| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Mallinckrodt Institute of Radiology, Washington University Medical Center, 510 S Kingshighway Blvd, St Louis, MO 63110. Received June 24, 1998; revision requested August 5; final revision received April 26, 1999; accepted July 28.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: A computer search of US reports completed between 1991 and 1994 helped identify 383 patients whose reports contained the word "hemangioma." One hundred eleven patients were excluded because the lesion's appearance was atypical (n = 16) or because the patients had a high risk of malignancy (prior history or current evidence of extrahepatic malignancy or chronic hepatic disease [n = 95]). Fifty-nine patients were excluded because they were lost to follow-up (n = 41) or had clinical follow-up of less than 2 years (n = 18). The conditions of the remaining 213 patients with typical-appearing hemangiomas and a low risk of malignancy were analyzed. One hundred twenty-one patients underwent imaging follow-up or histopathologic confirmation. Ninety-two had clinical follow-up of more than 2 years (mean, 46 months).
RESULTS: Of the 213 patients, four had benign lesions other than hemangiomas. One patient who subsequently developed a malignancy (neuroendocrine metastases from primary colonic carcinoma diagnosed 22 months after initial US) potentially had an early metastasis that was misdiagnosed as a hemangioma.
CONCLUSION: On the basis of these results, the authors no longer recommend follow-up studies in their patients with a low risk of malignancy and a typical-appearing hemangioma at US.
Index terms: Angioma, gastrointestinal tract, 761.3194 • Liver, US, 761.1298 • Liver neoplasms, 761.30, 761.33, 761.3194 • Liver neoplasms, US, 761.1298
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Therefore, the care of patients with typical-appearing hemangiomas first demonstrated at US varies and depends on their risk for a primary or secondary hepatic malignancy. Most patients with a prior history or current evidence of an extrahepatic malignancy and/or chronic hepatic disease should undergo a confirmatory examination such as computed tomography (CT), magnetic resonance (MR) imaging, or technetium 99m–labeled red blood cell single photon emission CT (SPECT) to distinguish between a hemangioma and a hepatic malignancy (16,17).
However, there are conflicting recommendations for patients with no known risk factors for hepatic malignancy; they range from performing confirmatory examinations such as MR imaging or scintigraphy, to obtaining a follow-up US image in 6 months to confirm stability, to performing no further imaging examinations (1,8,16–21). Our anecdotal experience with typical-appearing hemangiomas in patients with no risk factors for hepatic malignancy suggests that the yield of follow-up studies in establishing diagnoses other than hemangioma is extremely low.
The purpose of our study was to determine if follow-up imaging is indicated in patients with typical-appearing hepatic hemangiomas at US and with no known risk factors for malignancy.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
After the reports were reviewed by one of the authors (D.M.L.), it was clear that the word "hemangioma" was used not only when the lesion in question was typical of hemangioma, but also when the lesion was atypical. For the purposes of this study, a typical hepatic hemangioma was considered to be present when the lesion was (a) well defined, homogeneously hyperechoic with or without central hypoechoic areas, and without posterior shadowing (Fig 1) or (b) hypoechoic or isoechoic with a hyperechoic rim (Fig 2). The presence of posterior acoustic enhancement was not considered in this analysis.
|
|
|
The remaining 272 patients were considered to be at low risk for malignancy and were considered for inclusion in the study if they had appropriate follow-up. Forty-one patients were lost to follow-up entirely. Eighteen patients had clinical follow-up of less than or equal to 24 months and were excluded because it would be difficult to completely dismiss the possibility of a very slow growing malignancy if follow-up was not at least 24 months. The remaining 213 patients composed the study group of patients with a low risk for hepatic malignancy and US findings typical of a hemangioma.
To confirm or exclude the diagnosis of hemangioma in our study group, all pertinent imaging reports, histopathology records, and medical records (obtained from private physicians' offices or from hospital-based records departments) were reviewed by one of the authors (D.M.L., C.O.M., or J.R.L.). We accepted several forms of follow-up.
Histopathologic Analysis
Any histopathologic analysis of the lesion in question was considered definitive.
Imaging
A confirmatory imaging examination other than US was performed within 6 months of US, without further follow-up. Images from these confirmatory examinations were reviewed by one of the authors (D.M.L.) when the report did not describe findings diagnostic of a hemangioma.
Helical or conventional contrast material–enhanced dynamic CT findings considered diagnostic of hemangioma were peripheral nodular enhancement or delayed fill-in of the lesion with typical attenuation. Heavily T2-weighted and gadolinium-enhanced MR imaging characteristics diagnostic of a hepatic hemangioma included a hyperintense lesion of equal or nearly equal signal intensity as cerebrospinal fluid and/or characteristic enhancement similar to that described for CT. Tagged red blood cell SPECT demonstrated lack of early angiographic flow with delayed fill-in of the lesion.
Lesions were also considered to be hemangiomas if follow-up US, CT, or MR images showed that the lesion had not changed in size after 6 months or longer. Images from follow-up examinations were reviewed by one of the authors (D.M.L.), and a consensus was arrived at whenever the dictated report was not descriptive enough to confirm the stability of the lesion.
Clinical Follow-up
Review of patient medical records was performed by one of the authors (D.M.L., C.O.M., or J.R.L.) when follow-up images were not available. Records were reviewed to determine if a diagnosis of malignancy was established after the diagnosis of a hepatic hemangioma was made on the basis of the initial US findings.
