|
SEATTLE,
WA— The radiologic evaluation of the hepatic mass
with ultrasound, CT and MRI is in a continuing state of evolution.
These three imaging modalities have dramatically increased the detection
of focal hepatic pathol-ogy and, in some instances, allowed clinicians
to almost characterize the nature of the focal lesion, Paul D. Russ,
MD, said at the annual meeting of the American College of Gastroenterology.
“Cross-sectional imaging has a relatively high sensitivity,
but if you look at the radiologic literature, it’s hard to
dissect out the true sensitivity of each modality,” said Dr.
Russ, Professor of Radiology, University of Colorado Health Sciences
Center, Denver—a major liver transplant centre.
These modalities have a sensitivity of about 85 per cent, he said.
As for specificity (i.e. the ability to identify what a lesion is),
the range is anywhere from 15 to 85 per cent, depending on the modality,
clinical circumstances, pre-test probabilities and type of contrast
agent administered during the scan. Both sensitivity and specificity
are increased with exogenous contrast mediums.
Dr. Russ stressed that these imaging modalities are unpredictably
complementary. “A patient asks me which type of imaging to
have and I say, ‘I really don’t know.’ You can’t
predict in advance in an individual patient which one will give
you the answer.”
Nor will pathology detected by ultrasound necessarily show up on
the most sophisticated CT and MRI scans. “That’s why at
our institution we practise multimodality imaging. If there’s
a strong index of suspicion, say, that a patient has hepatocellular
carcinoma and one test doesn’t show the abnormality, we keep
on going.”
Ultrasonography is an inexpensive screening tool that helps get
an initial handle on patients suspected of having focal pathology.
It is an excellent method of differentiating fluid collections from
solids and hence detecting cysts or cystic lesions, he said. With
a specificity for diagnosing an incidental hepatic cyst exceeding
99 per cent, it performs much better than CT or MRI. “But if
we see a solid lesion on ultrasound, I can’t tell you at all
what that lesion is.”
Moreover, ultrasound has only limited ability to penetrate gas,
fat and bone. “Ultrasound, in fact, is actually the most difficult
cross-sectional imaging that we do.
“It is extremely operator dependent, extremely difficult to
perform, and very difficult to interpret. Results are quite variable
and depend on the skill of the technologist and the interpreting
radiologist.”
It has a sensitivity of 80 to 85 per cent for detecting a mass and
a variable specificity that depends on the number of cysts in the
cohort.
The essential workhorse of hepatic imaging and focal mass evaluation
is CT, he said.
“CT has tremendous spatial resolution but unfortunately does
require the use of oral and IV contrast agents because of its mediocre
contrast resolution. And it is contrast resolution that allows us
to differentiate a focal mass from background liver parenchyma.”
The advent of spiral and multi-detector technology allows dual and
multiphase CTs, which are very useful for finding a variety of pathologies.
CT is also the workhorse for doing image-guided interventional procedures
such as biopsy to establish liver lesion
histology.
MRI is no more expensive than CT, Dr. Russ said. State-of-the-art
MRI can image the entire liver in a single breathhold. MRCP can
image the biliary tree in seven seconds. A lot of information is
gathered in a 15- or 20-minute scan. And unlike CT, which is limited
to iodine, MRI uses a variety of contrast agents.
Even without ionizing radiation, the spatial resolution of MRI competes
with that of advanced CT scanners, while its contrast resolution
is intrinsically better than contrast-enhanced CT. Exogenous MR
contrast agents only improve this resolution.
“It should be noted that all MRI contrast agents have a better
biosafety profile than iodinated CT contrast medium,” he said.
“So if I have a choice between having a scan with iodine or
gadolinium, I’d go with gadolinium every time.”
Dr. Russ outlined the radiologic appearance of several common hepatic
lesions, noting that in some cases imaging features are highly suggestive
of the histological diagnosis. “But in many cases, a focal
hepatic mass has no distinct radiologic characteristics.”
Hepatic cysts
Liver cysts occur in five to 10 per cent of patients 50 years and
older and are thought to derive mainly from intrahepatic bile duct
hamartomas. Ultrasound is best for determining if the lesion is
fluid-filled and hence a cyst. At sonography, a simple cyst appears
round to oval, is anechoic (homogeneously black) and well demarcated
with a thin, discrete wall. It is associated with increased through-transmission.
A routine simple hepatic cyst does not require biopsy for diagnosis.
Only two per cent of malignant metastases are cystic and these occur
in patients with pancreatic or ovarian cystadenocarcinomas. Primary
intrahepatic biliary cystadenomas and cystadenocarcinomas are rare.
Simple cysts are more common than cystic neoplasms.
Hemangioma
Cavernous hemangiomas occur in 5 to 15 per cent of all adult livers
and are multiple in about 10 per cent of cases. Usually 2 to 5cm
in diameter, hemangiomas are composed of large venous sinusoids
separated by thin fibrous septations.
