JOURNAL OF MAGNETIC RESONANCE IMAGING 18:181–188 (2003)
Original Research
Liver Metastases From Pancreatic
Adenocarcinoma: MR Imaging Characteristics
Ioana-Maria Danet, MD,1 Richard C. Semelka, MD,1* Larissa L. Nagase, MD,1
John T. Woosely, MD,2 Polytimi Leonardou, MD,1 and Diane Armao, MD1
Purpose: To describe the magnetic resonance imaging
spectrum of appearances of liver metastases from pancre-
atic ductal adenocarcinoma.
Materials and Methods: We retrospectively evaluated the
MRI exams performed between July 1996 and August 2001
in all patients who had liver metastases from pancreatic
adenocarcinoma and histopathologic diagnosis from either
the primary pancreatic tumor, liver metastases, or both.
Sixteen patients were included in the study. All MR studies
were performed at 1.5 T with a standard protocol including
T1- and T2-weighted images and serial post-gadolinium
spoiled gradient echo (SGE) images. Location, size, num-
ber, signal characteristics on T1- and T2-weighted images,
and pattern of enhancement on serial gadolinium-en-
hanced SGE images were assessed.
Results: The diameter of metastases ranged from a few
millimetres to 4 cm, and 12 patients (75%) had only lesions
of 1.5 cm or less. Capsular-based liver metastases were
found in 13 patients (81%) and three patients had only
capsular-based lesions with a diameter under 1.5 cm. Hy-
pervascular lesions were found in six patients (38%) and
hypovascular lesions in 10 patients (62%). Perilesional en-
hancement was present in 10 patients (62%), with six pa-
tients (38%) having ring perilesional enhancement and
eight patients (50%) having wedge-shaped perilesional en-
hancement.
Conclusion: On MR imaging, hepatic metastases from
pancreatic adenocarcinoma show a range of enhancement
patterns. Hypervascular metastases are not rare. Capsular
based distribution, small diameter, and perilesional en-
hancement are common features. This retrospective study
describes the MR imaging spectrum of appearances of liver
metastases from pancreatic adenocarcinoma in patients
with histopathologic confirmation of the diagnosis.
1
Department of Radiology, University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina.
2
Department of Pathology, University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina.
*Address reprint requests to: R.C.S., Department of Radiology, Univer-
sity of North Carolina at Chapel Hill, CB# 7510, Chapel Hill, North
Carolina 27599-7510. E-mail: richsem@med.unc.edu
Received November 8, 2002; Accepted April 8, 2003.
DOI 10.1002/jmri.10337
Published online in Wiley InterScience (www.interscience.wiley.com).
© 2003 Wiley-Liss, Inc. 181
Key Words: pancreas, MR; pancreas, neoplasm; liver, MR;
liver, neoplasm; MR contrast agents
J. Magn. Reson. Imaging 2003;18:181–188.
© 2003 Wiley-Liss, Inc.
DUCTAL ADENOCARCINOMA of the pancreas repre-
sents 95% of malignant pancreatic tumors and is the
fourth most common cause of cancer death in the
United States (1). The prognosis of pancreatic adeno-
carcinoma is very poor with a 5-year survival rate of 5%
(1). Upon initial investigation, 65% of the patients have
advanced local disease or hematogeneous metastases,
with the liver being the most common site of distant
metastases (2). On computed tomography (CT) and
magnetic resonance imaging (MRI), pancreatic adeno-
carcinoma is considered a hypovascular lesion since it
enhances to a lesser extent than the normal pancreatic
tissue on the arterial phase (1–3). It is generally believed
that metastases have the same angiographic character-
istics as the primary tumor (4), and therefore liver
metastases from pancreatic adenocarcinoma are con-
sidered hypovascular (4,5). To our knowledge, hyper-
vascular liver metastases from pancreatic adenocarci-
noma have been only occasionally reported (4,6).
