What’s new?

C The availability of three-dimensional imaging allowing recon-

structions in any plane

C Newer, more liver-specific magnetic resonance imaging (MRI)

contrast agents

INVESTIGATIONS

Imaging the liver andbiliary tractMo Malaki

Kamarjit Mangat

C Diffusion-weighted MRI and positron emission tomography-

computed tomography for further characterization of focal

lesions

AbstractSeveral imaging modalities are used to investigate the liver and biliary

tract. The commonest is ultrasound, which is safe, cheap and readily

available. It is often used as a screening modality. Computed tomography

or magnetic resonance imaging or both are often the next modalities of

choice, and are used to explore any unusual findings detected on ultra-

sound. Fluoroscopic imaging procedures, such as percutaneous cholangi-

ography and endoscopic retrograde cholangiopancreatography, are often

used when intervention is required, usually for therapeutic reasons.

In this article, we describe the role of the each of these imaging modal-

ities in benign and malignant hepatobiliary disease, summarize their use

in commonly encountered conditions, such as gallstones, cirrhosis and

focal liver lesions, and delineate their advantages and disadvantages.

Often, a combination of different modalities is required to reach the

final diagnosis. We also describe the complementary role of other less

commonly used modalities.

Keywords bile ducts; computed tomography; diffuse liver disease;

focal liver disease; gallstones; liver function tests; magnetic resonance

imaging; ultrasound

Imaging of the gallbladder and biliary tree

Plain radiography

This is rarely used in investigating liver and biliary pathology.

Only 10e20% of gallstones are calcified enough to be seen on

a plain abdominal radiograph. Calcifications may be due to

gallstones, a ‘porcelain gallbladder’, which predisposes to

cancer, or milk of calcium bile. Occasionally, the biliary system is

seen to contain air, which may be caused by endoscopic retro-

grade cholangiopancreatography (ERCP), biliary enteric-fistula

or biliary infections.

Ultrasonography (US)

Biliary disease: US is the commonest imaging investigation

used in the initial assessment of hepatobiliary disease. It is

Mo Malaki MB BChMRCS FRCR is a Specialist Registrar in Clinical Radiology

at Queen Elizabeth Hospital, Birmingham, UK. Interests are

Hepatobiliary Imaging and Interventional Radiology. Conflicts of

interest: none declared.

Kamarjit Mangat MB BCh FRCS FRCR is a Consultant Interventional

Radiologist with Specialist Interest in Hepatobiliary Imaging and

Interventional Radiology at Queen Elizabeth Hospital, Birmingham, UK.

Conflicts of interest: none declared.

MEDICINE 39:9 511

non-invasive, quick, and easy. No ionizing radiation is involved.

It is the ideal modality for gallstones, biliary dilatation and for

image-guided procedures.

The gallbladder (GB) can be visualized clearly in the fasting

state in most patients (Figure 1). It appears as an oval hypoechoic

structure with a thin uniform smooth wall. It may be invisible

because of a contracted GB, agenesis of the GB, calcified GB wall

or the presence of intramural or intraluminal gas.

On US, most gallstones are highly reflective, cause an acoustic

shadow and are generally mobile (Figure 2). False-negatives

occur when stones are smaller than 3 mm, hidden behind a GB

fold or cystic duct, if there is sludge present within the GB, or due

to technical factors, such as patient habitus or operator inexpe-

rience. False-positives are uncommon but occur because of

polyps, folds (spiral valve at the neck of the GB) or cholester-

olosis of the GB. US provides no information about GB or cystic

duct pathway function.

US has a positive predictive value and a negative predictive

value of 92% and 95% respectively in patients with a clinical

diagnosis of cholecystitis and a positive Murphy’s sign1

(Figure 2). It is also of value in diagnosis of complications of

cholecystitis, such as GB perforation, gangrenous cholecystitis

and emphysematous cholecystitis. Other GB pathologies identi-

fiable on US include carcinoma and adenomyomatosis. In

chronic cholecystitis, the GB is contracted, fibrosed and thick-

walled, and differentiation from GB neoplasm may be difficult.

