Joint Program in Nuclear Medicine
Imaging Workup of Hepatic Hemangiomas
Jac D. Scheiner, MD
Kevin J. Donohoe, MD
September 10, 1996
Presentation
A 53 year old woman with a history of a messencymal tumor of the thigh
presented with hepatomegally.
Imaging Findings
The ultrasound of the liver showed a hyperechoic mass in the dome
of the right lobe (seen a sagital and an oblique images; arrows
show lesions on sagital and on transverse images). In addition
there was a
complex lession at the inferior tip of the right lobe of the liver
(shown by arrow).
First past images from the blood pool study (above) show no areas of increased
flow. Planar delayed images show two areas of increase uptake
(the lesion in the dome of the right lobe corresponding to the
hyperechoic lesion is shown by arrow heads; the lesion at the tip of the
right lobe is shown by arrows; also seen are the blood pools of the
heart, H, and the spleen, S). Coronal SPECT images
show these two areas of increased blood pool.
Discussion
Hemangiomas are the most common benign tumor of the liver, noted
in up to 7% of patients on autopsy series. Up to 20% are
multiple. Histology demonstrates an overgrowth of enlarged
endothelial lined vascular spaces. While almost always
asymptomatic, these tumors have been associated with abdominal
pain, fullness, and belching. Complications are similarly rare,
but have included bleeding, infarction, necrosis, and
thrombocytopenia. The clinical importance of these incidentally
discovered tumors relates to distinguishing them from
malignancies.
Among the most common imaging modalities used to evaluate
hemangiomas are ultrasonography (US), computed tomography (CT),
magnetic resonance imaging (MRI) and technetium-99m labeled red
blood cell single photon emission computed tomography (Tc-99m
RBC SPECT).
Ultrasonography
The classic finding of hemangioma on US is that of a well
defined, echogenic mass. This is seen in approximately 50-60%
of hemangiomas (1). However, this appearance can also be seen
in adenomas, hepatocellular carcinomas, and metastatic disease.
The specificity of US findings for hemangioma is increased if
one also sees increased through transmission, although this is
not always seen. Stable finding over several US exams
can be used to suggest the lesion is benign. In a study of 21
hemangiomas in 21 patients seen on US or CT, follow-up imaging
in 5-84 months (mean = 26 months) demonstrated size stability in
90% (19/21) (2). Of the 2 that had increased in size, one had
increased by 1 cm in 3 months and then remained stable at 3 year
follow-up, while the other had increased by 2 cm in 10 months.
In the latter case, the patient developed thrombocytopenia (an
extremely rare complication) and the hemangioma was resected.
Computed Tomography
The classic finding of hemangioma on CT is that of a round, low
attenuation lesion with slight lobulations on the non-contrast
images. Following contrast injection, "peripheral puddling" of
contrast (of an attenuation matching the aorta) is seen. The
classic findings are not always evident, however. A study of 58
hemangiomas in 47 patients who prospectively underwent CT with
and without contrast, using a more stringent criteria which
required the lesion to have an attenuation greater than or equal
to the liver on delayed images, demonstrated classic findings in
only 55% (32/58) of the hemangiomas (3).
Tc-99m Labeled RBC SPECT
The classic findings of hemangioma on Tc-99m labeled RBC SPECT
are a lesion that is undetectable or demonstrates decreased
intensity compared to the liver on the flow images, and
increased intensity to the liver on the 30-60 minute delayed
images. These findings are nearly 100%
specific for hemangioma. In the literature, there have
been only 8 reported false positive cases of hemangioma (4-9).
Two of these were cases of hepatocellular carcinoma in which
there was increased tracer localization on the flow images.
Another 2 were cases of hepatocellular carcinoma in patients
with chronic liver disease. The remaining 4 were in cases of
intrahepatic extramedullary hematopoiesis, metastatic gastric
carcinoid, metatstatic colon cancer, and angiosarcoma. Of these
8 cases, only one (metastatic colon cancer) had characteristic
hemangioma features (on both flow and delayed images) and could
be considered a common tumor.