Although both tumor vascularity and a high histologic
grade have been related to a more aggressive tumoral
behavior and a poorer prognosis in pancreatic adeno-
carcinoma, the relationship between these two factors
is still poorly understood (2,7,8). MRI is very sensitive in
analyzing the pattern of enhancement. Administration
of an extracellular space contrast agent such as gado-
linium chelate can provide estimates of the physiologic
properties of tumor microvessels such as perfusion,
diffusion, and distribution of contrast material within
the interstitial space (9). Moreover, the degree of vascu-
larization of a liver tumor is assessed on the arterial
phase of enhancement (10,11). To our knowledge, nei-
ther the spectrum of appearances of liver metastases
from pancreatic adenocarcinoma on MRI nor the fre-
quency of hypervascular liver metastases have been
described. Therefore, we performed this retrospective
study to describe the MRI spectrum of appearances of
liver metastases from pancreatic adenocarcinoma in
patients with histopathologic confirmation of the diag-
nosis.

182
MATERIALS AND METHODS
Patient Population
Over a period of five years (from July 1996 to August
2001) using the central computerized patient manage-
ment system, we retrospectively collected all patients
with liver metastases from pancreatic cancer who had
abdominal MRI studies and a histopathologic confirma-
tion of pancreatic ductal adenocarcinoma. A total of 16
patients were accrued (six men, 10 women; mean age
63 years). At the time the MRI studies were performed,
none of the patients had received treatment for liver
metastases, ensuring that imaging features were reflec-
tive of the natural state of the lesion. In 15 patients, the
diagnosis of pancreatic cancer with synchronous liver
metastases was established at the time of the MRI
study. In one patient, the liver metastases were dem-
onstrated on a two month follow-up MRI examination
after a Whipple procedure for the adenocarcinoma of
the head of the pancreas. Histopathologic material was
obtained from the pancreas in nine patients, from the
liver lesion in 10 patients, and from both in three pa-
tients. Liver metastases were evaluated by fine needle
aspirations (FNA) during exploratory laparoscopy/lap-
arotomy in five patients, and by radiologically guided
FNA in five patients. The pancreatic histopathologic
material was obtained as follows: 1) by radiologically
guided FNA in five patients; 2) at the time of autopsy in
two patients; and 3) after a Whipple procedure in one
patient. Review of the pathologic specimen was done by
an experienced pathologist in our center. The pancre-
atic tumor was located in the head of the pancreas in
seven patients, in the body of the pancreas in five pa-
tients, and in the tail of the pancreas in four patients. In
addition to the presence of metastases, 14 patients had
other criteria of advanced stage of disease as evidenced
by one or more of the following criteria: encasement of
the superior mesenteric artery or the celiac axis (eight
patients); peripancreatic adenopathy (five patients);
and invasion of adjacent organs (one patient).
MRI
All patients were imaged at the University of North
Carolina at Chapel Hill. MRI of the upper abdomen was
performed using a 1.5 T VISION (Siemens Medical Sys-
tems, Iselin, NJ). All MR examinations were performed
using a set protocol including: 1) precontrast T1-
weighted images acquired as breath-hold spoiled gradi-
ent echo (SGE) (TR ⫽ 120 –170 msec, TE ⫽ 4.0 – 4.5
msec, flip angle 80 –90°), and 2) T2-weighted half-Fou-
rier acquisition single-shot turbo spin echo (HASTE)
sequence (TR ⫽ infinite, effective TE ⫽ 90 msec, 2–3
acquisitions). Section thickness was 7–10 mm, and ma-
trix size was 128 –192 ⫻ 256 (phase ⫻ frequency encod-
ing) for all sequences. Gadolinium chelate (Magnevist,
Berlex, Wayne, NJ or Omniscan, Nycomed, Princeton,
NJ) was injected in a dosage of 0.1 mmol/kg as a rapid
bolus injection. Serial SGE images were acquired at 18
seconds (hepatic arterial-dominant phase) and at
45– 60 seconds (venous phase). A 90 –120 seconds
postinjection fat-suppressed SGE sequence was ac-
quired in all patients.
