The British Journal of Radiology, 85 (2012), 1029–1037

Contrast-enhanced ultrasound of intrahepatic

cholangiocarcinoma: correlation with pathological examination
1,2 2 1 1 1
H-X XU, MD, PHD, L-D CHEN, MD, L-N LIU, MD, Y-F ZHANG, MD, L-H GUO, MD and 1C LIU, MD

1
Department of Medical Ultrasound, Tenth People’s Hospital of Tongji University & Shanghai Tenth People’s Hospital,
Medical Imaging Center of Tongji University, Shanghai, China, and 2Department of Medical Ultrasonics, First Affiliated
Hospital, Sun Yat-Sen University, Guangzhou, China

Objective: To investigate the correlation between enhancement patterns of

intrahepatic cholangiocarcinoma (ICC) on contrast-enhanced ultrasound (CEUS) and
pathological findings.
Methods: The CEUS enhancement patterns of 40 pathologically proven ICC lesions
were retrospectively analysed. Pathologically, the degree of tumour cell and fibrosis
distribution in the lesion was semi-quantitatively evaluated.
Results: 4 enhancement patterns were observed in the arterial phase for 32 mass-
forming ICCs: peripheral rim-like hyperenhancement (n519); heterogeneous
hyperenhancement (n56); homogeneous hyperenhancement (n53); and
heterogeneous hypo-enhancement (n54). Among the four enhancement patterns, the
differences in tumour cell distribution were statistically significant (p,0.05). The
hyperenhancing area on CEUS corresponded to more tumour cells for mass-forming
ICCs. Heterogeneous hyperenhancement (n52) and heterogeneous hypo-enhancement
(n52) were observed in the arterial phase for four periductal infiltrating ICCs. In this
subtype, fibrosis was more commonly found in the lesions. Heterogeneous Received 18 June 2011
hyperenhancement (n51) and homogeneous hyperenhancement (n53) were observed Revised 19 October 2011
in the arterial phase for four intraductal growing ICCs. This subtype tended to have Accepted 25 October 2011
abundant tumour cells.
DOI: 10.1259/bjr/21653786
Conclusion: The CEUS findings of ICC relate to the degree of carcinoma cell
proliferation at pathological examination. Hyperenhancing areas in the tumour always ’ 2012 The British Institute of
indicated increased density of cancer cells. Radiology

Intrahepatic cholangiocarcinoma (ICC) originates in Mass-forming ICC spreads between hepatocyte plates and
the small bile duct and is grouped according to the expands via the hepatic sinusoidal spaces. It often invades
International Classification of Diseases code, with hepa- the adjacent peripheral branches of the portal vein.
tocellular carcinoma (HCC) being the primary liver Periductal-infiltrating ICC tends to spread along the bile
tumour. It is the second most common primary liver duct wall via the nerve and perineural tissue of Glisson’s
tumour and is highly malignant. Although ICC is a capsule towards the porta hepatis. Intraductal-growing
relatively rare tumour, interest in this disease is rising ICCs are usually small or polypoid and do not invade
because incidence and mortality rates for ICC are deeply into the submucosal layer, often spreading super-
increasing steadily worldwide [1–5]. ficially along the mucosa surface. Characterisation of the
ICC is notoriously difficult to diagnose and is usually tumours in terms of their growth pattern is necessary for
fatal, owing to its late clinical presentation and the lack of optimal treatment planning and prognosis assessing. The
effective non-surgical therapeutic modalities. It tends to prognosis for mass-forming and periductal-infiltrating
present with non-specific symptoms such as malaise, cholangiocarcinoma is generally unfavourable, but is much
weight loss and abdominal pain. Most patients have better for the intraductal-growing type after surgical
unresectable disease at presentation and die within 12 resection, and long-term patient survival can be expected
months from the effects of cancer cachexia and a [7, 8].
subsequent rapid decline in performance status. Contrast-enhanced ultrasound (CEUS) has been in-
According to growth characteristics, ICC is subcate- creasingly applied in liver imaging. By administration of
gorised into mass-forming, periductal infiltrating or ultrasound contrast agents, CEUS can display dynamic
intraductal-growing types by the Liver Cancer Study blood flow perfusion and microcirculation of liver le-
Group of Japan [6]. These subtypes show different sions [9], similar to CT and MRI. In previous studies,
biological behaviours and have different clinical outcomes. CEUS had a similar diagnostic accuracy for ICC to CT
and was suggested as an alternative diagnostic option
Address correspondence to: Professor Hui-Xiong Xu, Department when CT examination was not available for patients with
of Medical Ultrasound, Tenth People’s Hospital of Tongji
University & Shanghai Tenth People’s Hospital, Medical Imaging
iodine allergy or impaired renal function [10]. It
Center of Tongji University, No. 301, Yanchangzhong Road, was confirmed that CT and/or MRI findings of ICC
Shanghai, China. E-mail: [email protected] were correlated with pathological findings; that is, the