Figure 1 A normal gallbladder (GB) in the fasting state is distended and

has a thin wall.

� 2011 Elsevier Ltd. All rights reserved.

Figure 2 Acute cholecystitis with gallstone and acoustic shadow. Note the thickened wall (**) of the gallbladder and the acoustic shadowing caused by the

gallstone (arrow). Not all stones cause a shadow on ultrasound as demonstrated in b (arrow).

INVESTIGATIONS

US is also reliable in assessing for biliary duct dilatation. The

common hepatic duct (CHD) measures 4e5 mm in diameter and

is seen in almost all patients. The intrahepatic ducts (IHD) are

seen only if they are dilated (Figure 3). The common bile duct

(CBD) is usually seen within the head of the pancreas (Figure 4).

The presence of gas in overlying stomach or transverse colon

may hinder good visualization.

The CBD usually measures up to 6 mm in diameter up to the

age of 60 with a further allowance of 1 mm per decade of age

thereafter.2 Previous surgery in the region, in particular chole-

cystectomy, can also account for non-pathological increases in

diameter. Other causes of duct dilatation include intestinal

hypomobility, hyperalimentation and prolonged fasting.

If true dilatation is found, an obstructive lesion should be

considered. A ‘double duct sign’ refers to a dilated pancreatic

duct and a dilated CBD, and is usually due to ampullary or

Figure 3 Normally the intrahepatic ducts (IHD) are not seen on ultrasound.

When visualized it is usually indicative of biliary obstruction (arrow). Note

the adjacent portal vein (pv). The proximity of the dilated IHD to the pv

produces the so-called parallel sign.

MEDICINE 39:9 512

pancreatic head neoplasia. Conversely, bile duct obstruction

without dilatation is seen in patients with low-grade or inter-

mittent obstruction resulting from stricture or small stones and in

patients with sclerosing cholangitis.

US is only 21e63% accurate for detecting bile duct stones3,4;

if no obvious cause for obstruction is seen further investigation is

mandatory.

Liver disease: US is used as a screening tool for suspected diffuse

or focal liver disease. It allows the architecture of the liver to be

easily assessed. A normal liver has a uniform homogenous

echotexture. The echogenicity is affected by several factors, such

as the equipment, the transducer frequency and the setting on

the US. It is best to compare the echogenicity against internal

structures, most commonly the right renal cortex. The liver is

slightly hyperechoic or isoechoic with respect to the renal cortex,

provided the latter is normal (Figure 5).

US is highly sensitive in differentiating cystic from solid

lesions, but not as sensitive as CT or MRI in characterizing focal

solid liver lesions (Figure 6). The sensitivity for detecting liver

metastases is 40e70%.5 It is generally unable to detect lesions

Figure 4 The normal portal vein (pv) and common bile duct (**).

� 2011 Elsevier Ltd. All rights reserved.

Figure 5 The liver is normally slightly isoechoic or hyperechoic to the right kidney (a). In disease states, such as a fatty liver (b), the liver appears brighter

than the adjacent kidney. Fatty liver is a non-specific appearance and can result from a range of conditions such as diabetes, hypercholesterolaemia and

excessive alcohol intake.

INVESTIGATIONS

smaller than 1 cm. The presence of diffuse liver disease lowers

the sensitivity of US for detection of focal lesions.

Doppler and colour US: colour flow andDuplex are used to assess

hepatic vessels and tumour vascularity. The velocity and wave-

form of the portal veins, hepatic vein and the hepatic artery can be

reliably assessed. The portal veins are of low velocity (10e20 cm/

second) with a respiratory variation (Figure 7). In portal hyper-

tension, there is loss or attenuation of the portal vein flow. On

Duplex US, the hepatic veins have a triphasic flow pattern similar

to the inferior vena cava (IVC) with respiratory and cardiac cycle

variation. The normal flow in the hepatic veins is towards the IVC.