15222586, 2003, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.10337 by CochraneItalia, Wiley Online Library on [10/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Danet et al.
Image Interpretation
Two radiologists retrospectively examined the MR im-
ages to determine location, size, morphology, and signal
characteristics of liver metastases on precontrast T1-
weighted and T2-weighted images. The enhancement
pattern was assessed on serial post-gadolinium images.
Agreement regarding the lesion’s signal characteristics
and enhancement pattern was reached by consensus.
All MR exams were of good quality and showed an ad-
equate timing of the contrast material according to the
three phases of liver enhancement. A capsular-based
metastasis was defined as one in contact with the liver
capsule. All metastatic lesions were grouped into two
categories according to their size: those with a diameter
less than 1.5 cm and those greater than 1.5 cm. In
seven patients, metastases were too numerous to
count, thus four representative lesions were analyzed
for each patient. Hyperintensity on T2-weighted images
was determined subjectively as near isointense if the
signal intensity was slightly brighter than the liver pa-
renchyma, moderate if the signal intensity approached
that of the spleen, or markedly hyperintense if the sig-
nal intensity approached that of the cerebrospinal fluid.
Three phases of enhancement were determined based
on the location of gadolinium within the various hepatic
vessels. The arterial dominant phase was defined as the
phase in which contrast material was present in he-
patic arteries and portal veins and not within hepatic
veins. The portal venous phase showed contrast in ar-
teries, portal veins, and hepatic veins. The interstitial
phase of enhancement was acquired between two to
three minutes postinjection. Contrast enhancement of
the lesions was evaluated with attention to the pattern
of contrast material uptake on the arterial dominant
phase of enhancement and the presence of centripetal
progression of enhancement and washout on the portal
venous and interstitial phases. On the arterial domi-
nant phase of enhancement the lesions were defined as
hypervascular or hypovascular. A hypervascular lesion
was defined as a lesion that showed greater enhance-
ment than the liver parenchyma approximating the sig-
nal of normal pancreas or renal cortex on the arterial
dominant phase of enhancement. A hypovascular le-
sion was defined as one that showed less gadolinium
uptake than the liver parenchyma. Perilesional en-
hancement was defined as enhancement that occurs
beyond the lesion margins in the surrounding liver pa-
renchyma on past contrast images when compared to
images obtained precontrast administration. Perile-
sional enhancement was further defined as wedge-
shaped or circumferential. The pattern of enhancement
was considered either diffuse or peripheral on the arte-
rial dominant phase of enhancement. The progression
of enhancement over time was determined on the portal
and interstitial phases of enhancement. It was consid-
ered as incomplete when there was lack of enhance-
ment of the central portion of the tumor or complete
when the total volume of the tumor showed enhance-
ment, either isointense or hyperintense relative to liver
parenchyma. Lesional washout was considered when
the lesion showed a decrease in enhancement to hy-
pointensity in comparison to liver parenchyma.
 15222586, 2003, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.10337 by CochraneItalia, Wiley Online Library on [10/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Liver Metastases From Pancreatic Adenocarcinoma 183

Figure 1. Hypovascular liver metastases from pancreatic ductal adenocarcinoma in a 67-year-old female patient. Precontrast
T2-weighted HASTE image (a), T1-weighted SGE images (b), postcontrast arterial dominant phase (c), and interstitial phase (d)
images. On T2-weighted images the lesions are of minimal high signal intensity (a) and of low signal intensity on T1-weighted
images (b). Negligible enhancement of the lesion is demonstrated on early and late postcontrast images (c,d). Transient
moderately intense circumferential perilesional enhancement is identified on the arterial dominant phase. Perilesional enhance-
ment is demonstrated as enhancement beyond the lesion margins on the arterial dominant phase (c) when compared to the
precontrast T1-weighted images (b).

