DOI:10.2214/AJR.05.1416
AJR 2006; 187:W622-W629
© American Roentgen Ray Society
Hepatic Visceral Larva Migrans of Toxocara canis: CT and Sonographic Findings
Samuel Chang1,
Jae Hoon Lim1,
Dongil Choi1,
Cheol Keun Park2,
Nam-Hee Kwon3,
Seung-Yull Cho4 and
Dong-Chull Choi3
1 Department of Radiology and Center for Imaging Science, Samsung Medical
Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Kangnam-ku,
Seoul, South Korea 135-230.
2 Department of Pathology, Samsung Medical Center, Sungkyunkwan University
School of Medicine, Seoul, South Korea 135-230.
3 Department of Medicine, Samsung Medical Center, Sungkyunkwan University School
of Medicine, Seoul, South Korea 135-230.
4 Department of Parasitology, Samsung Medical Center, Sungkyunkwan University
School of Medicine, Seoul, South Korea 135-230.
Received August 14, 2005;
accepted after revision November 3, 2005.
Address correspondence to D.-C. Choi
(dcchoi{at}smc.samsung.co.kr).
WEB
This is a Web exclusive article.
Abstract
OBJECTIVE. The aim of this study was to describe the CT and
sonographic findings of hepatic visceral larva migrans of Toxocara
canis.
MATERIALS AND METHODS. Fifty-four patients (44 men, 10 women; age
range, 30-80 years; mean age, 53 years) with serologically confirmed visceral
larva migrans of Toxocara canis underwent evaluation of the liver
with CT (n = 25), sonography (n = 48), or both. Two
radiologists used consensus for retrospective evaluation of CT and sonographic
findings. Correlation between the presence and severity of hepatic
abnormalities on images and the degree of peripheral eosinophilia was
assessed.
RESULTS. Seventeen (68%) of 25 patients who underwent CT had single
or multiple ill-defined, oval or elongated, small, low-attenuating lesions in
the liver. Eighteen (38%) of 48 patients who underwent sonography had single
or multiple small, poorly defined, oval or elongated, hypoechoic scattered
focal lesions in the liver. In the 19 patients who underwent both CT and
sonography, the two techniques had no significant difference in rate of
detection of hepatic lesions (p = 0.375, McNemar test). The lesion
numbers on CT and sonography showed excellent linear correlation (r =
0.844, p = 0.001) by Pearson's correlation test. An independent
samples t test showed that eosinophil count and percentage in the
peripheral blood were significantly higher in patients with hepatic lesions on
CT and sonography than in patients without lesions.
CONCLUSION. CT and sonographic findings of hepatic visceral larva
migrans of T. canis are multiple, ill-defined, oval or elongated,
small, nodular lesions scattered in the liver parenchyma. The presence of
hepatic lesions on images was associated with higher peripheral eosinophil
count and percentage.
Keywords: CT • infectious diseases • liver • sonography
Introduction
The term visceral larva migrans (VLM) is used to describe the migration of
second-stage larvae of nematodes through the tissue of human viscera
[1,
2]. Toxocara canis is
the most common etiologic agent, although Toxocara cati, Baylisascaris
procyonis, Capillaria hepatica
[3], Ascaris suum
[4] and some
Ancylostoma species
[5] have been reported to cause
the disease. Humans are infected by T. canis by ingestion of
infective eggs from soil or of encapsulated larvae in the uncooked tissues of
a paratenic animal host [5].
The larvae are liberated in the intestine, burrow into the intestinal wall,
enter the portal flow, and reach the liver, where they may stay or from where
they may continue to be distributed to other tissues
[6]. In humans, the larvae are
deposited in the liver, lungs, eye, heart, and brain. The living or dead
larvae form abscesses or granulomas, which can cause serious complications
such as hepatomegaly, endophthalmitis, and neurologic disturbances
[5]. A specific diagnosis is
made when the larvae are identified in tissue sections or with serologic
testing [7,
8]. The role of imaging in the
diagnosis of hepatic VLM of T. canis has not been well
established.
VLM appears to have a worldwide distribution and has been reported in the
United States, Great Britain, continental Europe, Egypt, West Africa, South
Africa, Iran, Australia, India, China, Taiwan, Japan, and Brazil
[6,
8-12].
The rate of seropositive results is 5-40% with enzyme-linked immunosorbent
assay (ELISA) with T. canis excretory/secretory antigen from the
general population or patient groups
[13-15].
There have been series and case reports of VLM
[3,
4,
6,
8,
10-12,
16-18],
and the radiologic findings have been described briefly in several case
reports [4,
6,
10,
12,
19-21].
