distal portion of the common bile duct was seen on MR cholangio-

left hepatic tions.

lobe

due

in the region of the to heart pulsa-

The location of biliary could be demonstrated with

MR

bated

well

obstruction

in all patients

cholangiognaphy

proximal seen on

to the obstruction MR chobangiograms,

and comebiliamy tract was well but the

portion

of the

bile

with

CT. The

common

duct

dis-

tab to the obstruction was not regubarby visualized. Thus, in five cases, MR cholangiography could not help in differentiating between severe stenosis and complete occlusion. In the two patients with Klatskin tumors, both the left and right bibiary tract

could

be visualized

with

MR cholan-

giography; in one patient, only the bile ducts of the might lobe were seen after puncture of a right hepatic bile duct during PTC. The other patient had two biliary stents, and the left

and seen. dorsal

right

hepatic

However, segment

bibiary

not drained and could ized with conventional phy.

The

dilated

with

both

MR

trees

were

the bile duct of the of the right lobe was

duct

not

be visual-

chobangiogracould be seen

chobangiography

flrnihr

and

1 qqi

duced with spin-echo sequences. Branches of the portal vein that lie adjacent to the bile ducts appear dark on images obtained with the CE-FAST sequence. Image quality was not degraded by flow artifacts from overlying vessels

in any

of our

cases.

Postprocessing of the individual sections with the maximum intensity projection algorithm simplified the evabuation of MR chobangiography. Although

the dilated

bile ducts

obstruction

Figure 1. Carcinoma of the papilla of Vater and obstruction of the common bile duct. (a) Conventional cholangiogram obtained during percutaneous transhepatic drainage after the tumor was passed with a catheter. The obstruction is demonstrated (arrow). (b) MR cholangiogram well as the

dilated

obtained bile

with ducts.

maximum-intensity

projection

shows

2). In one of the patients with sclenosing cholangitis, the three-dimensional MR cholangiogram revealed a discontinuity between the left and right common hepatic duct. This, however, was not seen with CT, but US demonstrated a high-grade stenosis of the common left and right

It was and

not

PTC

possible

in this

pa-

tient.

In four underwent the biliany

use US.

of the eight patients who ERCP, the visualization

of

system proximal to the tuwas not possible because of cornobstruction of the common bile The intrahepatic bile ducts in patients were visualized with

mon plete duct. these

of MR

cholangiography,

CT,

and

In the patient with concrements of the common bile duct, MR cholangiography showed a complete obstruction. When contrast medium was injected with pressure at ERCP, howeven,

it was

contrast tion.

In the

demonstrated

medium

that

passed

volunteers,

the

the

the obstruc-

common

bile

duct and the left and right hepatic ducts were visualized in only two cases. In these subjects, the diameter of the left and right common hepatic bile ducts was determined to be less than 5 mm with US. Peripheral parts of the intrahepatic biliary tree could

not

be seen

with

either

cholangiography. The was seen in all volunteers imaging modalities.

obstruction

(arrow)

as

images

in

necessary

CT (Fig

hepatic bile ducts. to perform ERCP

the

US or MR gallbladder with

both

We have demonstrated the application of a rapid acquisition gradient-echo sequence to study the anatomy of the

struction

biliary

intensity.

in

patients

with

biliary

ob-

struction. This method combines the use of a heavily T2-weighted CE-FAST sequence with sequential two-dimensional acquisition and three-dimensional display. The individual sections are acquired during sequential breathhold periods, and the whole study can be performed in only 10-15 minutes. Thus, the technique can be included in routine imaging protocols of the liver for additional evaluation of the biliary system. MR cholangiography accurately demonstrated the anatomy of the dilated intrahepatic and extrahepatic bile ducts and the location of obstruction. In patients with obstructive jaundice, bile fluid has a long T2 compared with that of the surrounding liver tissue (5). With the CE-FAST sequence, this effect is emphasized due to the heavy T2-weighting, and there is good contrast between dilated bile ducts and liver tissue. Unlike other gradient-echo sequences, with