RESULTS On T1-weighted images, all patients had mildly hy-

A total of 50 metastatic lesions (size range from 5 mm to
4 cm, mean of 1.2 cm) were assessed in 16 patients. All
patients had lesions under 1.5 cm, and 12 patients
(75%) had only lesions equal to or less than 1.5 cm. A
capsular-based and central distribution of the lesions
was found in 11 (69%) patients. Three patients (19%)
had only capsular-based lesions and two patients (13%)
had only central lesions. All of the patients with metas-
tases involving only the capsular surface had only le-
sions smaller than 1.5 cm.
pointense lesions and additionally three patients had
lesions isointense relative to the liver parenchyma. In
patients with lesions under 1.5 cm, these lesions were
difficult to visualize on T1-weighted images. On T2-
weighted images, the lesions were of moderately high
signal intensity in 13 patients and isointense in three
patients. The patterns of enhancement are summarized
in Table 1. On the arterial dominant phase, a hypovas-
cular pattern of enhancement was encountered in 10
patients (27 lesions) (Fig. 1). It was further character-
 15222586, 2003, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.10337 by CochraneItalia, Wiley Online Library on [10/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
184 Danet et al.

Figure 2. Hypervascular liver metastases from pancreatic carcinoma in a 61-year-old male patient. Precontrast T2-weighted
HASTE image (a), T1-weighted SGE image (b), postcontrast arterial dominant phase (c), and interstitial phase (d). On T2-
weighted images the lesions are almost isointense to the liver parenchyma and of low signal intensity on T1-weighted images.
Early enhancement is either homogeneous (arrow, c) or peripheral ring (arrowhead, c) and there is progression to isointensity on
delayed images (d). An adenocarcinoma of the tail of the pancreas is demonstrated on postcontrast images (curved arrow c,d).

ized as follows: 1) a faint ring of enhancement in five
patients, and 2) negligible enhancement identified in
five patients. In this latter group, a peripheral ring of
enhancement was identified on the portal phase in
four patients. On the interstitial phase, eight of the
10 patients with hypovascular metastases showed
either an isointense homogeneous enhancement (one
patient) or a persistent nonenhancing central core
(seven patients). In one patient, the two types of cen-
tripetal progression were encountered, and in the last
patient negligible enhancement was identified in all
phases of enhancement. A hypervascular pattern of
enhancement was encountered in six patients (23
lesions) on the arterial dominant phase (Fig. 2). Pe-
ripheral ring of enhancement was identified in all
patients. In addition, lesions under 1.5 cm in diam-
eter (four patients) also showed homogeneous en-
hancement. On the portal venous phase or the inter-
stitial phase, the hypervascular lesions showed either
a persistent nonenhanced central core (five patients),
progression to complete isointense (three patients), or
hyperintense homogeneity (four patients) relative to
the liver parenchyma. In three patients, both pat-
terns of progression were observed. In two patients,
peripheral washout was identified on the interstitial
phase. Pronounced late enhancement was identified
in seven patients (six patients with hypervascular
metastases and one patient with hypovascular me-
tastases). No patient possessed both hypervascular
and hypovascular metastases.
 15222586, 2003, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.10337 by CochraneItalia, Wiley Online Library on [10/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Liver Metastases From Pancreatic Adenocarcinoma 185

Figure 3. Wedge-shaped perilesional enhancement in a 59-year-old male patient. Precontrast T2-weighted HASTE image
(a),T1-weighted SGE images (b), postcontrast arterial dominant phase (c), and interstitial phase (d). Wedge-shaped early
perilesional enhancement associated with a metastatic lesion is clearly depicted on immediate post-gadolinium image (c) that
fades rapidly (d).

Circumferential perilesional enhancement was DISCUSSION

present in six patients, and wedge-shaped perilesional
enhancement in eight patients (Fig. 3). The perilesional
enhancement was most conspicuous on the arterial
dominant phase of enhancement and showed fading to
isointensity on the subsequent phases of enhancement.