To our knowledge, there has been no large-series study of the imaging features
of hepatic VLM of T. canis. The purpose of this study was to describe
the CT and sonographic findings of a systematic retrospective analysis of the
cases of a relatively large number of patients with serologically proven
hepatic VLM of T. canis.
View larger version (82K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A —44-year-old man with hepatic visceral larva migrans of
Toxocara canis. Transverse contrast-enhanced CT scan obtained at
arterial phase shows faint, rim-enhancing lesion (arrow) in Couinaud
segment VII of liver.
|
|
View larger version (83K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —44-year-old man with hepatic visceral larva migrans of
Toxocara canis. CT scan obtained at portal venous phase shows
multiple small, ill-defined, oval or elongated, low-attenuating lesions
scattered throughout liver.
|
|
View larger version (108K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1D —44-year-old man with hepatic visceral larva migrans of
Toxocara canis. Oblique subcostal sonogram of right lobe of liver
shows multiple, small, oval or elongated, hypoechoic lesions (arrows)
with indistinct margins.
|
|
Materials and Methods
Patient Selection
Our institutional review board did not require its approval or patients'
informed consent for this retrospective study. During the period October 2001
to January 2005, 103 patients came to the outpatient department of medicine of
our institution with hypereosinophilia in peripheral blood (> 500
cells/µL) and were enrolled in the hypereosinophilia registry. Thirty-three
patients were excluded because of following diagnoses: Paragonimus
westermani and Clonorchis sinensis infection according to
positive results of parasite ELISA or stool examinations in 18 cases,
bronchial asthma in two cases, and allergic rhinitis according to positive
results of methacholine provocation tests and mild eosinophilia (< 1,500
cells/µL) in one case, drug reaction according to positive history of drug
ingestion and spontaneous resolution after cessation of drug ingestion in one
case, other etiologic factors in four cases, and idiopathic factors in seven
cases.
The diagnosis of T. canis infection was made in 70 cases and was
based on ELISA results. The medical records and laboratory data were reviewed,
and demographic and clinical data were collected by a physician. Twenty-eight
(40%) of the 70 patients reported symptoms such as cough (n = 10),
itching (n = 4), weakness (n = 2), myalgia (n = 3),
and fever (n = 1), but there were no symptoms or signs suggesting the
possibility of a specific disease. Forty-two (60%) of the 70 patients had a
history of ingestion of uncooked bovine liver. Six (9%) of the 70 patients had
underlying malignant disease: stomach cancer (n = 3), renal cell
carcinoma (n = 1), malignant melanoma (n = 1), and prostate
cancer (n = 1). All of these patients were tumor free when they
underwent CT or sonography because of T. canis infection. No patient
had hepatic cirrhosis. Among the 70 patients, 19 underwent both CT and
sonography, six underwent CT only, and 29 underwent sonography only. Among the
19 patients who underwent both CT and sonography, the median time interval
between CT and sonography was 12 days (mean, 25.7 days; SD, 34.1 days; range,
0-120 days). Sixteen patients who did not undergo imaging studies were
excluded. Thus 54 patients (44 men, 10 women; age range, 30-80 years; mean
age, 53 years) formed the basis of the study. Blood tests and imaging studies,
primarily CT when both CT and sonography were available, were performed on the
same day for 30 patients and within 66 days for 24 patients (mean, 5.2 days;
SD, 12.5 days). The eosinophil count and percentage of eosinophils in the
peripheral blood at imaging ranged from 0 to 23,540/µL (mean, 2,870/µL;
SD, 4,466/µL) and 0 to 88.0% (mean, 21.9%; SD, 18.5%). Sonographically
guided liver biopsy performed on seven patients disclosed focal eosinophilic
infiltration (n = 1) or eosinophilic abscess (n = 6) in the
liver parenchyma. In the department of medicine of our institution, the
indications for therapy for hepatic VLM of T. canis are hepatic
dysfunction or infection of another major organ. Twelve patients were treated
with a steroid only (n = 6), albendazole combined with prednisolone
(n = 4), or praziquantel (n = 2).
ELISA
An ELISA kit (Bordier Affinity Products) was used for the diagnosis of
human toxocariasis. This assay is used to detect human IgG antibodies to
Toxocara excretory/secretory antigens. This kit has been reported to
have a sensitivity of 91% and a specificity of 86%
[22]. The titers for positive
results varied according to the daily reference control. Although some cross
reaction occurs in other types of human helminthiasis, such as trichinosis,
fascioliasis, and strongyloidosis, the titers in these types of helminthiasis
are lower than those in positive control serum of patients with T.
canis infection [22].