The British Journal of Radiology, August 2012 1029

 H-X Xu, L-D Chen, L-N Liu et al

hyperenhancing areas always indicated a large num- patients, with a 4V1 vector transducer (frequency range,
ber of tumour cells and the areas of delayed enhance- 1.0–4.0 MHz). A contrast-specific mode of contrast pulse
ment corresponded to fibrotic stroma at pathological sequencing (Siemens Medical Solutions), which allowed
examination. In addition, different morphological sub- continuous real-time observation (mechanical index
types tended to exhibit distinct enhancement character- range, 0.15–0.21), was applied. Eight patients were
istics on CT [7, 8, 11–13]. On CEUS, besides the specific scanned with an Aplio XV (Toshiba Medical Systems,
feature of peripheral rim-like hyperenhancement, di- Tokyo, Japan) with a 375BT convex transducer (fre-
verse imaging findings of ICC were reported [9, 10, 14– quency range, 1.9–6.0 MHz). The contrast imaging mode
17]. These different CEUS appearances may reflect the was Contrast Harmonic Imaging (Toshiba Medical
differences in pathological subtypes or components of Systems) and the mechanical index range was 0.05–
ICC. The aim of this study was to investigate the cor- 0.08. The contrast agent was SonoVueH (BR1) (Bracco,
relation between the enhancement pattern of ICC on Milan, Italy), which was a sulphur hexafluoride-filled
CEUS and pathological findings. This information may microbubble.
be useful for diagnosis, treatment planning and prog- Two experienced radiologists with more than 3 years’
nostic evaluation of ICC. experience in liver CEUS performed the scanning. The
examinations were performed by both radiologists
separately prior to biopsy or surgical resection. Both
Methods and materials radiologists had received training in advance to ensure
that all the procedures were performed in the same
fashion. Before CEUS was undertaken, the radiologists
Patients
were told that the patients already had focal liver lesions
The study was approved by the ethics committee of detected by previous imaging studies but were blind to
the First Affiliated Hospital of Sun Yat-Sen University the results. Firstly, the whole liver was scanned with
and written informed consent was obtained from each baseline ultrasound (BUS) and the target lesion was
patient prior to CEUS. determined. Then the imaging mode was changed to
From March 2004 to April 2007, 59 patients with ICC CEUS with the transducer kept in a stable position. The
underwent CEUS examination in the hospital. The mechanical index settings and initial gain were adjusted
diagnoses of all patients were confirmed by pathological to suppress the echoes from the tissue and minimise
examination. The specimens were obtained by ultra- microbubble loss in the anatomical region of interest.
sound-guided percutaneous biopsy in 9 patients and by Subsequently, a volume of 2.4 ml of SonoVue was
surgery in 50. Among the 50 surgery cases, tumours in 3 administered with a bolus injection (within 1–2 s)
cases that were located in the left lobe of the liver (mean through the antecubital vein, followed by a flush of
size, 8.1¡3.4 cm; range, 4.6–11.4 cm, measured via 5 ml of normal saline using a 20-gauge cannula. The
ultrasound) were unresectable owing to abdominal wall whole CEUS process was observed continuously for
(n51) or lesser omentum (n52) infiltration. Partial more than 240 s and all cine clips were stored digitally on
hepatectomy was performed in 47 cases. Among them, a hard disk and subsequently transferred to a personal
seven patients had multiple lesions (two lesions in three computer for retrospective analysis. After 240 s, the
patients, three lesions in one and more than five lesions contrast agents washed out progressively and CEUS
in three); these patients were excluded because, in this was stopped when the contrast agents were nearly
retrospective study, it was uncertain whether or not invisible on the screen.
the lesion observed via CEUS was identical to the one
observed via pathological examination. 40 patients (21
males and 19 females: age range, 28–76 years; mean age, Image analysis
59¡12 years) who had solitary lesions were included in
our study for retrospective analysis. Two experienced readers, who had more than 3 years’
According to the cholangiocarcinoma classification experience in liver CEUS, assessed the CEUS images
proposed by the Liver Cancer Study Group of Japan [6], subjectively in consensus. The two readers were inde-
the lesions were classified into one of the three groups as pendent of the two who performed the ultrasound
either a mass-forming type (n532), in which the tumour is examination. Information about other imaging results,
an apparent nodule in the liver parenchyma and the clinical presentations, laboratory tests and histopathol-
borders between the cancerous and non-cancerous portions ogy of the patients was hidden to the readers. The CEUS
are distinct; a periductal-infiltrating type (n54), in which process was classified as arterial (8–30 s after the contrast
the tumour extends mainly along the peribile duct area and agent injection), portal (31–120 s) or late (121–240 s) phase
is elongated, spiculated or branchlike; or an intraductal- according to the European Federation of Societies for
growing type (n54), in which the tumour develops into the Ultrasound in Medicine and Biology guidelines for
bile duct lumen and forms an intraluminal occlusive lesion. CEUS [9]. The enhancement levels (hyper, iso- or hypo-
These classifications were based on the gross morphologi- enhancement, in comparison with adjacent liver par-
cal findings of the surgical specimens. enchyma) and their dynamic changes along with the
three phases were recorded. The enhancement patterns
in the arterial phase were subcategorised as follows:
CEUS examination peripheral rim-like hyperenhancement—a continuous
contrast hyperenhancing ring at the periphery of the le-
CEUS was performed with an ACUSON SequoiaTM 512 sion; heterogeneous hyperenhancement—enhancement
(Siemens Medical Solutions, Mountain View, CA) in 32 with different levels, with the highest level being