A metastatic lesion was identified in all cases with
wedge-shape perilesional enhancement.
In two cases, the MRI report included hepatic hamar-
tomas in the differential diagnosis. The lesions mea-
sured less than 1 cm in both cases and showed a pe-
ripheral ring of enhancement on the early and
interstitial phase.
Liver metastases from pancreatic adenocarcinoma have
been described as typically multiple and of small size
(12,13). This is confirmed in our study as 81% of the
patients had multiple lesions, with 38% having over 10
lesions. All patients had small lesions under 1.5 cm and
the majority (81%) had only lesions under 1.5 cm. The
size of a tumor is related to the degree of cellular pro-
liferation, and the presence of neovascularization or
angiogenesis (defined as the formation of new blood
vessels from preexisting microvessels) is considered es-
sential for tumor growth beyond a millimeter in solid
tissues (9,14). It has been suggested that a relatively

186
Table 1.
Enhancement Patterns of Liver Metastases From Adenocarcinoma of the Pancreas
Vascularity Arterial phase
Hypervascular, 6 patients Intense homogeneous
Intense peripheral rim
Hypovascular, 10 patients Faint peripheral rim
Negligible
low angiogenetic activity in pancreatic adenocarcinoma
may be responsible for the small size of the metastatic
lesions (13,15). The small size of the liver metastases
may also relate to the aggressiveness of their malig-
nancy such that patients do not live long enough to
develop large metastases.
In our study, capsular based hepatic metastases
were noteworthy and were identified in a large number
of the patients (88%), while 18% (3) showed capsular-
based metastases exclusively. The capsular-based dis-
tribution may be related to dissemination of pancreatic
adenocarcinoma through direct seeding of the perito-
neal cavity and serous coat of the liver. Alternatively,
since pancreatic carcinoma shows a penchant for inva-
sion of the adjacent vascular structures, subcapsular
metastases may reflect the hematogeneous spread of
tumor via the hepatic artery that reaches the surface of
the organ and ends in its fibrous coat in stellate plex-
uses (16).
Generally, it has been considered that the vascularity
of the metastases is similar to the vascularity of the
primary focus (4,5). From the radiological point of view,
pancreatic ductal adenocarcinoma is considered a hy-
povascular tumor since, on the arterial phase, it en-
hances to a lesser extent pancreatic tissue uninvolved
by tumor (1,8). This feature was present in all patients
in our study. Thus, metastases from pancreatic adeno-
carcinoma have been generally considered as hypovas-
cular (4,17). However, in 38% of the cases in our study,
the vascularity of the liver metastases did not correlate
with the enhancement of the primary tumor, as these
metastases showed a hypervascular pattern of en-
hancement. Angiographic studies have demonstrated
irregular and tortuous small vessels in up to 75% of the
cases of pancreatic adenocarcinoma (18). Endothelial
immunohistochemical staining measurements of an-
giogenetic activity have shown relatively low microves-
sel density (MVD) counts for ductal adenocarcinoma of
the pancreas in comparison with other neoplasms such
as bladder or gastric carcinoma (15). However, a hyper-
vascular subgroup of pancreatic adenocarcinoma with
high MVD counts has been described in up to 45% of
15222586, 2003, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.10337 by CochraneItalia, Wiley Online Library on [10/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Danet et al.
Patients
Portal venous and interstitial phase
N ⫽ 16
Hyperintense homogeneous 1
Isointense/hyperintense homogeneity 3
Incomplete central progression or 6
progression to isointense/hyperintense
homogeneity
Peripheral washout 2
Centripetal progression with lack of 5
central enhancement and/or
progression to isointense/hyperintense
homogeneity
Ring enhancement with centripetal 4
progression with lack of central
enhancement and/or progression to
isointense/hyperintense homogeneity
Negligible 1
the cases in two series in the pathology literature (8,15).
