AJR 2005; 184:1541-1548
© American Roentgen Ray Society
MRI Findings of Focal Eosinophilic Liver Diseases
Young-Kon Kim1,
Chong Soo Kim1,
Woo Sung Moon2,
Baik-Hwan Cho3,
Sang Yong Lee1 and
Jeong Min Lee4
1 Department of Diagnostic Radiology, Chonbuk National University Hospital,
Chonju, Korea.
2 Department of Pathology, Chonbuk National University Hospital, Chonju,
Korea.
3 Department of General Surgery, Chonbuk National University Hospital, Chonju,
Korea.
4 Department of Radiology and Institute of Radiation Medicine, Seoul National
University Hospital and College of Medicine, 28, Yongon-dong, Chongno-gu,
Seoul 110-744, South Korea.
Received April 8, 2004;
accepted after revision August 14, 2004.
Address correspondence to J. M. Lee.
Abstract
OBJECTIVE. The purpose of this study was to describe the MRI
findings of focal eosinophilic infiltration and eosinophilic abscess of the
liver.
CONCLUSION. MRI shows characteristic findings of focal eosinophilic
liver disease that can be helpful in differentiating lesions of focal
eosinophilic liver disease from other focal liver lesions. In addition,
eosinophilic abscess and focal eosinophilic infiltration showed different MRI
findings from each other.
Introduction
Focal eosinophilic liver disease is a relatively common disease and is
associated with a variety of disorders such as parasitic infestations,
allergic conditions, internal malignancies, drug hypersensitivity, and
hypereosinophilic syndrome
[1–5].
Histopathologically, focal eosinophilic infiltration in the periportal space,
eosinophilic abscess, and eosinophilic granuloma are placed in the category of
focal eosinophilic liver disease
[4–10].
Several reports regarding findings of this condition on imaging techniques
such as CT and sonography have been published
[1,
11], and the main histologic
findings of previous studies were focal eosinophilic infiltration in the
periportal space both with and without necrotic foci
[1,
11–13].
Several reports have described the occurrence of focal eosinophilic liver
infiltration in patients with gastric malignancy
[2–4];
recently, the coexistence of focal eosinophilic liver infiltration with
hepatocellular carcinoma within the same or a different hepatic segment has
also been reported [5].
Although relatively specific radiologic findings, as depicted on sonography
and CT, of focal eosinophilic infiltration in liver have been described
[1,
11–13],
it is sometimes still difficult to differentiate these foci from liver tumors,
especially in patients with known primary cancer.
Liver MRI has become a widely used technique and has been accepted as the
most accurate imaging technique for the detection and characterization of
focal liver lesions. To our knowledge, MRI findings of eosinophilic liver
disease are limited to a small number of case reports
[5]. We are not aware of any
studies that have correlated the imaging findings with the two different
histologic features of eosinophilic abscess and focal eosinophilic
infiltration.
Accordingly, in this study, we analyzed the MRI findings of the patients
with pathologically proven eosinophilic abscess and focal eosinophilic
infiltration to identify the distinguishing features from those of other focal
liver lesions. In addition, the imaging findings of focal eosinophilic
infiltration and eosinophilic abscess were correlated with their histologic
findings.
Materials and Methods
Patients
Between November 1999 and December 2003, 19 consecutive patients with
peripheral blood eosinophilia who were thought to have focal liver lesions on
the basis of previously performed sonography or dynamic helical CT underwent
liver MRI. Written informed consent was obtained from each patient before
being entered into the study, and the study was approved by the institutional
review board of our hospital. We excluded seven of these patients because of
the absence of histologic proof. The remaining 12 patients (11 men and one
women; age range, 35–66 years; mean age, 50 years) with pathologically
confirmed focal eosinophilic liver diseases were enrolled as our study
population: 11 eosinophilic abscesses in six patients and 27 focal
eosinophilic infiltrations in six patients.