1030 The British Journal of Radiology, August 2012

CEUS of ICC: correlation with pathological examination

hyperenhancement; homogeneous hyperenhancement— After administration of contrast agent, 4 enhancement

hyperenhancement with similar levels; and heterogeneous patterns were observed in the arterial phase as follows:
hypo-enhancement—enhancement with different levels, peripheral irregular rim-like hyperenhancement in 19
with the highest level being hypo-enhancement. The (59.4%) lesions, heterogeneous hyperenhancement in 6
intratumoral blood vessels displayed during the arterial (18.8%), homogeneous hyperenhancement in 3 (9.4%)
phase were also evaluated. and heterogeneous hypo-enhancement in 4 (12.5%). 9
(28.1%) lesions displayed intralesional blood vessels in
the arterial phase (Figures 1–4).
Histopathological examination In the portal phase, 1 (3.1%) mass-forming ICC
showed iso-enhancement; the other 31 (96.9%) ICCs
All 40 resected specimens were fixed in 10% formalin showed hypo-enhancement. During the late phase, all 32
and .4 thin tissue sections were cut. For the mass- (100%) lesions faded out to show hypo-enhancement
forming type, two separate regions were chosen to cut in (Figures 1–4).
the largest dimension of the specimens: the peripheral
portion and the central portion, avoiding any large
vessels and large areas of necrosis or bleeding. Sub-
sequently, they were embedded in paraffin, cut into 4 mm Periductal-infiltrating intrahepatic
thick slices and stained with haematoxylin and eosin. A cholangiocarcinoma
pathologist with 6 years’ experience in liver disease who Periductal-infiltrating ICCs were observed in 4 (10%)
was blinded to the CEUS and other imaging findings lesions (mean size, 6.0¡3.5 cm; range, 3.0–9.8 cm). On
reviewed the histological specimens retrospectively. The BUS, 1 (25.0%), 1 (25.0%) and 2 (50.0%) lesions showed
stained sections were observed under a light microscope hyper, iso- and hypo-echogenicity, respectively. All
at 6200 magnification. The degree of tumour cell and (100.0%) of these were irregular and ill-circumscribed.
fibrosis distribution in the lesion was semi-quantitatively 2 (50.0%) showed vicinal bile duct dilation and 1 (25.0%)
evaluated with reference to previous reports [18, 19], showed stones in the lesion.
using a scoring system based on a 3-point scale in 10 On CEUS, heterogeneous hyperenhancement in 2
randomly chosen fields. This was defined as follows: 1, (50.0%) lesions and heterogeneous hypo-enhancement
little (,10% of field); 2, moderate (10–50%); 3, abundant in 2 (50.0%) lesions were observed in the arterial
(.50%). The CEUS enhancement patterns and the phase. The 2 (50.0%) hyperenhancing lesions displayed
densities of tumour cells and fibrosis in three types of intralesional blood vessels. All four lesions showed
ICC were compared. hypo-enhancement in both the portal and late phases.