CE-FAST

imaging,

the level on

the

of

individ-

the

sagittab

for

evaluation

plane

were of

not

the

biliary

system. MR chobangiography demonstrated the level of obstruction in all but two patients, but the cause of obstruction could not be regularly determined. The dilated bile ducts proximal to the ob-

DISCUSSION

tree

and

visible

ual sections, they were difficult to evaluate because each section showed only a small part of the bile ducts. The postprocessed images provided a good view of the whole biliary tract on one image, and the bevel of obstruction, as well as the different parts of the biliary system, could be easily identified. The intrahepatic and extrahepatic bile ducts were well visualized in the coronal plane; the

b.

a.

were

the

protons

(which are excited by one radio-frequency pulse) are refocused by the subsequent radio-frequency pulse and form an echo before the third radio-frequency pulse. Flowing protons in blood vessels therefore have low signal intensity, since they do not stay in the image plane long enough to experience both radio-frequency pulses. This produces flow void comparable with that pro-

the

had

high

obstructed

distal

to

signal

area

the

intensity,

and

the

obstruction

The

had

obstruction

ducts

low

signal

itself

poorly tiveby

outlined, mainly bow signal-to-noise

FAST

images.

In

and

bile

this

was

due to the rebaratio on CEpreliminary

study,

the

thickness of the individual sections was 5 mm. This was determined from prior experience with MR angiography of abdominal

vessels

giognaphy,

thin

to optimize especially vessels

(9,11).

sections

flow-related for visualization

with

slow

In are

MR

an-

necessary

enhancement, of smaller

flowing

blood;

how-

ever, the high signal intensity of dilated bile ducts on images obtained with the CE-FAST sequence is independent from time-of-flight effects and solely due to T2-weighted contrast. It therefore may be useful to increase section thickness to obtain better signal-to-noise ratios and better delineation of the obstruction. MR cholangiography was not abbe to help differentiate between high-grade stenosis

and

complete

occlusion.

Due

to

the limited spatial resolution, the narrowed part of the bile duct could not be visualized with MR chobangiography. Although the diameter of the poststenotic was

section usually

fluid

in this

sity

on

area

same,

sponsibbe

this

poststenotic

section,

good

of bile

runoff

Since

in the of

effects

the

bow signal

fluid

areas

bile

5 mm,

images.

flow for

common

had

of bile

poststenotic

the than

CE-FAST

position

the

of more

the

are

pre-

biliary

theme

fluid,

intenthe tract

is

meIn

is usually

whereas

corn-

and

probably

phenomenon.

duct bile

the a

in the

a.

b.

Figure

2.

the left and not drained

and right

Klatskin right

tumor hepatic

by the stent. biliary system,

and biliary

(c) MR cholangiogram as well as the dilated

prestenotic area, only minimal flow effects are present. As mentioned earlier, flow void effects may explain the different contrast behavior of bile fluid in the pre- and poststenotic sections of the biliary

system.

This

effect

may

be

dimin-

ished by increasing the thickness of the individual sections. With thicker sections, the protons stay in the image plane for a longer time, and flow void effects are reduced. MR chobangiography demonstrated the anatomy of the biliary tract in only two of the five volunteers. In these two subjects, only the central part of the biliany system (common left and right intrahepatic bile ducts) could be visualized; the more peripheral areas were not seen. At US, the common intrahepatic bile ducts measured less than 5 mm in diameter in these subjects, and the smaller peripheral bile ducts could not be seen; the cause for the poor visualization of the peripheral biliary tree with MR cholangiography is the limited spatial resolution and the partial volume averaging of the bile ducts with surrounding

tissue,

which

leads

to

poor

contrast (5). The lack of visualization of the larger bile ducts in healthy volunteens is probably due to flow effects, as described earlier for the poststenotic biliamy tracts in patients with obstructive jaundice. In addition, there is a wide range of 12 of bile fluid in healthy subjects,

since

the

composition

changes, depending other physiologic influence

fluid

the

of bile

on nutritional factors. This also

visualization

on CE-FAST

images.

c.

left and right external tree. The stents are

of

the

fluid

and may bile

biliary patent.

obtained right dorsal

drainage (flush stents). (a) Conventional cholangiogram shows mild (b) CT scan shows that the right dorsal segmental bile duct is dilated in the coronal plane provides a good overview and shows mild dilatation

hepatic

bile duct

(arrows).