On ultrasound, these lesions appear round to oval, are well demarcated
and may be mildly lobular. With the strong acoustic interfaces of
the multiple septations, these lesions are highly echogenic (bright
white). If the clinical suspicion for other pathology is low, then
a presumed ultrasonography-detected hemangioma needs only periodic
imaging surveillance to ensure stability.
On dynamic CT and MRI, a typical hemangioma has fluffy and rounded
discontinuous enhancing nodules during the arterial phase of contrast
administration. As the contrast centripetally fills the lesion to
the point of homogeneity, enhancement follows that of the blood
pool of other intra-abdominal vascular structures. As with ultrasound,
these findings on CT or MRI indicate conservative management of
the presumed hemangioma.
In equivocal cases, suspected hemangiomas should be evaluated with
a technetium-99m red blood cell scan in conjunction with single
photon emission CT.
Due to slow blood flow in the venous sinusoids, hemangiomas tend
to sequester the tagged RBCs, producing a focal hot spot on SPECT
and achieving a specificity of 95 per cent for hemangioma.
Focal nodular hyperplasia
Focal nodular hyperplasia (FNH) is found in about one per cent of
the population. It is a benign proliferative mass of normal cellular
elements surrounding a central congenital ateriovenous malformation.
Named the “stealth lesion” for its similarity to background
liver, FNH is best characterized by dynamic CT or MRI, but the exact
specificity of these modalities in the diagnosis of FNH is not well
established.
With IV contrast, the vascular malformation contained in a small
central scar of loose connective tissue quickly enhances and perfuses
the well-demarcated, somewhat lobular lesion centrifugally. The
central vasculature can have a spokewheel configuration.
The accumulation of contrast medium in the extracellular space of
the scar supports a diagnosis of FNH, but is not patho-gnomonic.
Depending on imaging features and clinical setting, FNH can be monitored,
biopsied or surgically removed.
Adenoma
Hepatic adenomas mostly occur in women with an incidence of three
or four per 100,000.
Believed to be potentiated by oral contraceptive use, this lesion
is composed of well-differentiated hepatocytes and is known to hemorrhage
spontaneously. Due to their slight potential for malignancy, adenomas
are surgically removed when this is technically feasible.
By ultrasound, CT and MRI, hepatic adenomas have few characteristic
imaging features, appearing usually as nondescript focal masses.
Since imaging alone cannot distinguish adenoma from hepatocellular
carcinoma, biopsy and ultimately excision are required to confirm
the diagnosis.
Hepatocellular carcinoma
In the United States, the overall incidence of hepatocellular carcinoma
(HCC) is two or three cases per 100,000. Many cases are now associated
with hepatitis C-induced cirrhosis. In the setting of HCV and cirrhosis,
any focal solid mass is considered HCC until proven otherwise.
Since HCC has no unique ultrasound features, dual-phase IV-enhanced
CT or MRI are the imaging modalities of choice for detecting HCC
in patients with end-stage liver disease. On CT or MRI, small HCCs
(2 to 3cm) appear as hypervascular (bright) nodules during the arterial
phase, fading to inconspicuous in the later portal vein phase. HCCs
of greater than 3cm usually show some internal degeneration or necrosis.
At dual-phase imaging, these lesions have a subtler, heterogeneous
enhancement during the arterial phase and become more conspicuous
during the portal vein phase.
HCCs can have a central scar consisting of dense avascular collagen,
which does not enhance or retain contrast. HCC should be suspected
when a hepatic mass is associated with portal vein thrombosis. Between
30 and 70 per cent of HCCs are associated with invasion of the portal
vein, which occurs only rarely with other focal hepatic masses.
As for fibrolamellar HCC—which can occur de novo in young adults
with no hepatocellular pathology—the radiologic features are
distinct from those in patients with end-stage liver disease. Seventy-five
per cent of fibrolamellars occur in the left hepatic lobe. Fibrolamellar
carcinoma may resemble a large FNH on dynamic CT or MRI. It often
has a central scar that does not enhance. A crucial finding is the
presence of punctate, dystrophic calcifications within the scar,
which occur in 50 per cent of fibrolamellars but fewer than five
per cent of FNHs. Therefore, central calcification in a focal mass
suggests fibrolamellar carcinoma until this can be ruled out.
Metastasis
Metastases can look like just about anything, so solitary metastasis
is included in the radiologic differential for virtually every focal
hepatic lesion. It is unusual for an imaging characteristic to suggest
the true nature of the lesion. Psammomatous calcification suggests
metastasis from a mucinous GI carcinoma. Neuroendocrine metastases
may be inconspicuous on non-contrast CT but appear hyperintense
on non-contrast T2-weighted MRI.
Although all hepatic masses can be multifocal, metastases are most
likely to present as several to innumerable to confluent lesions
randomly distributed throughout the liver. Image-guided fine-needle
aspiration to determine cytopathology is usually diagnostic, Dr.
Russ said.
|