Statistical analysis
Intraductal-growing intrahepatic
The quantitative data were expressed as mean ¡ cholangiocarcinoma
standard deviation. The comparison between the CEUS
results and pathological appearance was tested using the x2 Intraductal-growing ICCs were observed in 4 (10%)
test. A p-value of ,0.05 was considered statistically lesions (mean size, 3.9¡1.3 cm; range, 2.0–4.8 cm). On
significant. The statistical analyses were performed using BUS, all of the four intraductal-growing ICCs were
the SPSSH v. 13.0 software package (SPSS Inc., Chicago, IL). hyperechoic. 3 (75.0%) lesions were irregular and 1
(25.0%) was regular. 3 (75.0%) lesions had well-circum-
scribed boundaries and 1 (25.0%) had an ill-circum-
scribed boundary. Local bile duct dilation was observed
Results
around all (100.0%) lesions and a stone was observed in 1
(25.0%) lesion.
Ultrasound and contrast-enhanced ultrasound On CEUS, heterogeneous hyperenhancement in 1
appearances (25.0%) and homogeneous hyperenhancement in 3
In total, 40 lesions were observed. The mean size was (75.0%) lesions were observed in the arterial phase. No
6.7¡3.1 cm, with a range from 2.0 to 15.5 cm. The depth lesion showed intralesional blood vessels. All four
from the body surface to the bottom of the lesion ranged lesions showed hypo-enhancement in both the portal
from 3.7 to 16.2 cm (mean, 8.2¡2.8 cm). and late phases.

Mass-forming intrahepatic cholangiocarcinoma Histopathological findings

There were 32 (80%) mass-forming ICCs (mean size, Mass-forming ICCs were adenocarcinomas with vari-
7.1¡3.1 cm; range, 2.1–15.5 cm). On BUS, hyper, iso- and able differentiation: 6 (18.8%) of 32 lesions were well
hypo-echogenicity were seen in 5 (15.6%), 5 (15.6%) and 22 differentiated; 11 (34.4%) were moderately differentiated;
(68.8%) of the 32 lesions, respectively. 10 (31.2%) lesions and 15 (46.9%) were poorly differentiated. Glandular
were regular and 22 (68.8%) were irregular. 24 (75.0%) structures grew in cellular nests that were separated
lesions were well defined and 8 (25.0%) were ill-defined. 11 by variable fibrous stroma. Carcinoma cells were more
(34.4%) lesions showed peripheral bile duct dilation around actively proliferating in peripheral portions of the
the tumour and 7 (21.9%) showed stones in the masses. tumour, while the central portions were more sclerotic

The British Journal of Radiology, August 2012 1031

 H-X Xu, L-D Chen, L-N Liu et al

(a) (b) (c)

(d) (e) (f)

Figure 1. Intrahepatic cholangiocarcinoma. (a) Baseline ultrasound shows an ill-defined iso-echoic mass (arrows) in the right
lobe of the liver. (b) Contrast-enhanced ultrasound shows peripheral rim-like hyperenhancement (arrowheads) around the
tumour (arrows) during the arterial phase, but hypo-enhancement in the centre portion. (c) In the portal phase, the whole lesion
(arrows) shows hypo-enhancement. (d) The lesion (arrows) continues to show hypo-enhancement in the late phase. (e)
Microscopically, pathological examination reveals abundant tumour cells in the peripheral portion of the tumour
(haematoxylin–eosin stain, 6200 magnification). (f) There are scarce tumour cells in the centre portion of the tumour and
fibrosis is prominent (haematoxylin–eosin stain, 6200 magnification).

and hypocellular. Central necrosis was present in 15 distribution were statistically significant (p,0.05 at
(46.9%) lesions. both the peripheral and the central portions); however,
Of the 4 periductal-infiltrating ICCs, 2 (50.0%) were there were no significant differences in terms of fibrosis
well differentiated and the other 2 (50.0%) were distribution (p.0.05 at both the peripheral and the
moderately differentiated. Carcinoma cells infiltrated central portions). There was a trend that hyperenhancing
into the peribiliary glands, with considerable desmopla- areas of the lesion on CEUS corresponded to more
sia and thickening of the affected bile duct walls. tumour cells (Figures 1–4).
4 intraductal-growing ICCs were papillary adenocar- Of the periductal-infiltrating ICCs, one heteroge-
cinomas histologically; 3 (75%) lesions were well neously hyperenhanced lesion had a distribution of
differentiated and 1 (25.0%) was moderately differen- tumour cell score of 1; another had a score of 2. Both of
tiated. This type comprised innumerable frond-like these lesions had a distribution of fibrosis score of 3. The
infoldings of columnar epithelial cells. The tumours remaining two heterogeneously hypo-enhanced periduc-
were confined within the mucosal layer of the bile ducts tal-infiltrating ICCs were scored as 1 and 3, respectively,
and did not invade deep into the submucosal layer. with regard to the distribution of tumour cells and
fibrosis. In this subtype, fibrosis was more commonly
found in the lesion.
Correlation of contrast-enhanced ultrasound The degree of tumour cell distribution in three
enhancement patterns and pathological findings homogeneously hyperenhanced intraductal-growing
ICCs was scored as 3, and the degree of fibrosis dis-
The distribution of tumour cells and the scores in the tribution was scored as 1 in 1 (33.3%) and as 3 in 2
peripheral and central portions for the mass-forming (66.7%) lesions. The remaining heterogeneously hyper-
ICCs are presented in Tables 1 and 2. Among the four enhanced intraductal-growing ICC was scored as 2 with
enhancement patterns, the differences in tumour cell regard to the distribution of both tumour cells and