The relatively large number of patients with hypervas-
cular liver metastases in our study may be related to
the presence of a hypervascular type of pancreatic ad-
enocarcinoma. However, complex series of interactions
between the tumor and the host microenvironment
have been described that influence the degree of vascu-
larization of primary tumor and metastases. We believe
this may be the major contribution to explain the dif-
ference in vascularization between the primary tumor
and the metastatic lesions (19,20). The native liver pro-
vides a richly vascularized environment with its dual
blood supply and sinusoidal architecture. This ana-
tomic feature together with angiogenetic factors gener-
ated by both host and metastatic tumor cells (21) may
work in concert to render metastases hypervascular.
On precontrast images, liver metastases with a diam-
eter greater than 1.5 cm were generally identified.The
most common appearance was that of moderately high
signal intensity on T2-weighted images and a moder-
ately low signal intensity on T1-weighted images. These
signal characteristics have been commonly described
for liver metastases in general (10,22). Metastases with
a diameter less than 1.5 cm were inconsistently shown
on noncontrast images. On the arterial phase, hyper-
vascular metastases showed greater enhancement rel-
ative to the liver parenchyma with the most common
appearance being that of a peripheral rim of enhance-
ment. As previously observed, less than 1.5 cm metas-
tases often showed homogeneous enhancement (23).
Peripheral washout, considered as a specific sign of
metastases and reflecting rapid arterio-venous transit
in the periphery of the metastasis (24), was identified in
only two patients with hypervascular metastases. Hy-
povascular lesions showed either a faint peripheral ring
or negligible enhancement on the arterial phase. Re-
gardless of the vascularity of the metastases, 94% of the
patients in our study demonstrated incomplete centrip-
etal progression with persistence of a nonenhanced
core on delayed images. Centripetal progression of en-
hancement on delayed images is a relatively common
feature of liver metastases and is suggestive of slow flow

Liver Metastases From Pancreatic Adenocarcinoma
or slow passage of contrast material into the extracel-
lular space in the central aspect of the tumor. The lack
of total central enhancement may be related to necro-
sis, fibrosis, compression of central vessels (25), or slow
centripetal flow. Pronounced late enhancement was ob-
served in metastases in 44% of the patients in our
study. Adenocarcinoma of the pancreas has a high con-
tent of desmoplastic or fibrous tissue (1) that may con-
tribute to the prolonged enhancement on delayed im-
ages. Perilesional enhancement on dynamic contrast
enhanced MR images has been observed in hepatic
metastases most commonly from colorectal carcinoma.
Semelka et al (26) showed that intense perilesional en-
hancement on early gadolinium SGE images correlates
with peritumoral desmoplastic reaction, inflammatory
cell infiltration, and vascular proliferation. In our
study, perilesional enhancement was present in 63% of
the patients and the wedge-shaped pattern was the
most common. We believe the wedge-shaped pattern of
enhancement is related to a vascular imbalance be-
tween the arterial and the portal supply with resultant
increased hepatic arterial enhancement. However, the
precise underlying pathophysiologic cause remains un-
certain. It is usually most conspicuous on the arterial
dominant phase and results in an increase in the re-
gional arterial flow. Although a metastatic lesion was
found within all wedge-shaped areas of enhancement,
the presence of the metastasis itself may not be the only
factor responsible for this vascular imbalance, as it
does not seem to be a common feature in other meta-
static lesions. Since increased perilesional enhance-
ment was observed in both hypovascular and hypervas-
cular metastases, an increase in hepatic arterial supply
of metastases is not solely responsible for this effect.
Certain features of pancreatic adenocarcinoma, such
as its tendency to be highly desmoplastic and to invade
vessels, may be responsible in part for this focal wedge-
shaped vascular imbalance. Another consideration to
account for this feature may be compression or inva-
sion of adjacent portal veins causing compensatory in-
creased hepatic arterial supply.