Imaging Technique
CT—The CT techniques were not standardized because of the
relatively long span of the study and the rapid development of CT scanners. CT
was performed with one of four helical CT scanners (HiSpeed, LightSpeed QX/i,
LightSpeed Ultra, or LightSpeed 16; GE Healthcare). Unenhanced scans were
obtained with 5- to 7.5-mm reconstruction intervals and slice thickness.
Contrast-enhanced CT scans were obtained after injection of 120 mL of nonionic
iodinated contrast material (iopamidol, Iopamiro 300, Bracco) at a rate of 3-4
mL/s through an antecubital vein with 2.5- to 7.5-mm reconstruction intervals
and slice thickness. Arterial, portal venous, and equilibrium phase scans were
obtained 30, 70, and 180 seconds, respectively, after the start of injection
of contrast material. Four patients underwent unenhanced, arterial, portal
venous, and equilibrium phase scanning; seven patients underwent arterial,
portal venous, and equilibrium phase scanning; two patients underwent
unenhanced, arterial, and portal venous phase scanning; and 12 patients
underwent routine abdominal CT in the portal venous phase only. Thus all
patients underwent portal venous phase imaging.
Sonography—Routine transabdominal sonographic examination of
the upper abdomen was performed with the patient in the supine or right
anterior oblique position. Because the study was retrospective, many abdominal
radiologists with various lengths of experience were involved in sonographic
scanning. All sonographic examinations were performed with commercially
available high-end sonography units (Acuson XP 10, Acuson; HDI 3000, 3500,
Advanced Technology Laboratories; Logic 700, GE Healthcare) and 2- to 5-MHz
probes. No patient underwent color Doppler or contrast-enhanced sonographic
examination.
Image Analysis
All CT and sonographic images were reviewed retrospectively by consensus of
two board-certified radiologists. Portal venous phase CT images were evaluated
for number of lesions, predominant shape (round or elongated) of lesions,
distinctiveness of margins, attenuation, rim enhancement, maximum diameter of
largest lesion, central or peripheral location, periportal distribution, and
traversing vessel. In the cases in which they were available, arterial and
equilibrium phase images were evaluated for number of lesions, lesion
conspicuity, distinctiveness of margins, and enhancement. When lesions in the
same patient had different imaging findings, the predominant pattern was
recorded. Lower lung fields included in CT of the liver or abdomen were
reviewed for focal pulmonary infiltration. CT images were reviewed in the lung
window setting on monitors of a PACS workstation (Centricity 2.0, GE
Healthcare). Sonography reports were reviewed for lesion number and maximum
diameter of largest lesion. Sonographic images were reviewed for shape,
distinctiveness of margins, and echogenicity.
Statistical Analysis
In the case of patients who underwent both CT and sonography, the rates of
detection of hepatic lesions with the two techniques were compared by McNemar
test and for number of lesions by paired t test and Pearson's
correlation test. To assess whether there was a difference in the degree of
peripheral eosinophilia between the group with and the group without hepatic
lesions, the means of eosinophil count and percentage in the peripheral blood
were compared by independent samples t test. To assess whether there
was correlation between degree of peripheral eosinophilia and severity of
hepatic abnormality on CT or sonography, bivariate analysis between eosinophil
count and percentage and number of hepatic lesions on CT or sonography was
performed with Pearson's correlation test. Statistical analysis was performed
with a statistical software package (SPSS for Windows, version 13.0, SPSS). A
statistically significant difference was p < 0.05.
Results
Imaging Characteristics of Hepatic Lesions
Seventeen (68%) of the 25 patients with CT scans obtained at the portal
venous phase had single (n = 2) or multiple (n = 15)
ill-defined, oval or elongated, low-attenuating lesions in the liver
parenchyma (Figs. 1A,
1B,
1C, and
1D). The portal venous phase CT
characteristics of hepatic lesions of VLM are summarized in
Table 1. The median number of
hepatic lesions was seven (mean ± SD, 18.1 ± 23.8; range, 1-93).
The lesions were less than 2 cm in diameter in 12 patients and 2-4 cm in four
patients. One patient had multiple ill-defined, low-attenuating lesions larger
than 5 cm at the periphery of the liver, some of which were crossed by portal
vein branches (Fig. 2). Rim
enhancement was seen at the portal venous phase in seven patients. The lesions
tended to be in the periphery of the liver (n = 10) and along the
portal vein branches (n = 10). In some cases, the portal vein
traversed the lesions (n = 6)
(Fig. 2). Abnormalities were
not seen in the other eight patients.