1032 The British Journal of Radiology, August 2012

CEUS of ICC: correlation with pathological examination

(a) (b) (c)

(d) (e) (f)

Figure 2. Intrahepatic cholangiocarcinoma. (a) Baseline ultrasound shows an ill-defined mass (arrows) with mixed echogenicity
in the right lobe of the liver. (b) Contrast-enhanced ultrasound shows heterogeneous hyperenhancement (arrows) during the
arterial phase. (c) In the portal phase, the lesion (arrows) becomes slightly hypo-enhancing. (d) The lesion (arrows) becomes
hypo-enhancing in the late phase. (e, f) Microscopically, abundant tumour cells are present in both the peripheral portion (e)
and the centre portion (f) (haematoxylin–eosin stain, 6200 magnification).

fibrosis. In total, the lesions in this subtype tended to which showed hyperattenuation in the arterial phase,
have abundant tumour cells. with higher attenuation during the portal and late
phases. The bile ducts proximal to the tumour were
dilated owing to the narrowed or obliterated duct
Discussion lumens [7]. Intraductal tumours could appear as a
hypo-attenuating mass in dilated bile ducts, with a very
ICC arises from the bile ducts peripheral to the low enhancement relative to the liver.
secondary bifurcation of the left or right hepatic duct. In previous studies, four different enhancement patterns
The role of imaging procedures in early detection and of ICC on CEUS were reported. These were periphe-
diagnosis is important. Before the introduction of CEUS, ral irregular rim-like hyperenhancement, heterogeneous
ultrasound was used only for screening in focal liver hyperenhancement, homogeneous hyperenhancement and
lesion detection; characterisation of the tumour usually heterogeneous hypo-enhancement [10, 14–17]. In the
required CT or MRI [7, 8]. present study, the most common type of ICC was the
An outstanding advantage of CT over BUS is that mass-forming type (80%). The common appearance of
administration of the contrast medium enables CT to mass-forming ICCs on BUS was a single hypo-echoic
demonstrate the microcirculation perfusion of the mass with irregular borders. On CEUS, 87.5% of them
tumour. For instance, the mass-forming type of ICC were hypervascular and 28.1% showed blood supply
has a distinctive appearance on CT; that is, a thin, mild, vessels within the tumour. The same four enhancement
incomplete rim-like contrast enhancement at the periph- patterns were demonstrated in this type. Among them,
ery and hypo-attenuating centres during both the arterial peripheral irregular rim-like hyperenhancement was
and the portal venous phases, but a mild or marked seen in 59.4% of lesions. This appearance corresponds
increase in tumour attenuation with progressive and well with the important pathological characteristic of
centripetal filling of contrast medium in the delayed ICC: marked fibrous stroma in the centre and abundant
phase [7, 8, 12]. The involved bile duct walls in carcinoma cells at the edge. In the homogeneous or
periductal-infiltrating tumours were diffusely thickened, heterogeneous hyperenhancing ICCs, carcinoma cells

The British Journal of Radiology, August 2012 1033

 H-X Xu, L-D Chen, L-N Liu et al

(a) (b) (c)

(d) (e) (f)

Figure 3. Intrahepatic cholangiocarcinoma. (a) Baseline ultrasound shows an iso-echoic mass (calipers) in the right lobe of the
liver. (b) The lesion (arrows) shows homogeneous hyperenhancement during the arterial phase. (c) The lesion (arrows) becomes
slightly hypo-enhancing during the portal phase. (d) The lesion (arrows) becomes hypo-enhancing during the late phase. (e, f)
Microscopically, abundant tumour cells are present in both the peripheral portion (e) and the centre portion (f) (haematoxylin–
eosin stain, 6200 magnification).