These patients had flulike signs or abdominal pain and discomfort in the
right upper quadrant. None had a history of allergies or of drug use. None of
these patients had positive results on serology or stool examination for
parasites, including Taenia solium (cysticercosis), Paragonimus
westermani (paragonimiasis), Clonorchis sinensis
(clonorchiasis), Sparganum mansoni (sparganosis), Anisakis
simplex (anisakiasis), Fasciola hepatica (fascioliasis),
Schistosoma (schistosomiasis), and Toxoplasma gondii
(toxoplasmosis). One of these patients had concomitant pneumonia with typical
imaging findings of eosinophilic pneumonia. One patient fulfilled the criteria
for the diagnosis of hypereosinophilic syndrome. Four patients with
eosinophilic abscesses had a single nodular hepatocellular carcinoma with
underlying chronic hepatitis B virus infection or viral cirrhosis, and two
patients with focal eosinophilic infiltrations had gastric cancer in one case
and colon cancer in the other. Most of these patients had mild hepatic
dysfunction with mildly elevated serum levels of aspartate aminotransferase
(range, 34–182 U/L; mean, 73 U/L), alanine aminotransferase (range,
36–161 U/L; mean, 80 U/L), or alkaline phosphatase (range, 227–419
U/L; mean, 269 U/L). In four patients with hepatocellular carcinoma, an
elevated serum 
-fetoprotein level (
250 ng/mL) was found.
Peripheral blood eosinophilia (range, 8.7–55.1%; mean, 27%) was found in
all patients.
Lesion Confirmation
Histopathologic confirmation was obtained in eight patients by
sonographically guided percutaneous core needle biopsy using 19-gauge
automated biopsy guns. Hepatic surgery was performed in four patients who had
hepatocellular carcinoma. Sonographically guided biopsy was performed for only
one or two liver lesions in patients with multiple lesions because the imaging
findings of multiple lesions were identical in each patient. In four patients
with hepatocellular carcinomas and eosinophilic abscesses, segmentectomy
(n = 3) or enucleation (n = 1) was performed on the basis of
preoperative image analysis and intraoperative sonography. During hepatic
surgery, five eosinophilic abscesses (which had been considered on MRI to be
daughter nodules of hepatocellular carcinomas) and four hepatocellular
carcinomas were found on intraoperative sonography (Fig.
1A,
1B,
1C,
1D,
1E,
1F,
1G,
1H). In the eight patients who
underwent sonographically guided percutaneous biopsy, histopathologic
diagnoses included eosinophilic infiltration in the periportal area and the
hepatic lobule in six patients (Fig.
2A,
2B,
2C,
2D,
2E) and eosinophilic abscess
in two patients.
View larger version (131K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A. —48-year-old man with nodular hepatocellular carcinoma and
eosinophilic abscess in left lateral segment of liver. Respiratory-triggered
T2-weighted turbo spin-echo image shows 3-cm-diameter hyperintense nodule
(arrow) in left lateral segment of liver.
|
|
View larger version (131K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B. —48-year-old man with nodular hepatocellular carcinoma and
eosinophilic abscess in left lateral segment of liver. Respiratory-triggered
T2-weighted turbo spin-echo image obtained at slightly higher level than
A shows another small, slightly hyperintense nodule (arrow)
considered to be daughter nodule.
|
|
View larger version (109K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1C. —48-year-old man with nodular hepatocellular carcinoma and
eosinophilic abscess in left lateral segment of liver. On dynamic arterial
phase image obtained after administration of gadobenate dimeglumine, nodule
(arrow) in hepatic segment II shows strong nodular enhancement.