SPECT significantly improves the sensitivity for hemangioma
detection over planar imaging, especially for lesions less than
5.0 cm. A study of 77 hemangiomas in which planar imaging and
single head SPECT were performed showed the sensitivity of these
2 techniques for different size lesions to be as follows (10):
| Size (cm) | Planar (%) | SPECT (%)
|
|---|
| <=2.0 | 6/31 (19) | 9/26 (35)
|
| 2.1-3.0 | 12/21 (57) | 16/19 (84)
|
| 3.1-5.0 | 13/16 (81) | 15/15 (100)
|
| >5.0 | 9/9 (100) | 9/9 (100)
|
Another study of 108 hemangiomas less than 5.0 cm, in which
planar imaging and single head SPECT was performed, gave the
following sensitivities (11):
| Size (cm) | Planar (%) | SPECT (%)
|
|---|
| <=1.0 | 0/20 (0) | 0/20 (0)
|
| 1.1-2.0 | 3/29 (10) | 11/29 (38)
|
| 2.1-3.0 | 15/35 (43) | 29/35 (83)
|
| 3.1-4.0 | 8/13 (62) | 12/13 (92)
|
| 4.1-5.0 | 10/11 (91) | 11/11 (100)
|
This same study also looked at 45 hepatocellular carcinomas
sized 1.5-5.0 cm. None of these malignancies met the
scintigraphic criteria for hemangioma.
Two headed SPECT study of 56 hemangiomas resulted in slightly
improved sensitivities (13):
| Size (cm) | Planar(%) | SPECT1(%) | SPECT2(%)
|
|---|
| <=1.0 | 1/11 (9) | 2/11 (18) | 9/11 (92)
|
| 1.1-2.0 | 3/14 (21) | 7/14 (50) | 13/14 (93)
|
| 2.1-3.0 | 6/11 (55) | 9/11 (82) | 11/11 (100)
|
| 3.1-4.0 | 5/6 (83) | 6/6 (100) | 6/6 (100)
|
| >=4.1 | 14/14 (100) | 14/14 (100) | 14/14 (100)
|
SPECT1 refers to conventional SPECT images, whereas SPECT2
refers to SPECT images interpreted on a 3-D digital display.
A study in which 3 headed SPECT was performed on 34 hemangiomas
(0.5-6.8 cm, mean = 2.2 cm) gave the following sensitivity
results (14):
| Size (cm) | SPECT (%)
|
|---|
| <=0.8 cm | 1*/5 (20)
|
| 0.9-1.3 cm | 3/9 (33)
|
| >=1.4 cm | 20/20 (100)
|
| * this was a 5 mm hemangioma
|
False negative exams most commonly occur because of small
hemangioma size. Other reported causes include the hemangioma
being adjacent to a vascular structure (such as the heart,
portal vein, or IVC) (13), perceived increased activity on the
flow images (reported in 6% and 8% of cases in 2 studies)
(10,15), and extensive fibrosis within the tumor itself (15).
MRI
The classic findings of hemangioma on MRI are that of a
homogeneous, well defined lesion with slight lobulations that is
low signal relative to liver on the T1 weighted images, becomes
progressively brighter (greater than or equal to cerebral spinal
fluid) on the T2 weighted images as the time to echo (TE) is
lengthened, and demonstrates peripheral puddling of contrast on
the enhanced images (of a signal intensity matching the aorta).
Approximately 5% of malignant tumors mimic hemangiomas on these
long TE images, with the most common mimics being sarcomas,
endocrine tumors, and cystadenocarcinomas (16). The common
liver metastases (colon, lung, pancreas) are poor mimics on MRI.
One of the useful distinguishing features on MRI is that,
whereas both hemangiomas and malignancies tend to have high
signal on conventional T2 weighted images, the signal from
malignancies tends to decrease as the TE is lengthened, whereas
the signal from hemangiomas tends to increase.
MRI was compared to Tc-99m RBC SPECT
in 69 lesions seen on CT and/or US ranging in size from 1.0
to 13.0 cm (12). All patients underwent Tc-99m RBC SPECT (single
head) and 0.5 Tesla MRI (TR/TE = 2000/50 and 2000/100). The
final diagnosis was hemangioma in 64 of these cases. The
sensitivity results were as follows:
| Size (cm) | SPECT (%) | MRI (%)
|
|---|
| 1.0 - 1.9 cm | 14/24 (58) | 20/24 (83)
|
| 2.0 - 2.9 cm | 12/13 (92) | 11/13 (85)
|
| 3.0-13.0 cm | 24-27*/27 (89-100) | 27/27 (100)...
|
the 3 false negatives on SPECT were ‘hot’ on
the
delayed images and the flow images.
|
Summary
Tc-99m RBC SPECT is a highly specific examination in the
diagnosis of hemangioma, with only 8 reported false positive
cases in the literature. While the sensitivity of MRI is
greater than that of SPECT, it lacks SPECT’s high specificity.
- For possible hemangiomas greater than 2.0
cm, where the sensitivity of SPECT has been reported
to range from 83% to 100%, Tc-99m RBC SPECT is the study of
choice.
- For possible hemangiomas < 1.0 cm, the sensitivity of SPECT
has been reported to range from 0% to 20%. Since small size often
prohibits diagnostic characterization on MRI,
biopsy or follow-up imaging to document stability are the
recommended.