In tricuspid regurgitation, there is reverse flow in the hepatic veins.

In cirrhosis, there is loss of the triphasic pattern since the encased

hepatic vein is less compliant. The hepatic artery has a high

Figure 6 Ultrasound is sensitive in detecting solid from cystic lesions, but

not in characterizing them. This lesion is hypoechoic compared to the rest

of the liver. It was thought to be due to focal fatty sparing, but needed

magnetic resonance imaging to confirm its nature and avoid the patient

any unnecessary biopsy.

MEDICINE 39:9 513

diastolic flow because of the low impedance and is in the same

directions as normal portal vein. In patients with transplanted

livers, the waveform of the hepatic artery is important in detecting

stenosis or thrombosis in the blood supply of the allograft. Colour

flow imaging and Duplex are also of use in detecting thrombus,

fistulas and aneurysms within the hepatic vasculature in patients.

Despite the potential advantages of US contrast agents, their

use remains limited in clinical practice, largely due to the need

for intravenous injection and the narrow window of time within

which the scan has to be performed after contrast injection.

Applications include the assessment of flow from tumours such

as hepatocellular carcinoma (HCC) and the detection and char-

acterization of hepatic lesions.6

Intraoperative US and laparoscopic US are highly sensitive for

detecting liver lesions not seen on routine pre-operative imaging

and are regularly used before hepatic resection. Endoscopic US

(EUS) is useful in evaluating the left lobe of the liver and lymph

nodes in the gastro-hepatic ligament, and for fine needle aspi-

ration of liver lesions. EUS is also useful for assessing lesions

near the ampulla, pancreatic or hilar lesions and it can also be

used to look for small CBD stones or biliary sludge.

Figure 7 The flow in the normal portal vein is monophasic. Colour Doppler

with the cursor placed in the middle of the portal vein.

� 2011 Elsevier Ltd. All rights reserved.

Figure 8 Multiple metastases (*) become more conspicuous compared to

the normal liver parenchyma in the portal venous phases, where they

appear as lower attenuation.

INVESTIGATIONS

Computed tomography (CT)

Advances such as helical CT and multi-detector CT (MDCT) have

enabled rapid image acquisition, improved spatial resolution and

the ability to image the liver and biliary tree in multiple phases of

contrast enhancement in a more precise manner. Advances in

post-processing images have allowed the acquisition of 3D

images of the hepatic vasculature (CT angiography), which is

important in mapping the vascular anatomy and tumour volume.

Intravenous iodinated contrast, routinely used for CT, increases

the relative attenuation difference between focal lesions and

vessels from the normal hepatic parenchyma, making them more

conspicuous. Enhancement of the liver is dependent on the phase

of contrast delivery during which scanning occurs. It also helps to

characterize focal lesions based on the enhancement pattern

during the various phases of contrast circulation in the liver.When

performed properly, CT will suffice for most clinical indications.

The various phases used are:

Unenhanced: unenhanced CT is not done routinely unless hae-

morrhage or calcification is suspected within liver lesions.

Figure 9 Dilated common bile duct (**) on US. The cause on US was unknown.

are of lower density compared to the adjacent intrahepatic branches of the po

the cause of the obstruction in this patient as being due to malignancy of th

MEDICINE 39:9 514

Arterial phase: the blood supply of the liver is via the hepatic

artery (25%) and the portal vein (75%). Most tumours have

a predominantly arterial supply.

Imaging the liver 20 seconds after the start of the injection of

contrast is ideal for detecting most lesions. It is a very short phase

because iodinated contrast diffuses rapidly from the intravas-

cular compartment into the extravascular space. Highly vascular

lesions are highlighted against a background of unenhanced

normal liver parenchyma. Such lesions will become isodense if

scanned later during the equilibrium phase, and may be missed.