were prominent in the centre as well as at the the late phase [11, 12]. However, ultrasound contrast agents
periphery. By contrast, in heterogeneous hypo-enhan- have different pharmacokinetics. They are blood pool tracers
cing ICCs, only a small number of carcinoma cells were and are confined to the intravascular space. Therefore,
scattered at the edge and in the centre. It is well known instead of implicating the degree of fibrosis of the tumour,
that ultrasound contrast agents perform as blood pool CEUS demonstrates the real blood flow perfusion within the
tracers; accordingly, CEUS could provide information tumour.
about the degree of lesion vascularity. Pathophysio- Periductal-infiltrating ICC is pathologically identical
logically, carcinoma cells would proliferate densely in to infiltrating hilar cholangiocarcinoma, but has a
the areas with abundant blood supply. Therefore, as different location [7, 8]. It results in irregular narrowing
was demonstrated in this study for mass-forming ICC, of the involved bile duct and eventual obstruction. Four
the hyperenhancing areas on CEUS correspond with periductal-infiltrating ICCs in our study showed a
abundant carcinoma cells, pathologically, and different scirrhous type of growth, with extensive stromal
distributions of carcinoma cells may display different desmoplasia and sparse tumour cell infiltration. On
CEUS enhancement patterns. BUS, instead of a well-defined mass, an irregular,
It was reported that the areas of delayed or prolonged confounding mixed-echoic area in the liver with ambig-
enhancement in ICC at CT imaging, which corresponded to uous boundary may be found. On CEUS in our study,
fibrotic stroma at histopathological examination [7, 8, 11, 12], two lesions showed heterogeneous hypo-enhancement
could reflect the degree of fibrosis within the tumour. On the with no supplying vessels; the other two showed
contrary, no lesion in this group showed delayed enhance- heterogeneous hyperenhancement and tumour vessels
ment on CEUS, and there was no correlation between the were visualised. The latter enhancement may have been
CEUS enhancement patterns and the degree of fibrosis in associated with the fact that periductal-infiltrating
ICC. The delayed enhancement characteristics of ICC seem tumours invaded the bile duct wall or hepatic parench-
to be due to the CT contrast material being rapidly cleared yma in close association with or involving surrounding
from the blood pool and its retention within the large periductal blood vessels, providing a relatively abundant
interstitial space in the fibrous stroma of the tumour during vascular supply.

1034 The British Journal of Radiology, August 2012

CEUS of ICC: correlation with pathological examination

(a) (b) (c)

(d) (e) (f)

Figure 4. Intrahepatic cholangiocarcinoma. (a) Baseline ultrasound shows an ill-defined iso-echoic mass (arrows) in the left lobe
of the liver. (b) The lesion (arrows) shows heterogeneous hypo-enhancement during the arterial phase. (c) The lesion (arrows)
remains hypo-enhancing during the portal phase. (d) The lesion (arrows) continues to be hypo-enhancing during the late phase.
(e, f) Microscopically, scarce tumour cells are present in both the peripheral portion (e) and the centre portion (f) (haematoxylin–
eosin stain, 6200 magnification).

Four intraductal-growing ICCs were found. The mass findings were papillary adenocarcinoma, which is a low-
appeared specifically as a hyperechoic intraductal mass grade malignancy that forms sessile or polypoid masses,
within focal dilated bile ducts, with an irregular and consisting of innumerable papillary frond-like infoldings
clear margin on grey-scale ultrasound. They presented as of proliferated columnar epithelial cells and fibrovascu-
homogeneous or heterogeneous hyperenhancement in lar cores. Owing to their growth pattern, the tumours
the arterial phase, and as hypo-enhancement in the spread superficially along the mucosa of the bile duct
portal and late phases. The corresponding pathological instead of invading adjacent tissues [7, 8].

Table 1. Correlation of Contrast-enhanced ultrasound enhancement pattern and pathological distribution of tumour cells in 32
mass-forming intrahepatic cholangiocarcinomas
Distribution of tumour cells in Distribution of tumour cells in the
the peripheral portiona central portiona

Enhancement patterns Score 1 Score 2 Score 3 Score 1 Score 2 Score 3

Peripheral irregular rim-like hyperenhancement (n519) 0 (0) 8 (42.1) 11 (57.9) 18 (94.7) 1 (5.3) 0 (0)
Heterogeneous hyperenhancement (n56) 0 (0) 4 (66.7) 2 (33.3) 1 (16.7) 5 (83.3) 0 (0)
Homogeneous hyperenhancement (n53) 0 (0) 1 (33.3) 2 (66.7) 1 (33.3) 0 (0) 2 (66.7)
Heterogeneous hypo-enhancement (n54) 2 (50.0) 2 (50.0) 0 (0) 3 (75.0) 1 (25.0) 0 (0)
Total (n532) 2 (6.3) 15 (46.9) 15 (46.9) 23 (71.9) 7 (21.9) 2 (6.3)
p-value 0.007b ,0.001b
Data are expressed as number (percentage) unless otherwise indicated.
a
Score 1, little (,10% of field); score 2, moderate (10–50% of field); score 3, abundant (.50% of field).
b
Indicates statistically significant difference.