MR cholangiography formed in patients with

cannot metallic

be perbiliary

Radiology

#{149}

1.

should

be performed

before

positive

finding

biliary

these bile

tract

of an occlusion

because

substances

of the

compared

4.

T2 of

with

that

of

fluid.

Motion

is another

problem

that

tios

due

reduced

to heart

the region

was

demonstrate

the volunteers,

the

some

not

Since able

biliary

patients

6.

ra-

especially

of the left lobe.

cholangiography

tently

signal-to-noise

pulsation,

in

MR

can be demonstrated

cholangiography termined.

MR cholangiography

has

with not

yet

7.

to consis-

tract

in

may

have

8.

slight dilatation of the biliary system that cannot be seen on MR cholangiograms. The minimal grade of dilatation

that

5.

may

diminish image quality in MR imaging of the biliary system. There may be sig-

nal loss and

2.

of the

shorter

9.

MR been

de10.

has the capabil-

ity for noninvasive imaging of the biliary tree in patients with obstructive jaundice. The major limitations at that stage are the limited spatial resolution and relatively low signal-to-noise ratio, so that the cause of obstruction cannot regularly be determined. Further improvements of the technique are

Weill F, Eisenacher

A, Zeltner

F.

Ultra-

sonic

3.

any other procedure. Another possible pitfall of MR cholangiography is the presence of blood clot or sludge in the biliary tract. In these cases, MR cholangiography might demonstrate a false-

needed before the method can be implemented as a routine clinical test. 808

References

stents because of artifacts and the possible induction of currents in the implant. The presence of air in the bile ducts also makes MR cholangiography of the biliamy system impossible. MR examination,

therefore,

dilatation of (arrows) and is of the left

11.

study of the normal and dilated biliary tree. Radiology 1978; 127:221-224. Cooperberg PL. High-resolution real-time ultrasound in the evaluation of the normal and obstructed biliary tract. Radiology 1978; 129:477-480. Taylor KJW, Rosenfield AT, Spiro HM. Diagnostic accuracy of gray-scale ultrasonography for the jaundiced patient: a report of 275 cases. Arch Intern Med 1979; 139:60-63. Goldberg HI, Filly RA, Korobkin M, et al. Capability of CT body scanning and ultrasonography to demonstrate the status of the biliary ductal system in patients with jaundice. Radiology 1978; 129:731-737. Dooms GC, Fisher MR, Higgins CB, et al. MR imaging of the dilated biliary tract. Radiology 1986; 158:337-341. Buxton RB, Edelman RR, Rosen BR, et at. Contrast in rapid MR imaging: TI- and T2weighted imaging. J Comput Assist Tomogr 1987; 11:7-16. Winkler MI, Ortendahl DA, Mills TC, et at. Characteristics of partial flip angle and gradient reversal MR imaging. Radiology 1988; 166:17-26. Martin JF, Edelman RR. Fast imaging. In: Edelman RR, HesselinkJR, eds. Clinical magnetic resonance imaging. Philadelphia: Saunders, 1990; 183-220. Kim D, Edelman RR, Kent C, Porter DH, Skiliman JJ. Abdominal aorta and renal artery stenosis: evaluation with MR angiography. Radiology 1990; 174:727-731. Edelman RR, Zhao B, Liu C, et al. Magnetic resonance angiography and dynamic flow evaluation of the portal venous system. AJR 1990; 154:937-946. Wallner B, Edelman RR, Kim D, Finn JP. Assessment of thoracic and abdominal aortic aneurysms using MR angiography. Fortschr ROntgenstr 1991; 154: 1 1--16.

U

December

1991