Nodule was confirmed as hepatocellular carcinoma.
|
|
View larger version (105K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1D. —48-year-old man with nodular hepatocellular carcinoma and
eosinophilic abscess in left lateral segment of liver. On dynamic arterial
phase image, second nodule (arrow) (seen on B) also shows
nodular enhancement.
|
|
View larger version (108K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1E. —48-year-old man with nodular hepatocellular carcinoma and
eosinophilic abscess in left lateral segment of liver. On equilibrium phase
image obtained 3 min after contrast injection, main mass (arrow)
shows low signal intensity with capsular enhancement.
|
|
View larger version (101K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1F. —48-year-old man with nodular hepatocellular carcinoma and
eosinophilic abscess in left lateral segment of liver. On equilibrium phase
image, second nodule (seen on B and D) becomes isointense
compared with surrounding liver parenchyma.
|
|
View larger version (98K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1G. —48-year-old man with nodular hepatocellular carcinoma and
eosinophilic abscess in left lateral segment of liver. Photograph of gross
specimen shows both nodular hepatocellular carcinoma (large arrow)
and eosinophilic abscess (small arrow) in same hepatic segment.
|
|
View larger version (176K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1H. —48-year-old man with nodular hepatocellular carcinoma and
eosinophilic abscess in left lateral segment of liver. Photomicrograph of
specimen (indicated by small arrow in G) obtained at surgery
shows that normal hepatocytes are replaced by numerous inflammatory cell
infiltrates predominantly composed of eosinophils (arrows). (H and E,
x150)
|
|
View larger version (111K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A. —55-year-old man with multifocal eosinophilic infiltrations in
liver. T1-weighted gradient-echo image shows focal faintly low signal
intensities in subcapsular area (arrow) of right hepatic lobe.
|
|
View larger version (137K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B. —55-year-old man with multifocal eosinophilic infiltrations in
liver. Respiratory-triggered T2-weighted turbo spin-echo image shows poorly
defined faint hyperintensity (arrow) at same location as in
A.
|
|
View larger version (107K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2C. —55-year-old man with multifocal eosinophilic infiltrations in
liver. Gadolinium-enhanced dynamic portal phase image shows irregularly low
signal intensities (arrow) at same location as in A and
B.
|
|
View larger version (127K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2D. —55-year-old man with multifocal eosinophilic infiltrations in
liver. Superparamagnetic iron oxide–enhanced breath-hold T2*-weighted
fast image obtained with steady-state procession shows poorly defined focal
hyperintense lesion (arrow). Conspicuity of lesion is remarkably
improved compared with that seen on unenhanced images (A and
B).
|
|
View larger version (165K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2E. —55-year-old man with multifocal eosinophilic infiltrations in
liver. Photomicrograph of needle biopsy specimen shows numerous inflammatory
cell infiltrates predominantly composed of eosinophils (small arrows)
in periportal area. Large arrows indicate portal triad. (H and E,
x100)
|
|
The total number of lesions was determined by consensus review of all
imaging techniques including sonography, CT, intraoperative sonography, and
MRI by two experienced abdominal radiologists who were not involved in the
image analysis.
MRI Examination
All MRI was performed with a 1.5-T superconducting imager (Magnetom
Symphony, Siemens) with a phased-array multicoil for signal reception. The
liver was imaged in the axial planes in all of the imaging sequences. Baseline
MR images were obtained using a respiratory-triggered T2-weighted turbo
spin-echo, a breath-hold T2*-weighted fast imaging with steady-state
precession (FISP), and a breath-hold T1-weighted fast low-angle shot (FLASH)
sequence. Respiratory-triggered T2-weighted turbo spin-echo images were
obtained using a TR/TE of 3,300–5,500/85, echo-train length of 5, matrix
of 256 x 512, and signal average of 2. Breath-hold T2*-weighted FISP
images were obtained using a TR/TE of 180/12, flip angle of 30°, matrix of
96 x 256, and signal average of 1. Breath-hold T1-weighted FLASH images
were obtained using a TR/TE of 120/4, flip angle of 70°, matrix of 120
x 256, and signal average of 1. For all sequences, a 7-mm slice
thickness was used with a 10% intersection gap and a field of view of
35–40 cm, depending on the size of the liver.