View larger version (88K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2 —36-year-old man with subsegmental hepatic involvement of
visceral larva migrans of Toxocara canis. Transverse
contrast-enhanced CT scan obtained at portal venous phase shows multiple,
ill-defined, low-attenuating lesions at periphery of liver. Portal veins
(arrows) traverse lesions. Percutaneous liver biopsy disclosed
eosinophilic abscess.
|
|
View larger version (91K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —68-year-old man with hepatic visceral larva migrans of
Toxocara canis. Transverse contrast-enhanced CT scan obtained at
arterial phase shows small, ill-defined, enhancing nodule (long
arrow) in right hepatic lobe. Short arrow points to branch of portal
vein.
|
|
View larger version (96K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —68-year-old man with hepatic visceral larva migrans of
Toxocara canis. CT scan obtained at portal venous phase shows nodule
in A as low attenuating lesion (short arrow). Another
ill-defined lesion (long arrow) at anterolateral part of right lobe
is evident on portal venous phase image but not on arterial phase image.
|
|
View larger version (164K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3C —68-year-old man with hepatic visceral larva migrans of
Toxocara canis. Photomicrograph of needle biopsy specimen shows
eosinophilic abscess (arrowheads) and moderate to marked eosinophilic
infiltration (arrows) in hepatic sinusoids. (H and E, x100)
|
|
The lesions were best seen or only seen at the portal venous phase in six
(75%) of the eight patients who underwent unenhanced dynamic arterial, portal
venous, and equilibrium phase CT (Figs.
1A,
1B,
1C, and
1D). Even when seen on arterial
or equilibrium phase images, the lesions were less prominent, and the number
of visualized lesions was small. The arterial phase images of six patients
showed only multiple low-attenuating nodules with (n = 3) or without
(n = 3) peripheral rim enhancement (Figs.
1A,
1B,
1C, and
1D). Images of one patient
showed multiple enhancing nodules only (Figs.
3A,
3B, and
3C). The other patient had both
low-attenuating nodules with peripheral rim enhancement and enhancing nodules.
Images at the portal venous phase showed multiple low-attenuating nodules with
(n = 2) or without (n = 7) peripheral rim enhancement in all
eight patients, one of whom had lesions with both characteristics. Images at
equilibrium phase showed no visible lesions in four patients. In three of the
other four cases images showed ill-defined, slightly high-attenuating nodules;
in the other case images showed a few ill-defined, low-attenuating lesions.
Unenhanced CT images depicted only a few of the lesions that were visible on
contrast-enhanced CT scans. CT scans of seven (28%) of 25 patients showed one
or several small areas of round nodular infiltration with peripheral
ground-glass attenuation at the bases of the lungs. The pulmonary lesions
measured 1.0-1.2 cm in diameter.
Images of 18 (38%) of 48 patients who underwent sonography showed single or
multiple poorly defined, oval or elongated, hypoechoic, small scattered focal
lesions in the liver (Figs. 1A,
1B,
1C, and
1D). The sonographic
characteristics of hepatic lesions of VLM are summarized in
Table 2. Thirteen (72%) of the
lesions were less than 2 cm in diameter. In the patient who had multiple
large, low-attenuating lesions at the periphery of the liver on CT images, the
lesions appeared as multiple large hypoechoic lesions on sonography. Thirty
patients had no focal hepatic lesion detected on sonography.
Biopsy-proven eosinophilic infiltration (n = 1) and abscess
(n = 6) were seen as small, ill-defined, low-attenuating nodules at
the portal venous phase of CT and as small, ill-defined, hypoechoic nodules on
sonography and could not be differentiated.
Images of 11 (58%) of the 19 patients who underwent both CT and sonography
showed hepatic lesions with both techniques. The lesion numbers on CT and
sonography in the 11 patients were 17.2 ± 27.6 and 7.4 ± 5.7,
respectively, which showed excellent linear correlation (r = 0.844,
p = 0.001) by Pearson's correlation test and no statistical
difference (p = 0.183) by paired t test. Images of three
(16%) of the patients showed discrepancy between the CT and sonographic
results. CT showed 43 and 18 hepatic lesions only in two patients, and
sonography showed one hepatic lesion in one patient. The time intervals
between CT and sonography for these three patients were 12, 4, and 23 days,
respectively, and the patients were not treated pharmacologically between the
two examinations. In five cases no imaging technique showed hepatic lesions.
The two techniques had no statistically significant difference in detection
rate of hepatic lesions (p = 0.375, McNemar test).