The British Journal of Radiology, August 2012 1035

 H-X Xu, L-D Chen, L-N Liu et al

Table 2. Correlation of contrast-enhanced ultrasound enhancement pattern and pathological distribution of fibrosis in 32 mass-
forming intrahepatic cholangiocarcinomas
Distribution of fibrosis in the Distribution of fibrosis in the
peripheral portiona central portiona

Enhancement patterns Score 1 Score 2 Score 3 Score 1 Score 2 Score 3

Peripheral irregular rim-like hyperenhancement (n519) 11 (57.9) 8 (42.1) 0 (0) 3 (15.8) 9 (47.4) 7 (36.8)
Heterogeneous hyperenhancement (n56) 2 (33.3) 4 (66.7) 0 (0) 1 (16.7) 3 (50.0) 2 (33.3)
Homogeneous hyperenhancement (n53) 2 (66.7) 1 (33.3) 0 (0) 1 (33.3) 2 (66.7) 0 (0)
Heterogeneous hypo-enhancement (n54) 2 (50.0) 2 (50.0) 0 (0) 2 (50.0) 1 (25.0) 1 (25.0)
Total (n532) 17 (53.1) 15 (46.9) 0 (0) 7 (21.9) 15 (46.9) 10 (31.2)
p-value 0.716 0.693
Data are expressed as number (percentage) unless otherwise indicated.
a
Score 1, little (,10% of field); Score 2, moderate (10–50% of field); Score 3, abundant (.50% of field).

Peripheral irregular rim-like hyperenhancement and In conclusion, the imaging findings of ICC on CEUS
heterogeneous hypo-enhancement are specific enhance- were related to the degree of carcinoma cell proliferation
ment patterns of the mass-forming type of ICC shown at pathological examination. hyperenhancing areas in the
on CEUS. Some metastatic liver cancers may show rim- tumour always indicated increased density of cancer
like hyperenhancement too, so exclusion of an extra- cells. Understanding the relationship between CEUS
hepatic primary site is the most important clue for enhancement and pathological examination could be
discrimination. hypervascular ICCs should be distin- helpful in diagnosis, treatment planning and prognostic
guished from HCCs. Information such as the back- evaluation for patients with ICC.
ground of liver cirrhosis or higher serum a-fetoprotein
values may be helpful [20]. On BUS, biliary sludge or
muddy stones can produce intraluminal echoes that References
may be difficult to distinguish from intraductal growing
ICCs. After injection of contrast agents, the presence of 1. de Groen PC, Gores GJ, LaRusso NF, Gunderson LL,
enhancement can eliminate the possibility of sludge or Nagorney DM. Biliary tract cancers. N Engl J Med
stones. 1999;341:1368–78.
2. Khan SA, Thomas HC, Davidson BR, Taylor-Robinson SD.
There were some limitations in this study. The case
Cholangiocarcinoma. Lancet 2005;366:1303–14.
numbers of periductal-infiltrating ICCs and intraduc-
3. Khan SA, Taylor-Robinson SD, Toledano MB, Beck A,
tal-growing ICCs were small, so it was difficult to Elliott P, Thomas HC. Changing international trends in
reach a conclusion. Like other retrospective studies, mortality rates for liver, biliary and pancreatic tumours.
there was limited control in the selection of the patient J Hepatol 2002;37:806–13.
population. In addition, we did not investigate the 4. Patel T. Increasing incidence and mortality of primary
relationship between CEUS enhancement appearances intrahepatic cholangiocarcinoma in the United States.
and microvessel density. Although there was no Hepatology 2001;33:1353–7.
direct association between CEUS enhancement pat- 5. Taylor-Robinson SD, Toledano MB, Arora S, Keegan TJ,
terns and tumour cell distribution (because contrast Hargreaves S, Beck A, et al. Increase in mortality rates from
agents were trapped only in the blood vessels and intrahepatic cholangiocarcinoma in England and Wales
tumour cells were abundant in areas of rich vascular- 1968–1998. Gut 2001;48:816–20.
ity), we showed that the CEUS findings correlated 6. Liver Cancer Study Group of Japan. The general rules for
the clinical and pathological study of primary liver cancer.
with tumour cell distribution. Another limitation was
Jpn J Surg 1989;19:98–129.
the use of two scanners that might have caused bias in 7. Lim JH. Cholangiocarcinoma: morphologic classification
image evaluation. However, both scanners worked according to growth pattern and imaging findings. AJR Am
under a low mechanical index; therefore, their J Roentgenol 2003;181:819–27.
influence was minimised. In addition, no objective 8. Lim JH, Park CK. Pathology of cholangiocarcinoma.
quantification or measure of reproducibility was Abdom Imaging 2004;29:540–7.
performed in this study, although quantification was 9. Claudon M, Cosgrove D, Albrecht T, Bolondi L, Bosio M,
applied in a previous study [17]. Owing to the Calliada F, et al. Guidelines and good clinical practice
relatively small number of cases, we did not analyse recommendations for contrast enhanced ultrasound (CEUS)-
the difference in enhancement pattern by lesion update 2008. Ultraschall Med 2008;29:28–44.
depth, which may have influenced the observed 10. Chen LD, Xu HX, Xie XY, Lu MD, Xu ZF, Liu GJ, et al.
enhancement pattern; however, the influence would Enhancement patterns of intrahepatic cholangiocarcinoma:
comparison between contrast-enhanced ultrasound and
have been small because the current high-end
contrast-enhanced CT. Br J Radiol 2008;81:881–9.
machine would have depicted the lesion clearly at
11. Asayama Y, Yoshimitsu K, Irie H, Tajima T, Nishie A,
such depth. Finally, associated diffuse liver diseases Hirakawa M, et al. Delayed-phase dynamic CT enhance-
such as fatty liver or cirrhosis would have influenced ment as a prognostic factor for mass-forming intrahepatic
the liver haemodynamics. However, also because of cholangiocarcinoma. Radiology 2006;238:150–5.
the small number cases, this influence was not 12. Lacomis JM, Baron RL, Oliver JH, Nalesnik MA, Federle
analysed in the current study and further study is MP. Cholangiocarcinoma: delayed CT contrast enhance-
mandatory. ment patterns. Radiology 1997;203:98–104.