Dynamic imaging was performed with IV administration of gadopentetate
dimeglumine (Magnevist, Schering) at a dose of 0.1 mmol per kilogram of body
weight in five patients and gadobenate dimeglumine (MultiHance, Bracco
Diagnostics) at a dose of 0.05 mmol per kilogram of body weight in seven
patients. Rapid bolus injection of contrast agent at a rate of 2 mL/sec was
followed by a 20-mL saline flush. In gadolinium-enhanced dynamic imaging
sequences, images were acquired before contrast administration and
during the arterial phase (20–35 sec after contrast administration), the
portal phase (45–60 sec after contrast administration), and the
equilibrium phase (3 min after contrast administration). Gadopentetate
dimeglumine–enhanced dynamic images were obtained using a 2D breath-hold
T1-weighted FLASH sequence, and gadobenate dimeglumine–enhanced dynamic
images were obtained using 3D Fourier transform gradient-echo imaging (VIBE
[volumetric interpolated breath-hold examination], Siemens) using a TR/TE of
3.4/1.5, a flip angle of 12°, a bandwidth of 490 Hz/pixel, a matrix 256
(interpreted) x 120 (phase) x 64–72 (partition), an
effective slice thickness of 2.3 mm, and a field of view of 32–35 cm.
VIBE images were acquired during breath-hold, and sampling was done by 60% in
the direction of z and by 82% in the direction of phase encoding
using volumetric interpolation. Image reconstruction with a 6-mm thickness was
performed with source images at an MRI workstation.
After completion of the dynamic MRI examination, superparamagnetic iron
oxide (SPIO)–enhanced MRI was performed after a 24-hr interval.
SPIO-enhanced imaging comprised the respiratory-triggered T2-weighted turbo
spin-echo sequence and the breath-hold T2*-weighted FISP sequence, using the
same parameters as those used in baseline MRI. The SPIO agent (Feridex;
Advanced Magnetics), at a dose of 15 µmol of iron per kilogram of body
weight, was diluted in 100 mL of 5% dextrose solution and injected IV through
a specific 5-µm filter for 30 min; imaging was initiated approximately 70
min (range, 50–90 min) after the IV infusion of SPIO.
Imaging Analysis
All MR images, including the unenhanced T1- and T2-weighted images,
gadolinium-enhanced dynamic images, and SPIO-enhanced T2-weighted images, were
jointly evaluated by two gastrointestinal radiologists experienced in
interpreting liver MR images in their daily clinical practice for at least 3
years. The reviewers were unaware of the clinical history, laboratory
findings, and histologic diagnoses of the patients when they interpreted the
MR images. Although there were minimal discrepancies between the two reviewers
in the interpretations of the imaging findings of lesions, agreement was
easily reached by a consensus interpretation. The signal intensity of the
lesion relative to that of the surrounding liver parenchyma on each image was
recorded as one of three categories: hypointense, isointense or invisible, or
hyperintense. If the signal intensity of the lesion was heterogeneous, the
signal intensity of the major portion of the lesion was recorded. In addition,
the margin (poorly defined or well defined), shape (nodular, irregular, or
bizarre), and distribution (subcapsular or nonsubcapsular) of the lesions were
analyzed. The subcapsular distribution was regarded as the area within 1.5 cm
of the liver capsule. We also compared the MRI findings of two histologic
types to identify the differential features between them.
Results
The total number of lesions was 38, including 27 focal eosinophilic
infiltrations and 11 eosinophilic liver abscesses. The numbers of lesions for
each patient were as follows: a single lesion in three patients, two to five
lesions in eight patients, and 10 lesions in one patient. The mean diameter of
the lesion size was 1.7 cm (range, 0.7–2.6 cm). The mean diameters of
focal eosinophilic infiltrations and eosinophilic abscesses were 1.8 cm and
1.0 cm, respectively. The distribution of focal eosinophilic infiltrations was
subcapsular in 19 (70.4%) (Figs.