Portal venous phase: this the commonest phase scan performed

for the evaluation of the liver. It is performed with a delay of

50e70 seconds after contrast injection. There is maximum

enhancement of the liver parenchyma. Many liver metastases are

supplied primarily by the hepatic artery and may therefore

appear as hypodense compared with the normal surrounding

liver parenchyma during this phase (Figure 8).

Delayed phase: delayed imaging, usually obtained 10e20

minutes after injection of contrast, is rarely used; it may be useful

in detecting intrahepatic cholangiocarcinoma and metastases.

Retention of contrast may be seen in cholangiocarcinoma,

fibrous tumours or scar tissue.

General uses of CT: CT is useful for assessing the overall back-

ground of the liver as well as giving an overview of the

remainder of the abdominal contents. Multiphase MDCT is an

effective method for detecting liver tumours and characterizing

focal lesions.

Themain indication for CT in biliary disease is for the diagnosis

and staging of GB carcinoma and in the evaluation of complica-

tions of cholecystitis, such as perforation and pericholecystic

abscess. It is also used in identifying the cause of biliary

obstruction, such as periampullary pancreatic tumours and nodal

masses at the porta hepatis (Figure 9). Often, it will differentiate

benign from malignant causes of obstruction and provides guid-

ance for biopsy and radiological staging of biliary tumours.

Despite the immense usefulness of CT in modern medical

practice there are some limitations that should be borne in mind.

First, its high radiation dose, especially in MDCT with multiphase

On CT, there is intrahepatic bile duct dilatation (*). The dilated ducts on CT

rtal vein, which contain contrasts within them. CT was useful in identifying

e pancreatic head (not shown).

� 2011 Elsevier Ltd. All rights reserved.

INVESTIGATIONS

scanning; second, its low sensitivity for detecting and charac-

terizing lesions smaller than 1 cm; and third, iodinated contrast is

contraindicated in those with history of previous anaphylaxis

and should be used with extreme caution in patients with renal

impairment.

Magnetic resonance imaging (MRI)

In many centres, MRI is the modality of choice for investigating

complex hepatobiliary disease. The development of rapid

acquisition techniques and tissue-specific contrast agents, the

lack of ionizing radiation and the ability to produce non-invasive

cholangiograms make it an ideal investigation. It is very useful in

investigating focal liver lesions, the biliary tree, intrinsic biliary

lesions, and to a lesser extent, extrinsic lesions. Its disadvantages

include cost, limited availability and the usual general contrain-

dications to MRI imaging.

MRI is used to diagnose and characterize diffuse and focal liver

disease, and obviates the need for invasive procedures such as

biopsy. Biopsies should be avoided if at all possible, especially if

tumour is suspected, due to the riskof seedingalong theneedle track.

Most focal hepatic lesions can be reliably characterized with MRI.

MRI is also useful for mapping of tumours before treatment.

Various sequences are used; these usually include dynamic

gadolinium imaging, with images acquired at different time

points after contrast administration. Contrast enhancement

patterns can be specific to focal lesions and a knowledge of

which contrast agent has been injected is vital as there is an

increasing portfolio of agents, several of which are liver-specific,

and their actions are different (Table 1).

Various contrast agents are used in MRI imaging of the liver.7

Gadolinium is the most commonly used agent. It has an excellent

distribution and behaves in a similar fashion to iodinated agents

MRI characteristics of focal liver lesions

Benign hypodense/hypoenhancing lesions

Hepatic cysts

Regenerating/dysplastic nodules e cirrhosis

Lipoma

Bile duct hamartoma

Malignant hypodense/hypoenhancing lesions

Hypovascular metastases (e.g. colon, lung, prostate, stomach)

Lymphoma

Arterial hyperdense/hyperenhancing lesions

Benign e hepatic adenoma

Malignant e HCC, hypervascular metastases (e.g. thyroid,

melanoma, breast, carcinoid)

Hypervascular lesion with central scar e FNH, fibrolamellar HCC

Delayed enhancement

Haemangiomas, cholangiocarcinoma

Other patterns of enhancement

Ring enhancement e abscess, pseudocysts,

post-radiofrequency ablation

Peripheral washout e metastases (e.g. carcinoid, breast)

FNH, focal nodular hyperplasia; HCC, hepatocellular carcinoma.