1036 The British Journal of Radiology, August 2012

CEUS of ICC: correlation with pathological examination

13. Yoshida Y, Imai Y, Murakami T, Nishikawa M, Kurokawa SonoVue: initial experience. J Ultrasound Med 2006;25:
M, Yonezawa T, et al. Intrahepatic cholangiocarcinoma 23–33.
with marked hypervascularity. Abdom Imaging 1999;24: 17. Chen LD, Xu HX, Xie XY, Xie XH, Xu ZF, Liu GJ, et al.
66–8. Intrahepatic cholangiocarcinoma and hepatocellular carci-
14. Vilana R, Forner A, Bianchi L, Garcı́a-Criado A, Rimola J, de noma: differential diagnosis with contrast-enhanced ultra-
Lope CR, et al. Intrahepatic peripheral cholangiocarcinoma sound. Eur Radiol 2010;20:743–53.
in cirrhosis patients may display a vascular pattern similar 18. Okamura N, Yoshida M, Shibuya A, Sugiura H, Okayasu I,
to hepatocellular carcinoma on contrast-enhanced ultra- Ohbu M. Cellular and stromal characteristics in the
sound. Hepatology 2010;51:2020–9. scirrhous hepatocellular carcinoma: comparison with hepa-
15. Xu HX, Liu GJ, Lu MD, Xie XY, Xu ZF, Zheng YL, et al. tocellular carcinomas and intrahepatic cholangiocarcino-
Characterization of focal liver lesions using contrast- mas. Pathol Int 2005;55:724–31.
enhanced sonography with a low mechanical index mode 19. Maetani Y, Itoh K, Watanabe C, Shibata T, Ametani F,
and a sulfur hexafluoride-filled microbubble contrast agent. Yamabe H, et al. MR imaging of intrahepatic cholangio-
J Clin Ultrasound 2006;34:261–72. carcinoma with pathologic correlation. AJR Am J
16. Xu HX, Lu MD, Liu GJ, Xie XY, Xu ZF, Zheng YL, et al. Roentgenol 2001;176:1499–507.
Imaging of peripheral cholangiocarcinoma with low- 20. Bruix J, Sherman M. Management of hepatocellular
mechanical index contrast enhanced sonography and carcinoma. Hepatology 2005;42:1208–36.

The British Journal of Radiology, August 2012 1037