2A,
2B,
2C,
2D,
2E and
3A,
3B,
3C,
3D,
3E) and nonsubcapsular in
eight (29.6%) infiltrations. The distribution of most of the eosinophilic
liver abscesses was nonsubcapsular (n = 8, 72.7%) except in three
patients with eosinophilic abscesses in the subcapsular region.
View larger version (129K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B. —50-year-old man with multifocal eosinophilic infiltrations in
liver. Respiratory-triggered T2-weighted turbo spin-echo image shows poorly
defined, irregularly shaped, faintly high signal intensities (arrows)
in subcapsular region.
|
|
On unenhanced MR images, lesions were predominantly isointense on the
T1-weighted images and hyperintense on the T2-weighted images. Focal
eosinophilic infiltrations were shown as faint low-signal-intensity areas
(n = 8/27, 29.6%) on T1-weighted images and as faint
high-signal-intensity areas (n = 19/27, 70.4%) on T2-weighted images
(Figs. 2A,
2B,
2C,
2D,
2E and
3A,
3B,
3C,
3D,
3E), and the others were not
depicted. Of the 11 eosinophilic abscesses, seven (63.6%) were not depicted
and four (36.4%) were shown as slightly low-signal-intensity areas on
T1-weighted images; all 11 (100%) were shown as areas of slightly high signal
intensity on T2-weighted images (Figs.
1A,
1B,
1C,
1D,
1E,
1F,
1G,
1H and
4A,
4B,
4C,
4D). Signal changes in
eosinophilic liver diseases, including eosinophilic abscess or infiltration,
on unenhanced images were typically somewhat subtle despite their relatively
large size (Fig. 3A,
3B,
3C,
3D,
3E). The margin of most
(26/27, 96.3%) of the focal eosinophilic infiltrations appeared poorly defined
on all sequences (Figs. 2A,
2B,
2C,
2D,
2E and
3A,
3B,
3C,
3D,
3E), but eight (72.7%) of the
11 eosinophilic abscesses showed well-defined margins (Figs.
1A,
1B,
1C,
1D,
1E,
1F,
1G,
1H and
4A,
4B,
4C,
4D). The shapes of the focal
eosinophilic infiltrations were irregular or bizarre in 21 lesions (21/27,
77.8%) and nodular in six lesions (6/27, 22.2%); the shapes of all
eosinophilic abscesses were nodular (Figs.
1A,
1B,
1C,
1D,
1E,
1F,
1G,
1H and
4A,
4B,
4C,
4D).
View larger version (131K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A. —44-year-old man with hepatocellular carcinoma (not shown) and
single eosinophilic abscess. Respiratory-triggered T2-weighted turbo spin-echo
image shows nodular slightly high-signal-intensity lesion (arrow) in
right lobe.
|
|
View larger version (110K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4C. —44-year-old man with hepatocellular carcinoma (not shown) and
single eosinophilic abscess. On dynamic arterial phase image obtained after
administration of gadobenate dimeglumine, lesion shows bright nodular
enhancement with well-defined margin (arrow).
|
|
View larger version (100K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4D. —44-year-old man with hepatocellular carcinoma (not shown) and
single eosinophilic abscess. On dynamic equilibrium phase image, lesion is
shown as isointense compared with surrounding liver parenchyma.
|
|
On gadolinium-enhanced dynamic three-phase images, most focal eosinophilic
infiltrations (25/27, 92.6%) were shown as having poorly defined, irregularly
low signal intensity on the portal phase images and as isointense or nearly
isointense on the arterial and equilibrium phase images (Figs.