- In the 1.0 to 2.0 cm category, the sensitivity of SPECT has
been reported to range from 33% - 58%, whereas the sensitivity
of MRI has been reported to be approximately 83%. If the
possible hemangioma is adjacent to a highly vascular structure
(i.e. the heart, portal vein, or IVC), it may ‘blend’ into the
activity of the adjacent high activity and not be
distinguished. For lesions adjacent to a highly vascular
structure lesions, MRI will probably have a better diagnostic
yield. The main false positive MRI exams result from metastases
from hypervascular or cystic primary tumors. In patients with
a history of such primary tumors, Tc-99m RBC SPECT will
probably have a higher diagnostic yield. In cases in which
Tc-99m RBC SPECT is equivocal, MRI can be a useful adjunctive
study.
References
1. Prakash R, Gupta RK, Narayyanan R, et al. Tc-99m
radiocolloid scintigraphy, planar, and SPECT red blood cell
imaging and ultrasonography in the diagnosis of hepatic
hemangioma. Austr Radiol, 1989;33:237-244.
2. Mungovan JA, Cronan JJ, Vacarro J. Hepatic cavernous
hemangiomas: lack of enlargement over time.
Radiology,1994;191:115-117.
3. Freeny PC and Marks WM. Hepatic hemangioma: dynamic bolus
CT. AJR; 1986:147:711-719.
4. Tamm EP, Rabushka LS, Fishman EK, et al. Intrahepatic
extramedullary hematopoiesis mimicking hemangioma on Tc-99m red
blood cell SPECT examination. Clin Imag 1995;19:88-91.
5. Ginsberg F, Slavin JD and Spencer RP. Hepatic
angiosarcoma: mimicking of angioma on three-phase Tc-99m red
blood cell scintigraphy. JNM 1986;27:1861-1863.
6. Swayne LC, Diehl WL, Brown TD and Hunter NJ. False-positive
hepatic blood pool scintigraphy in metastatic colon carcinoma.
Clin Nucl Med 1991;16:630-632.
7. Farlow DC, Little JM, Gruenewald SM, et al. A case of
metastatic malignancy masquerading as a hepatic hemangioma on
labelled red blood cell scintigraphy. JNM 1993;34:1172-1174.
8. Rabinowitz SA, McKusik KA, Strauss HW. Tc-99m red blood cell
scintigraphy in evaluating focal liver lesions. AJR
1984;143:63-68.
9. Intenzo C, Kim S, Madsen M, et al. Planar and SPECT Tc-99m
red blood cell imaging in hepatic cavernous hemangiomas and
other hepatic lesions. Clin Nucl Med 1988;13:237-240.
10. Bonanno N, Baldari S, Cerrito A, et al. Diagnosis of
hepatic hemangiomas with Tc-99m labelled red blood cell
scanning: value of SPECT. J of N Biology and Medicine,
1991;35:135-140.
11. Kudo M, Ikekubo K, Yamamoto K, et al. Distinction between
hemangioma of the liver and hepatocellular carcinoma: value of
labelled RBC - SPECT scanning. AJR 1989;152:977-983.
12. Birnbaum BA, Weinreb JC, Megibow AJ, et al. Definitive
Diagnosis of Hepatic Hemangiomas: MR Imaging versus Tc-99m -
labeled red blood cell SPECT. Radiology 1990;176:95-101.
13. Krause T, Hauenstein K, Studier-Fisher B, et al. Improved
evaluation of Tc-99m RBC SPECT in hemangioma of the liver. JNM,
1993;34:375-380.
14. Zeissman HA, Silverman PM, Petterson J, et al. Improved
detection of small cavernous hemangiomas of the liver with high
resolution 3 headed SPECT. JNM 1991;32:2086-2091.
15. Rabinowitz SA, McKusick KA, Strauss HW. Tc-99m red blood
cell imaging in evaluating focal liver lesions. AJR
1984;143:63-68.
16. Weissleder R and Stark DD. Magnetic resonance imaging of
liver tumors. Seminars in US, CT, and MRI. 1989;10:63-77.
17. Egglin TK, Rummeny E, Stark DD, et al. Hepatic Tumors:
Quantitative tissue characterization with MRI. Radiology,
1990:176;107-110.
18. Choi BI, Han MC, Kim CW. Small hepatocellular carcinoma
versus small cavernous hemangioma: differentiation with MR
imaging at 2.0 T. Radiology 1990;176:103-106.
Click here to go
to Joint Program in Nuclear Medicine home page and Copyright
notice.
J. Anthony Parker, MD PhD, Tony_Parker@bidmc.harvard.edu