Table 1

MEDICINE 39:9 515

in CT. It diffuses out of the intravascular space into the liver

parenchyma, so rapid imaging is required to exploit differences

between the lesion and the surrounding liver parenchyma. In an

analogous manner to multiphase MDCT, imaging is obtained in

several timed phases. This dynamic contrast enhancement

pattern is useful in the characterization of many focal hepatic

lesions, including haemangiomas, hypervascular metastases,

HCC, focal nodular hyperplasia and adenomas.

Liver-specific agents, such as mangafadipir trisodium (taken

up by hepatocytes) and ferrumoxides (taken up by Kupffer cells),

demonstrate selective uptake in the liver and are used primarily

for lesion detection.8 Other agents that combine the effect of

gadolinium but are excreted primarily via the biliary tree are also

available and are used for characterization of focal lesions.

Fat suppression techniques are useful in detecting cellular fat;

they can differentiate pseudolesions, such as focal fatty infiltra-

tion and focal fat sparing, from pathological liver lesions. The use

of 3D gradient recall echo (GRE) sequences has improved MRI by

providing a dynamic, contrast-enhanced thin-section images

with fat saturation and high signal-to-noise ratio. The advantages

of 3D techniques over 2D dynamic imaging are the ability to

reformat in any plane, and high-quality thin-section imaging with

no gaps. In addition the data can be used to generate MR of the

vasculature (MR angiography).

Magnetic resonance cholangiopancreatography (MRCP) is

a non-invasive investigation that involves no contrast. It uses

heavily weighted T2 sequences to delineate the ductal system. It

involves acquiring thin-slice images, usually in a single breath

hold with reduced motion artefact, although more modern

machines are using respiratory gating sequences.

The diagnostic accuracy of MRCP is comparable to that of ERCP

in the evaluation of bile duct stones, malignant biliary obstruction

and anatomic variants, as well as the visualization of the pancreatic

Figure 10 Magnetic resonance cholangiopancreatography (MRCP) in

coronal plane. Note the dilated CBD and hepatic ducts. The obstruction

was due to a low CBD stone which appeared as a filling defect (arrow).

CBD stone on MRCP.

� 2011 Elsevier Ltd. All rights reserved.

Figure 11 Percutaneous transhepatic cholangiography demonstrating

dilated proximal CBD and intrahepatic ducts (thin arrows). The lack of

contrast in the mid-CBD was due to a stricture (thick arrow). Such

a ‘shouldered’ stricture is usually due to a malignant stricture.

INVESTIGATIONS

duct.9 MRCP is more sensitive than US and CT for imaging CBD

stones, and is especially accurate when these are symptomatic10,11

(Figure 10). Stones (as small as 2 mm) are seen as well-

circumscribed, low signal-intensity filling defects in the biliary tract.

MRCP has obvious advantages over direct cholangiog-

raphy: no contrast is injected, and there is no radiation and no

risk of the complications associated with ERCP, such as

pancreatitis, perforation or sepsis. There is also no require-

ment for sedation or post-sedation recovery. ERCP should be

reserved for those cases where a therapeutic option is

required at the same sitting.

The disadvantages ofMRCP are that gas, blood or abnormalities

within the duct can simulate stones, and signal loss can result from

surgical clips. Bright signal from fluid collections within the

abdomen, ascites or oedema may also interfere with the signal

from biliary fluid. Flow artefacts from adjacent vessels can also

Figure 12 This patient had undergone previous radiofrequency ablation (RFA) f

to determine whether the low-density lesion (*) was due to residual disease or

images from PET-CT (b) show that this lesion was highly metabolically active (þdue to residual disease activity rather than pure post-RFA changes.