2A,
2B,
2C,
2D,
2E and
3A,
3B,
3C,
3D,
3E). The remaining two focal
eosinophilic infiltrations were not clearly depicted on dynamic images. Seven
eosinophilic abscesses (7/11, 63.6%) in five patients showed nodular
enhancement during the dynamic phases, and exhibited high signal intensity
during the arterial phase (n = 7), isointensity (n = 5) or
slightly low signal intensity (n = 2) during the portal phase, and
isointensity (n = 7) during the equilibrium phase (Figs.
1A,
1B,
1C,
1D,
1E,
1F,
1G,
1H and
4A,
4B,
4C,
4D). Four eosinophilic
abscesses (36.4%) in one patient showed isointensity during the dynamic
phases.
On SPIO-enhanced T2-weighted images, all lesions were clearly depicted as
having focal high signal intensity (Fig.
2A,
2B,
2C,
2D,
2E), whereas the shapes and
margins of the lesions coincided with those seen on the unenhanced or dynamic
images.
Discussion
Recently, a few reports have shown the relatively high rate of coexistence
of focal eosinophilic liver disease and malignant diseases
[2–5].
The coexistence of focal eosinophilic liver diseases with primary malignancy,
such as stomach cancer and hepatocellular carcinoma, makes it difficult to
differentiate focal eosinophilic lesions from metastases or daughter nodules
in patients with hepatocellular carcinoma on CT and sonography, even when they
exist together with peripheral blood eosinophilia. Indeed, in our study, six
patients had the following types of cancer: gastric cancer (n = 1),
colon cancer (n = 1), or hepatocellular carcinoma (n = 4)
and coexisting focal eosinophilic liver diseases.
In this study, we retrospectively analyzed the MRI findings of focal
eosinophilic infiltration and eosinophilic abscess to find characteristic MRI
features that are helpful in differentiating them from the features of focal
liver tumors. On gadolinium-enhanced dynamic images in our study, two kinds of
imaging findings were shown: the first was the poorly defined low signal
intensity during the dynamic portal phase, seen as a main imaging finding of
focal eosinophilic infiltration, and the second was the focal nodular
enhancement, especially during the arterial phase, seen as a main imaging
finding of eosinophilic abscess. In our study, 25 lesions (92.6%) of focal
eosinophilic infiltration were shown as poorly defined low signal intensities
on dynamic portal phase images with iso- or near isointensity on arterial and
delayed phase images, findings that are in agreement with previous CT
findings. Depending on the equilibrium of the dynamic MR images, it may be
difficult to differentiate eosinophilic abscess from hypervascular hepatic
tumors such as hepatocellular carcinoma and liver metastasis, especially in
patients with underlying liver cirrhosis or other primary malignancy. However,
although arterial nodular enhancement similar to that seen in hepatocellular
carcinoma was seen in seven eosinophilic abscesses in five patients in our
study, most (9/11, 81.8%) of the eosinophilic abscesses were isointense on
both the portal and equilibrium phases. Furthermore, none of the eosinophilic
abscesses showed a prominent washout of contrast material in the lesions with
delayed capsular enhancement, which is a typical finding in hepatocellular
carcinoma on dynamic delayed phase images. Therefore, we believe that this may
be helpful for differentiating eosinophilic abscess from hepatocellular
carcinoma. In cases of focal eosinophilic infiltration, an irregular or
bizarre shape and a poorly defined margin make it possible to differentiate it
from liver metastasis.
Unenhanced T1- and T2-weighted images were helpful in accurately
characterizing focal eosinophilic liver disease. On unenhanced T1- and
T2-weighted images, visible focal eosinophilic infiltrations and eosinophilic
abscesses exhibit low and high signal intensities, respectively. Although the
signal intensities of focal eosinophilic liver diseases on unenhanced MR
images were similar to those of other hepatic tumors (both benign and
malignant), the signal changes in most of the lesions were subtle.