Although the ring enhancing pattern has been shown
to be specific for metastases (10,27), the differential
diagnosis to be considered in the presence of small liver
lesions with a persistent rim of enhancement is that of
hepatic hamartoma. This differential diagnosis was
mentioned at the time of the MRI interpretation for two
patients in our study. The presence of benign liver le-
sions in patients with a known primary tumor is com-
mon (28), and careful evaluation of the signal charac-
teristics and enhancement pattern is mandatory.
Hepatic hamartomas, which are present in approxi-
mately 3% of the population, are usually small (⬍1 cm)
lesions of high signal intensity on T2-weighted images
and demonstrate a thin rim of enhancement on early
and late postcontrast image (29). A distinction between
the two lesions may depend on the observation that
enhancement in metastases often progresses centrally,
whereas hamartomas do not show progressive en-
hancement. However, this distinction may be difficult
to make in lesions smaller than 1 cm, particularly when
metastases may be extremely hypovascular or cystic.
15222586, 2003, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.10337 by CochraneItalia, Wiley Online Library on [10/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
187
One important limitation in our study is the small
number of patients, which reflects both the single in-
stitution accrual and the limited number of concurrent
histopathologic examinations. On the basis of these
initial observations, it would be of interest to study a
larger population in order to correlate the metastatic
pattern of enhancement on MR images, tumoral MVD,
and patient outcome.
In conclusion, a variable pattern of enhancement of
metastases from pancreatic adenocarcinoma can be
demonstrated by MRI. Although adenocarcinoma of the
pancreas from the radiological point of view is consid-
ered a hypovascular tumor, hypervascular liver metas-
tases are common and are found in 38% of the patients.
In our study, hepatic metastases from pancreatic car-
cinoma generally presented as small and multiple he-
patic lesions. We observed that a capsular based distri-
bution is common, underscoring the fact that careful
attention should be paid to this location when staging
pancreatic adenocarcinoma. Perilesional enhancement
is also a common feature of these metastases with a
wedge-shaped pattern most commonly observed.
REFERENCES
1. Martin DR, Semelka RC. MR imaging of pancreatic masses. Magn
Reson Imaging Clin North Am 2000;:8:;787– 812.
2. Rosai J. Pancreas and ampullary region. In: Rosai J, editor. Acker-
man’s surgical pathology. 8th ed. St.Louis: Mosby; 1996. p 969 –
1013.
3. Semelka RC, Nagase LL, Armao D, et al. Pancreas. In: Semelka RC.
Abdominal-pelvic MRI. 2nd ed. New York: Wiley-Liss; 2002. p 373–
490.
4. Baum S. Hepatic arteriography. In: Baum S, editor. Abram’s an-
giography. 4th ed. New York: Little-Brown; 1997. p 1433–1456.
5. Nino-Murcia M, Olcott EW, Brooke R, et al. Focal liver lesions:
pattern-based classification scheme for enhancement at arterial
phase CT. Radiology 2000;215:746 –751.
6. Noda A, Inaba K, Sakuragi S, et al. Hypervascular liver metastases
from hypovascular ductal cell carcinoma of the pancreas. Intern
Med 2001;40:227–231.
7. Karademir S, Sokmen S, Terzi C, et al. Tumor angiogenesis as a
prognostic predictor in pancreatic cancer. J Hepatobiliary Pancreat
Surg 2000;7:489 – 495.
8. Ikeda N, Adachi M, Taki T, et al. Prognostic significance of angio-
genesis in human pancreatic cancer. Br J Cancer 1999;79:1553–
1563.
9. Padhani AR, Neeman M. Challenges for imaging angiogenesis. Brit-
ish J Radiol 2001;74:886 – 890.
10. Semelka RC, Braga L, Armao D, et al. Liver. In: Semelka RC.
Abdominal-pelvic MRI, 2nd ed. New York: Wiley-Liss; 2002:33–
318.