MEDICINE 39:9 516

mimic filling defects. It is very important to review all of the images

in both the axial and coronal planes to avoid these pitfalls.12

Diffusion-weighted MRI is a form of MRI in which benign

lesions can be delineated frommalignant ones. Malignant lesions,

such as HCC and metastases, have a low apparent diffusion coef-

ficient (ADC), whereas benign lesions have a high ADC. This

modality is also used to differentiate liver abscess from cystic and

necrotic tumours. Abscesses are markedly hyperintense whereas

cystic or necrotic tumours are hypointense on MRI-DWI.13

Direct cholangiography

This involves two techniques, percutaneous transhepatic chol-

angiography (PTC) and ERCP. These are fluoroscopic techniques

in which the biliary tree is opacified directly using iodinated

contrast. Normally, antibiotics are given before the procedure to

prevent sepsis, and any coagulopathy corrected.

These techniques are most useful when intervention (stent

placement, drainages, biopsy or brushings) is needed. They have

over a 95% success rate in determining the site of obstruction

and more than a 90% success rate in determining its cause.

Direct cholangiography is also of use in assessment of biliary

leaks and biliary fistulas.

In PTC, the biliary system is punctured peripherally using

a fine needle (Chiba-22G), under US and/or fluoroscopic guid-

ance with contrast being injected to opacify the ducts. PTC is

uniformly successful in experienced hands where the ducts are

dilated and allows therapeutic measures (e.g. stent placement,

biopsy, stone extraction) to be undertaken (Figure 11).

In ERCP, the biliary tree is accessed distally in a retrograde

fashion via an endoscope. It allows visualization and biopsy of the

ampulla, and therapeutic manoeuvres such as sphincterotomy,

stone extraction and stenting. The pancreatic duct is also assessed.

The choice of PTC versus ERCP depends on several factors.

ERCP is impossible in the presence of previous biliary-enteric

bypass surgery. In our institution, ERCP is the method of

choice in cases of low biliary obstruction, whereas PTC is

preferred in cases where obstruction is higher than the common

hepatic duct, or where ERCP has failed.14

However, the choice of PTC versus ERCP also depends on the

planned therapeutic procedure; if CBD stones are suspected,

or this malignant liver lesion. On the follow-up CT scan (a), it was not easy

post-RFA changes. PET-CT was used as a problem-solving tool. The fused

) compared with the surrounding liver and therefore confirmed that it was

� 2011 Elsevier Ltd. All rights reserved.

INVESTIGATIONS

ERCP with sphincterotomy and endoscopic stone extraction may

be the procedure of choice. ERCP is also preferred in a non-

dilated biliary duct system, low CBD strictures, or where the

primary suspected pathology is in the pancreas and opacification

of the pancreatic duct is needed for the diagnosis.

Complication rates of direct cholangiography are 1e2% and

include sepsis, biliary peritonitis, bleeding, inadvertent damage

to adjacent organs and pancreatitis.

Nuclear medicine

Radionuclide imaging of the biliary tree is a useful technique that

can demonstrate both anatomy and function. In cholescintig-

raphy, the patient is injected with technetium 99 (radioisotope)

labelled hydroxy iminodiacetic acid (99m Tc-HIDA). The tracer is

taken up by the liver and excreted in the bile. In acute cholecystitis,

the main finding is prompt biliary excretion of the tracer without

demonstration of the GB. This method is very sensitive and

specific for the diagnosis of acute cholecystitis. Despite this, this

test is not commonly used for diagnosing acute cholecystitis. False

positives occur in patients with abnormal bile flow due to liver

disease, or a prolonged fast with a distended sludge-filled GB.