Furthermore, 96.3% of the focal eosinophilic infiltrations showed poorly
defined margins, whereas 81.8% of the focal eosinophilic abscesses showed
well-defined margins. These subtle signal changes and the indistinct margins
of focal eosinophilic infiltrations were different from other malignant tumors
with sharply defined margins, and this feature may be helpful for the accurate
characterization of focal eosinophilic infiltrations as distinct from other
hepatic tumors such as metastases and hepatocellular carcinomas. However,
there are still some difficulties in differentiating between focal
eosinophilic abscesses and malignant liver tumors because of their overlapping
imaging findings, despite the subtle signal changes of eosinophilic
abscesses.
The terminology of eosinophilic liver disease is not strictly defined, and
several terms including "focal eosinophilic infiltration,"
"eosinophilic abscess," and "eosinophilic granuloma"
are based on histopathologic findings. The common histopathologic finding of
focal eosinophilic infiltration in the liver is periportal and lobular
infiltration of eosinophils with normal histologic architecture
[1,
6–8].
Some researchers have used the terms "eosinophilic liver abscess"
to refer to a lesion composed of massive eosinophils and destroyed liver
parenchyma with inflammation and "eosinophilic granuloma" to refer
to a lesion composed of central necrosis and mixed inflammatory, infiltrated,
mainly eosinophils and epithelioid histiocytes
[4,
9,
10]. Although eosinophilic
granuloma and eosinophilic abscess are considered to be caused by direct
parasitic invasion
[14–16],
several reported studies of eosinophilic liver abscess found no evidence of
direct parasitic invasion [4,
9]. Therefore, uniform use of
the term "focal eosinophilic infiltration" for description of
focal eosinophilic liver disease with no definitive cause may be
inappropriate.
In this study, we also found differences in imaging features between focal
eosinophilic infiltration and focal eosinophilic abscess. First, most focal
eosinophilic infiltrations were irregularly or bizarrely shaped with poorly
defined margins, and eosinophilic abscesses had well-defined nodular shapes.
Second, as previously mentioned, the main findings of focal eosinophilic
infiltrations were focal irregularly low signal intensity on portal phase
imaging, whereas eosinophilic abscesses showed nodular enhancement during all
the dynamic phases, especially the arterial phase. The different enhancement
patterns may be caused by the difference in degree of inflammation;
inflammation is minimal in focal eosinophilic infiltration compared with that
in eosinophilic abscess. Different shapes and margins depend on the degree of
preservation of the normal histologic architecture in focal eosinophilic
infiltration or on the amount of destroyed liver parenchyma in eosinophilic
abscess [1,
4,
11]. Finally, in terms of
lesion distribution, 70.4% of focal eosinophilic infiltrations were
subcapsular, and 72.7% eosinophilic abscesses were nonsubcapsular. Based on
our results, the term "eosinophilic liver abscess" should not be
used interchangeably with the term "focal eosinophilic
infiltration" because of different histopathologic and imaging findings.
For clear definition and classification of these conditions, further
evaluation of a large number of pathologically proven cases is warranted.
A major limitation of this study is that not all lesions had a
histologically proven diagnosis, so the correlation between the MRI findings
and the histologic composition of large numbers of lesions was not possible.
However, obtaining histologic confirmation of all the lesions would have been
difficult; acquiring specimens via imaging-guided biopsy is difficult because
of the small size of the lesions, nonvisualization of a large number of
lesions on sonography, and the fact that most of these lesions can easily be
diagnosed by clinical findings and on radiologic images except in patients
with concomitant primary malignancy. In addition, all patients enrolled in
this study had peripheral eosinophilia. Given that in clinical routines, most
patients undergoing MRI for evaluation of focal liver lesions do not have
suspected peripheral eosinophilia, the study population was slightly biased.
However, even in patients with peripheral eosinophilia, the differentiation
between focal liver malignancies and focal eosinophilic liver diseases may not
be easy if patients have malignancies in the liver or extrahepatic organs.