View larger version (89K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A —42-year-old woman with hepatic visceral larva migrans of
Toxocara canis. Transverse contrast-enhanced CT scan obtained at
portal venous phase shows multiple small, ill-defined, oval or elongated,
low-attenuating nodules throughout liver.
|
|
View larger version (87K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B —42-year-old woman with hepatic visceral larva migrans of
Toxocara canis. CT scan at same level as A 4 months after
A without antihelminthic treatment shows similar but fewer
low-attenuating nodules and different loci of lesions, suggesting migration of
larvae.
|
|
Among the 17 patients with hepatic lesions on CT, follow-up CT images were
obtained for nine patients. In eight cases hepatic lesions had disappeared on
follow-up CT images obtained 324 ± 270 days (median, 213 days; range,
50-814 days) after the original images. In the other patient, the number and
position of the nodular lesions on follow-up CT were different from those on
the initial CT images (Figs. 4A
and 4B). Follow-up sonography
was performed for eight of the 18 patients with hepatic lesions detected with
sonography. Hepatic lesions had disappeared in five patients on follow-up
sonography 164 ± 42 days (median, 155 days; range, 119-224 days) after
the original examination. In the other three patients, hepatic lesions
persisted with a change in number. Among the pharmacologically treated
patients, only three patients underwent follow-up examinations. One patient
was treated with steroid only and underwent follow-up CT. The other two
patients were treated with prednisolone and albendazole. One of these patients
underwent follow-up sonography, and the other underwent both imaging
techniques. The lesions had disappeared on follow-up examinations performed,
217, 224, and 50 days after the original images were obtained.
Relation Between Hepatic Abnormality on Images and Peripheral Eosinophilia
Eosinophil count and percentage of eosinophils in the peripheral blood at
CT or sonography were considerably higher in the patient group with hepatic
lesions than in the group without hepatic lesions, a finding that was
statistically significant by independent samples t test (Tables
3 and
4). However, eosinophil count
and percentage did not correlate with the number of hepatic lesions on CT
(r = 0.375, p = 0.138; r = 0.352, p =
0.165) or sonography (r = 0.265, p = 0.288; r =
0.444, p = 0.065) by Pearson's correlation test.
Discussion
VLM of T. canis is known to be mainly a disease of children.
Children who live with dogs in poor sanitary conditions are likely to ingest
contaminated soil, hair, and the infective form of the eggs of T.
canis. In the United States, approximately 80% of puppies younger than 6
months and 20% of dogs older than 6 months are infected, and 20-25% of soil in
parks and playgrounds is contaminated with the eggs of T. canis
[23]. Therefore, there is high
risk of infection among children. In addition to egg ingestion, larvae in
animal tissues can be transferred from host to host through predation
[5], which has been proven
experimentally [24], and from
animals to humans [5]. Some
adults ingest uncooked cow or cattle liver or meat containing the infective
form of T. canis, encapsulated larvae
[5,
6]. Ishibashi et al.
[6] reported the cases of two
adults with VLM of Toxocara caused by ingestion of raw meat or bird,
boar, or horse liver. Forty-two (60%) of 70 patients in the present study had
a definite history of ingestion of a considerable quantity of raw cow liver
within 6 months of the hospital visit.
The infective larvae of T. canis, which are 0.5 mm long, move
slowly from place to place (larva migrans) in the hepatic parenchyma. In the
wake of migration, the larvae cause eosinophilic infiltration followed by
abscess or granuloma formation
[8]. The common histopathologic
finding of focal eosinophilic infiltration in the liver is periportal and
lobular infiltration of eosinophils with normal histologic architecture
[25]. Eosinophilic abscess
refers to a lesion composed of massive eosinophils and destroyed liver
parenchyma with inflammation. Eosinophilic granulomas are typically multiple
and consist of central necrosis and mixed inflammatory cell infiltrate with
numerous eosinophils, varying numbers of neutrophils and lymphocytes, and a
palisade of epithelioid histiocytes or giant cells. Remnant parasites may be
identified [8].
CT and sonographic findings reflect the pathologic findings. Pathologic
changes such as periportal eosinophilic infiltration, abscess, or granuloma
with central necrosis and peripheral edema caused by slow migration of T.
canis have been depicted as multiple small, ill-defined, oval or
elongated, low-attenuating nodules on portal venous phase images of dynamic CT
[4,
6,
10,
20]. These CT findings seem to
be nonspecific and similar to the findings of other types of granuloma or
inflammatory lesions. However, the eosinophilic lesions were best seen or only
seen in the portal venous phase. This finding seemed to be somewhat
characteristic of eosinophilic infiltration, granuloma, and abscess. In most
cases, the enhancement pattern of hepatic lesions on dynamic CT could be
explained by variable degrees of arterial hyperemia from the periphery of the
hepatic lesions during the arterial phase or later and by complete central
fill-in during the equilibrium phase. Although several cases have been
confirmed on the basis of histopathologic findings, no difference in
enhancement pattern that could explain the pathologic difference was found
between eosinophilic infiltration and abscess. The lesions were small in most
cases, but larger, low-attenuating lesions were present in which portal vein
branches traversed the center of the lesion
(Fig. 2). Because larvae of
T. canis are distributed by portal blood flow, these larger lesions
may be attributed to occlusion or alteration of the microvascular portal blood
supply [26] in addition to
eosinophilic inflammation itself. Sonography showed multiple small, focal,
hypoechoic lesions in the liver parenchyma
[4,
6,
10,
12]. The lesions were usually
oval or elongated and sometimes were angulated or trapezoid rather than round,
and the margins were ill-defined. These sonographic findings were nonspecific
and could not be used to differentiate the lesions from other types of
granuloma and inflammatory lesions.