Positron emission tomography-CT (PET-CT)

This relativelymodernmethod of imaging has an important role in

the evaluation of metastatic disease. Most malignant lesions have

a higher metabolic activity than the surrounding tissues because

they have a greater uptake of glucose, labelled as 18F-fluoro-2-

deoxy D glucose (FDG) (Figure 12). This allows the identifica-

tion of malignant tumour foci.15 PET is highly sensitive and can

look at the entire body. The disadvantages are the cost, limited

availability and limited sensitivity for lesions less than 1 cm in

diameter. False positives occur in areas of increased metabolism,

due to inflammation or infection. PET-CT combines the advan-

tages of CT with the functional ability of PET by fusing the images.

This helps to localize the areas of abnormality accurately. A

REFERENCES

1 Ralls PW, Colletti PM, Lapin SA, et al. Real-time sonography in sus-

pected acute cholecystitis. Prospective evaluation of primary and

secondary signs. Radiology 1985 June; 155: 767e71.

MEDICINE 39:9 517

2 Wu CC, Ho YH, Chen CY. Effect of aging on common bile duct diam-

eter: a real-time ultrasonographic study. J Clin Ultrasound 1984

October; 12: 473e8.

3 Bortoff GA, Chen MY, Ott DJ, Wolfman NT, Routh WD. Gallbladder

stones: imaging and intervention. Radiographics 2000 May; 20:

751e66.

4 Stott MA, Farrands PA, Guyer PB, Dewbury KC, Browning JJ, Sutton R.

Ultrasound of the common bile duct in patients undergoing chole-

cystectomy. J Clin Ultrasound 1991 February; 19: 73e6.

5 Paulson EK. Evaluation of the liver for metastatic disease. Semin

Liver Dis 2001 May; 21: 225e36.

6 Wilson SR, Jang HJ, Kim TK, Burns PN. Diagnosis of focal liver

masses on ultrasonography: comparison of unenhanced and

contrast-enhanced scans. J Ultrasound Med 2007 June; 26:

775e87.

7 Hahn PF, Saini S. Liver-specific MR imaging contrast agents. Radiol

Clin North Am 1998 March; 36: 287e97.

8 Kamel IR, Bluemke DA. MR imaging of liver tumors. Radiol Clin North

Am 2003 January; 41: 51e65.

9 Reinhold C, Taourel P, Bret PM, et al. Choledocholithiasis: evaluation

of MR cholangiography for diagnosis. Radiology 1998 November;

209: 435e42.

10 Hartman EM, Barish MA. MR cholangiography. Magn Reson Imaging

Clin N Am 2001 November; 9: 841e55. vii.

11 Kim JH, Kim MJ, Park SI, et al. Using kinematic MR chol-

angiopancreatography to evaluate biliary dilatation. AJR Am J

Roentgenol 2002 April; 178: 909e14.

12 David V, Reinhold C, Hochman M, et al. Pitfalls in the interpretation of

MR cholangiopancreatography. AJR Am J Roentgenol 1998 April; 170:

1055e9.

13 Kamel IR, Bluemke DA, Ramsey D, et al. Role of diffusion-weighted

imaging in estimating tumor necrosis after chemoembolization of

hepatocellular carcinoma. AJR Am J Roentgenol 2003 September;

181: 708e10.

14 Silva MA, Tekin K, Aytekin F, Bramhall SR, Buckels JA, Mirza DF. Surgery

for hilar cholangiocarcinoma; a 10 year experience of a tertiary referral

centre in the UK. Eur J Surg Oncol 2005 June; 31: 533e9.

15 Zealley IA, Skehan SJ, Rawlinson J, Coates G, Nahmias C, Somers S.

Selection of patients for resection of hepatic metastases: improved

detection of extrahepatic disease with FDG pet. Radiographics 2001

October; 21 Spec No: S55e69.

� 2011 Elsevier Ltd. All rights reserved.