Furthermore, we believe that familiarity with the MRI findings of focal
eosinophilic liver diseases may encourage radiologists to consider checking
peripheral eosinophilic counts to add diagnostic confidence to their imaging
interpretations.
In summary, although MRI findings of focal eosinophilic liver disease may
be confused with those of liver malignancy, careful analysis of the shape,
margin, signal intensity, and enhancement pattern of the lesions at the
combined interpretations of unenhanced and dynamic images should make accurate
characterization possible.
References
- Kim GB, Kwon JH, Kang DS. Hypereosinophilic syndrome: imaging
findings in patients with hepatic involvement. AJR1993; 161:577
–580[Abstract/Free Full Text]
- Yoon IL. The eosinophil and gastrointestinal carcinoma.
Am J Surg 1959;97:195
–200[Medline]
- Iwasaki K, Torisu M, Fusimura T. Malignant tumor and eosinophils.
Cancer 1986;58:1321
–1327[Medline]
- Soon WH, Kim HG, Park CI, Lee SI. Eosinophilic liver abscess in
patients with gastric cancer. [in Korean]. Korean J
Pathol 1993;27:27
–33
- Won JH, Kim MJ, Kim BM, et al. Focal eosinophilic infiltration of
the liver: a mimick of hepatic metastasis. Abdom
Imaging 1999;24:369
–372[Medline]
- Fauci AS, Harley JB, Robert WC, Ferrans VJ, Gralnick HR, Bjornson
BH. The idiopathic hypereosinophilic syndrome: clinical, pathophysiologic, and
therapeutic considerations. Ann Intern Med1982; 97:78
–92
- Chusid MJ, Dale DC, West BC, Wolff SM. The hypereosinophilic
syndrome: analysis of fourteen cases with review of the literature.
Medicine 1975;54:1
–27[Medline]
- Foong A, Scholes JV, Gleich GJ, Kephart GM, Holt PR.
Eosinophil-induced chronic active hepatitis in the idiopathic
hypereosinophilic syndrome. Hepatology1991; 13:1090
–1094[Medline]
- Shanti CM, Lucas CE, Tyburski JG, Soulen RL, Lucas DR. Eosinophilic
abscess and eosinophilic pseudotumor presenting as bile duct masses: a report
of 2 cases. Surgery2001; 130:104
–108[Medline]
- Kaplan KJ, Goodman ZD, Ishak KG. Eosinophilic granuloma of the
liver: a characteristic lesion with relationship to visceral larva migrans.
Am J Surg Pathol2001; 25:1316
–1321[Medline]
- Lee WJ, Lim HK, Lim JH, Kim SH, Choi SH, Lee SJ. Foci of
eosinophil-related necrosis in the liver: imaging findings and correlation
with eosinophilia. AJR1999; 172:1255
–1261[Abstract/Free Full Text]
- White WL, Wahner HW, Brown ML, James EM. Sequential liver imaging
in the hypereosinophilic syndrome: discordant images with scintigraphy,
ultrasound, and computed tomography. Clin Nucl Med1981; 6:75
–77[Medline]
- Yoo SY, Han JK, Kim YH, Kim TK, Choi BI, Han MC. Focal eosinophilic
infiltration in the liver: radiologic findings and clinical course.
Abdom Imaging2003; 28:326
–332[Medline]
- Farid Z, Kilpatrick ME, Chiodini PL. Parasitic diseases of the
liver. In: Schiff L, Schiff ER, eds. Diseases of the
liver, 6th ed. Philadelphia, PA: Lippincott, 1993:1338
–1355
- Han JK, Choi BI, Cho JM, et al. Radiologic findings of human
fascioliasis. Abdom Imaging1993; 18:261
–263[Medline]
- Ishibashi H, Shimamura R, Hirata Y, Kudo J, Onizuka H. Hepatic
granuloma in toxocaral infection: role of ultrasonography in
hypereosinophilia. J Clin Ultrasound1992; 20:204
–210[Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