11. Baron RL. Understanding and optimizing use of contrast material
for CT of the liver. Am J Roentgenol 1994;163:323–331.
12. Stephens DH. Pancreatic adenocarcinoma: radiologic diagnosis
and staging. In: Margulis AR, Burhenne HS, editors. Alimentary
tract radiology, 5th ed. St. Louis, MO: Mosby; 1994. p 1107–1126.
13. Ellis LM, Takahashi Y, Fenoglio CJ, et al. Vessel counts and vas-
cular endothelial growth factor expression in pancreatic adenocar-
cinoma. Eur J Cancer 1998;34:337–340.
14. Klauber-DeMore N, Van Zee KJ, Linkov I, et al. Biological behavior
of human breast cancer micrometastases. Clin Cancer Res 2001;
7:2434 –2439.
15. Fujimoto K, Hosotani R, Wada M, et al. Expression of two angioge-
netic factors, vascular endothelial growth factor and platelet-de-
rived endothelial cell growth factor in human pancreatic cancer,
and its relationship to angiogenesis. Eur J Cancer 1998;34:1439 –
1447.
16. Gray H. The liver. In: Gray H, editor. Anatomy of the human body.
Philadelphia: Lea-Febiger; 1948. p 1233–1243.
17. Schnider G. Abdomen. In: Runge VM, editor. Clinical MRI. 1st ed.
Philadelphia: WB Saunders; 2002. p 289 –352.

188
18. Tylen U. Accuracy of angiography in the diagnosis of carcinoma of
the pancreas. Acta Radiol 1973;14:449 – 465.
19. Fidler IJ, Ellis LM. The implications of angiogenesis for the biology
and therapy of cancer metastasis. Cell 1994;79:185–188.
20. Passe TJ, Bluemke MD, Siegelman SS. Tumor angiogenesis: tuto-
rial on implications for imaging. Radiology 1997;203:593– 600.
21. Bruns CJ, Ellis LM, Radinsky R. Angiogenetic strategies in pancreatic
cancer. In: Evans DB, Pisters PWT, Abbruzzese JL, editors. Pancreatic
cancer. 1st ed. New York: Springer-Verlag; 2002. p 357–357.
22. Imam K, Bluemke DA. MR Imaging in the evaluation of liver me-
tastases. MR Clin N Am 2000;8:741–767.
23. Larson RL, Semelka RC, Bagely AS, et al. Hypervascular malignant
liver lesions: comparison of various MR imaging pulse sequences
and dynamic CT. Radiology 1994;192:393–399.
24. Mahfouz AE, Hamm B, Wolf KJ. Peripheral washout: a sign of
malignancy on dynamic gadolinium-enhanced MR images of focal
liver lesions. Radiology 1994;190:49 –52.
15222586, 2003, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jmri.10337 by CochraneItalia, Wiley Online Library on [10/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Danet et al.
25. Folkman J, Beckner K. Angiogenesis imaging. Acta Radiol 2000;7:
783–785.
26. Semelka RC, Hussain SM, Marcos HB, Woosley JT. Perilesional
enhancement of hepatic metastases: correlation between MR im-
aging and histopathologic findings-initial observations. Radiology
2000;215:89 –94.
27. Mitchell DG, Sani S, Weinreb J, et al. Hepatic metastases and
cavernous hemangiomas: distinction with standard- and triple-
dose gadoteriol-enhanced MR imaging. Radiology 1994;193:49 –57.
28. Noone TC, Semelka RC, Balci NC, Graham ML. Common occur-
rence of benign liver lesions in patients with newly diagnosed
breast cancer investigated by MRI for suspected liver metastases. J
Magn Reson Imaging 1999;10:165–169.
29. Semelka RC, Hussain SM, Marcos HB, Woosely JT. Biliary hamar-
tomas: solitary and multiple lesions on current MR techniques
including gadolinium enhancment. J Magn Reson Imaging 1999;
10:196 –201.