The rates of detection of hepatic lesions of VLM of T. canis with
CT (68%, 17/25) and sonography (38%, 18/48) in our study could not be compared
simply because the patient populations in each group were partly different.
Among 19 patients who underwent both CT and sonography, the detection rates
with the two techniques were not statistically different. When both techniques
were used, the numbers of lesions seen on CT and sonography were not
different. However, lesion site and size could not be compared lesion by
lesion because of incomplete sonographic data. The two patients with multiple
small hepatic lesions recognized only on CT had homogeneous and slightly
heterogeneous parenchymal echogenicity, respectively, on retrospective review
of sonograms. The time intervals between CT and sonography were short enough
to exclude resolution of the lesions. When lesions are so numerous the lesions
touch one another, sonographers may overlook and perceive ill-defined focal
hepatic lesions as heterogeneous parenchymal echogenicity. However, no
acceptable hypothesis other than physical differences in the imaging mechanism
between CT and sonography could explain the other two cases of
discrepancy.
Eosinophil count and percentage in the peripheral blood were considerably
higher in the patient group with hepatic lesions on imaging than in the
patient group without lesions. When there was no abnormality on CT or
sonography, peripheral eosinophilia was mild. However, the degree of
eosinophilia and the number of hepatic lesions on CT or sonography were not
proven to have positive linear correlation.
Several articles have described imaging findings of eosinophilic
infiltration in the liver in association with peripheral eosinophilia
[25-32].
Some investigators
[25-27,
31] have addressed the
hypothesis that the cause of eosinophilic infiltration is related to
idiopathic hypereosinophilia or to cancer through a tumor-associated
eosinophilotactic factor [27,
29,
30,
32]. Others investigators have
offered no explanation of causes
[30,
33]. The CT and sonographic
findings in the aforementioned articles
[25,
27-32]
appear to match the imaging findings in our series. According to a recent
pathologic report, 30% (13/43) of hepatic eosinophilic granulomas, excluding
cases of Langerhans' cell histiocytosis, had remnants of Toxocara
organisms in the tissue (n = 8) and had positive serologic results
for Toxocara species (n = 5)
[8]. Although the proportion of
focal hepatic eosinophilic lesions due to toxocariasis is not known, we
believe that VLM of T. canis is the likely cause of hepatic
eosinophilic abscess or granuloma in many patients.
Patients with mild infection usually have no symptoms. Heavy infection can
cause fever, nausea, epigastric discomfort, abdominal pain, weight loss, and
ocular and neurologic symptoms
[8,
20]. Pulmonary infiltrates are
also common in severe infection
[8]. The diagnosis of VLM of
T. canis is clinically important because the patients can be treated
with specific antihelminthic agents such as albendazole
[10]. Identification of an
eosinophilic lesion in the liver on imaging studies suggests the diagnosis of
VLM as the cause of hypereosinophilia. Physicians who find such a lesion
should initiate a prompt serologic and pathologic search for causative
organisms. In patients with hypereosinophilia and hepatic parenchymal nodules
on CT and sonography, ELISA with Toxocara excretory/secretory antigen
should be performed. Imaging studies are also useful for guiding percutaneous
biopsy, which may reveal eosinophilic inflammation or the remnants of T.
canis larvae. Percutaneous biopsy of this benign inflammatory lesion
under imaging guidance is necessary to differentiate the lesion from hepatic
metastasis when a patient has a history of malignancy and the hepatic lesion
cannot be differentiated with imaging alone. Hypereosinophilia and CT and
sonographic abnormalities resolve slowly after antihelminthic or steroid
treatment [10,
29].
Although in our study the time intervals between the initial examinations
and follow-up examinations were not regular, follow-up imaging studies were
valuable in showing changes in and resolution of hepatic lesions (Figs.
4A and
4B). Despite complete
resolution of all hepatic lesions, only three patients among the
pharmacologically treated patients underwent follow-up imaging studies, so
therapeutic response could not be evaluated and analyzed properly.
The clinical significance of hepatic lesions of VLM of T. canis
resides in the differential diagnosis from nodular hepatic metastasis. Some of
the patients in this series had a history of underlying malignant disease.
Because of mysterious beliefs, some patients with malignant or longstanding
wasting disease prefer to eat "health-promoting" foods, such as
uncooked flesh, organs, horns, and the blood of animals, which may have
nodular lesions of T. canis. Nodules in toxocariasis differ from
metastatic lesions in that nodules of VLM have fuzzy margins, are subtly low
attenuating, are oval or elongated, and are best seen or only seen in the
portal venous phase of CT
[27]. Rim enhancement at the
equilibrium phase is seen almost exclusively in metastasis but is seldom found
in eosinophilic necrosis [27].
This finding may be explained by the variable degree of centripetal
enhancement into tissues of different compositions. These findings of fuzzy
margins and nonspherical shape are criteria for differential diagnosis from
metastasis. Another element of clinical significance exists in evaluation of
the severity of toxocariasis and the therapeutic effect of antihelminthic
treatment.
This study had several limitations. When the patients were selected, the
results of ELISA, which has a sensitivity of 91% and a specificity of 86%,
were used as the standard of reference in determining the diagnosis.
Therefore, false-positive results caused by cross reaction with other
parasites might have been included in this patient group, obscuring the true
CT and sonographic findings of VLM of T. canis. However, common
parasites causing cross reaction were excluded by ELISA or stool examination,
and the remaining false-positive proportion would have been too small to alter
the major CT and sonographic findings. The presence or absence and the
severity of abnormality on images may depend on the quantity of ingested
uncooked bovine tissue and consequently on the worm burden, but we could not
substantiate this factor because there is no specific test or method for
quantifying worm burden. There also were limitations associated with the use
of different imaging techniques for this retrospective study. Although most of
our imaging studies were performed with reasonably modern CT, there was lack
of standardization in terms of imaging protocols. However, all patients
underwent portal venous phase CT, in which the lesions of VLM are best
observed. Sonographic examinations were performed by many radiologists with
varying expertise; therefore, the detectability of small hepatic nodular
lesions might have been affected. Another limitation of this study was that CT
and sonographic images were reviewed by two radiologists by consensus rather
than through independent evaluation. However, this study was performed not to
evaluate the accuracy of the diagnostic technique or interpreters but to
describe the imaging characteristics of hepatic VLM of T. canis.
In summary, VLM of T. canis is a cause of hypereosinophilia. In
patients with sustained hypereosinophilia showing multiple small, oval or
elongated, ill-defined, low-attenuating or hypoechoic nodular lesions in the
liver on CT and sonography, VLM of T. canis should be considered.
Acknowledgments
The authors thank John Roberts, Harrisco, for editorial assistance, and
Young Joo Moon, Department of Radiology, Samsung Medical Center, for
assistance in manuscript preparation.
References
- Beaver PC, Snyder CH, Carrera GM, Dent JH, Lafferty JW. Chronic
eosinophilia due to visceral larva migrans: report of three cases.
Pediatrics 1952;9
: 7-19[Abstract/Free Full Text]
- Beaver PC. Larva migrans. Exp Parasitol1956; 5:587
-621[CrossRef][Medline]
- Glickman LT, Magnaval JF, Domanski LM, et al. Visceral larva
migrans in French adults: a new disease syndrome? Am J
Epidemiol 1987; 125:1019
-1034[Abstract/Free Full Text]
- Hayashi K, Tahara H, Yamashita K, et al. Hepatic imaging studies on
patients with visceral larva migrans due to probable Ascaris suum
infection. Abdom Imaging 1999;24
: 465-469[CrossRef][Medline]
- Beaver PC, Jung RC, Cupp EW. Clinical parasitology,
9th ed. Philadelphia, PA, Lea & Febiger, 1984:320
-329
- 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]
- Glickman L, Schantz P, Dombroske R, Cypess R. Evaluation of
serodiagnostic tests for visceral larva migrans. Am J Trop Med
Hyg 1978; 27:492
-498[Abstract/Free Full Text]
- Kaplan KJ, Goodman ZD, Ishak KG. Eosinophilic granuloma of the
liver: a characteristic lesion with relationship to visceral larva migrans.
Am J Surg Pathol 2001;25
: 1316-1321[CrossRef][Medline]
- Beaver PC. The nature of visceral larva migrans. J
Parasitol 1969; 55:3
-12[Medline]
- Bhatia V, Sarin SK. Hepatic visceral larva migrans: evolution of
the lesion, diagnosis, and role of high-dose albendazole therapy.
Am J Gastroenterol 1994;89
: 624-627[Medline]
- Fan CK, Lan HS, Hung CC, et al. Seroepidemiology of Toxocara
canis infection among mountain aboriginal adults in Taiwan. Am
J Trop Med Hyg 2004; 71:216
-221[Abstract/Free Full Text]
- Baldisserotto M, Conchin CF, Soares MG, Araujo MA, Kramer B.
Ultrasound findings in children with toxocariasis: report on 18 cases.
Pediatr Radiol 1999;29
: 316-319[CrossRef][Medline]
- Hakim SL, Mak JW, Lam PL. ELISA seropositivity for Toxocara
canis antibodies in Malaysia, 1989-1991. Med J
Malaysia 1993; 48:303
-307[Medline]
- Uhlikova M, Hubner J. Seroprevalence of Toxocara canis
infection in Czech Republic. Cent Eur J Public Health1998; 6:195
-198[Medline]
- Radman NE, Archelli SM, Fonrouge RD, Guardis M, Linzitto OR. Human
toxocarosis: its seroprevalence in the city of La Plata. Mem Inst
Oswaldo Cruz 2000; 95:281
-285[Medline]
- Ochiai H, Shimizu T, Hamada E. A case of Toxocariasis
cati: laparoscopic observation of a hepatic lesion.
Gastroenterol Endosc 1990;32
: 109-114
- Rayes A, Teixeira D, Nobre V, et al. Visceral larva migrans
syndrome complicated by liver abscess. Scand J Infect
Dis 1999; 31:324
-325[CrossRef][Medline]
- Rodan KS, Buckley JJ. Infection with adult Toxocara cati.BMJ 1969; 1:188[Free Full Text]
- Hirata T, Yamasaki K, Li YG, Majima Y, Tsuji M. Demonstration of
hepatic granuloma due to visceral larva migrans by ultrasonography.
J Clin Ultrasound 1990;18
: 429-433[Medline]
- Hartleb M, Januszewski K. Severe hepatic involvement in visceral
larva migrans. Eur J Gastroenterol Hepatol2001; 13:1245
-1249[CrossRef][Medline]
- Jain R, Sawhney S, Bharava DK, Panda SK, Berry M. Hepatic
granulomas due to visceral larva migrans in adults: appearance on US and MRI.
Abdom Imaging 1994;19
: 253-256[CrossRef][Medline]
- Jacquier P, Gottstein B, Stingelin Y, Eckert J. Immunodiagnosis of
toxocariasis in humans: evaluation of a new enzyme-linked immunosorbent assay
kit. J Clin Microbiol 1991;29
: 1831-1835[Abstract/Free Full Text]
- Schantz PM, Glickman LT. Toxocaral visceral larva migrans.
N Engl J Med 1978;23
: 436-439
- Taira K, Saeed I, Permin A, Kapel CM. Zoonotic risk of Toxocara
canis infection through consumption of pig or poultry viscera.
Vet Parasitol 2004;7
: 115-124
- Kim GB, Kwon JH, Kang DS. Hypereosinophilic syndrome: imaging
findings in patients with hepatic involvement. AJR1993; 161:577
-580[Abstract/Free Full Text]
- Lim JH, Lee WJ, Lee DH, Nam KJ. Hypereosinophilic syndrome: CT
findings in patients with hepatic lobar or segmental involvement.
Korean J Radiol 2000;1
: 98-103[Medline]
- Jang HJ, Lee WJ, Lee SJ, Kim SH, Lim HK, Lim JH. Focal eosinophilic
necrosis of the liver in patients with underlying gastric or colorectal
cancer: CT differentiation from metastasis. Korean J
Radiol 2002; 3:240
-244[Medline]
- Cha SH, Park CM, Cha IH, et al. Hepatic involvement in
hypereosinophilic syndrome: value of portal venous phase imaging.
Abdom Imaging 1998;23
: 154-157[CrossRef][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 Imaging 2003;28
: 326-332[CrossRef][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. AJR 1999;172
: 1255-1261[Abstract/Free Full Text]
- Nam KJ, Jung WJ, Choi JC, et al. Hepatic involvement in
hypereosinophilia: sonographic findings. J Ultrasound
Med 1999; 18:475
-479[Abstract]
- Won JH, Kim MJ, Kim BM, et al. Focal eosinophilic infiltration of
the liver: a mimic of hepatic metastasis. Abdom
Imaging 1999; 24:369
-372[CrossRef][Medline]
- Iwasaki K, Torisu M, Fushimura T. Malignant tumor and eosinophils.
Part 1. Prognostic significance in gastric cancer.
Cancer 1986; 58:1321
-1327[CrossRef][Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