Eighteen patients with less than 24 months of follow-up were considered to have had insufficient clinical follow-up to exclude malignancy. However, their charts were still reviewed to determine if a hepatic malignancy was diagnosed within the available period of follow-up.
When medical records were not available, a confidential telephone interview with the patient was held by one of the authors (D.M.L., C.O.M., or J.R.L.) to review pertinent clinical history and to determine if a diagnosis of malignancy had been established at any time after the diagnosis of hemangioma was made on the basis of the initial US findings.
Patients were questioned as to whether pertinent imaging had been performed at another institution and whether there was a history of tumor, cancer, chronic hepatic disease, or other medical problems such as jaundice, ascites, or alcohol consumption. They also were questioned regarding a history of chemotherapy and/or radiation therapy. Our institutional review board approved this interview, and the patients provided informed consent.
The Fisher exact test was used to evaluate the null hypothesis that the proportion of patients with malignant hepatic tumors misdiagnosed as hemangiomas was equal to zero. The criterion for significance (
) was set at .05. The test was one-tailed, which means that only an effect in the positive direction was interpreted.
The test was performed with StatXact-Turbo: Statistical Software for Exact Nonparametric Inference (Cytel Software, Cambridge, Mass). The power of the test was determined with nQuery Advisor Study Planning Software, Release 2.0 (Statistical Solutions, Cork, Ireland). The 95% CI was calculated for the proportion of patients with malignant hepatic tumors misdiagnosed as hemangiomas with an add-on module to the software described earlier that calculated exact binomial CIs.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The lesions identified at US were 0.6–6.0 cm in diameter (mean size, 1.8 cm). One hundred fifty-seven lesions were larger than 1.5 cm, and 110 lesions were smaller than or equal to 1.5 cm. Two hundred sixty-one (98%) of the 267 lesions were homogeneous and hyperechoic. The rest were hypoechoic or isoechoic with a hyperechoic rim. Additional clinical information about the study patients is shown in the Table.
|
|
In this group of 74 patients, there was one identified patient in whom a hepatic metastasis was potentially misdiagnosed as a hemangioma. This patient had a hyperechoic, well-defined hepatic mass of 2.2 cm thought to represent a hemangioma. The patient subsequently received a diagnosis of diffuse hepatic metastatic disease from a neuroendocrine carcinoma of the colon 22 months later on the basis of contrast-enhanced CT results. Hepatic metastases were proved at percutaneous biopsy. Although it was difficult to separate the suspected hemangioma from all of the other lesions at CT, there was no CT evidence of a typical hemangioma in the expected position of the hemangioma. For this reason, this potentially represents a case in which a hepatic malignancy was misdiagnosed as a hemangioma on the original US image.
Three patients in this group of 74 with radiologic follow-up of greater than 6 months went on to develop colonic cancer. However, in all three patients, the original diagnosis of hepatic hemangioma was confirmed with contrast-enhanced CT performed 16, 54, and 72 months after the original US-aided diagnosis of hemangioma. One patient in this group with a 3.5-cm hyperechoic mass subsequently underwent MR imaging, with findings consistent with focal fatty infiltration. All of the others had lesions that were stable in size and/or had imaging features consistent with hemangiomas.
Forty-two patients underwent a confirmatory imaging examination, as previously described, less than 6 months after their initial US examination. The mean time was 1 month, with a range of 0–5 months. Confirmatory imaging examinations performed during this time included contrast-enhanced CT in 12 patients, contrast-enhanced MR imaging in six, and red blood cell scintigraphy in 24. One of these patients with multiple hyperechoic lesions of less than or equal to 15 mm had MR imaging findings consistent with focal fatty infiltration. All of the others had findings consistent with hemangiomas.
Ninety-two patients had clinical follow-up of greater than 2 years. The medical records of 66 patients were reviewed. The mean clinical follow-up time was 46 months (range, 24–72 months). Seven patients were followed up for 24–30 months; 40 patients, for 30–50 months; and 19 patients, for 50–72 months. All 66 of these patients were living at the time of their last follow-up. None of these patients developed clinical evidence of hepatic malignancy.
In 26 of the 92 patients with clinical follow-up longer than 2 years, medical records were not available or did not provide clinical follow-up data for an interval greater than 24 months from the initial US examination. These 26 patients were contacted by telephone for an interview, with a mean follow-up of 52 months (range, 39–84 months). None of the patients in this group developed clinical evidence of hepatic malignancy.
Fifty-nine patients were lost to follow-up or had insufficient follow-up. They included 18 patients with benign clinical follow-up findings at 12–24 months, none of whom had any clinical evidence of hepatic malignancy. The Table compares the 59 patients who were lost to follow-up with the 213 patients who were included in the study. These two groups were similar in all other respects.
Results of the Fisher exact test indicated that the proportion of patients whose conditions were misdiagnosed (one of 213) was not significantly different from 0% (P = .5). This test had an approximate power of 74% to demonstrate the difference in proportions. The 95% CI for the percentage of patients whose conditions were misdiagnosed, with the assumption of one misdiagnosis, was 0%, 2.59%. Because it was not clear that this was a true misdiagnosis of hemangioma in a patient with a malignant hepatic tumor, the 95% CI was also calculated, with the assumption that there were 0 of 213 misdiagnoses. The resultant 95% CI was 0%, 1.40%.
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Most lesions in adults remain stable in size and in US appearance (10,23). In one study, 19 of 21 (90%) lesions remained stable in size over a follow-up of 5–84 months (23). The authors of this study suggested that growth of hemangiomas is unusual and that when it occurs it should alert the radiologist to the possibility of an alternate diagnosis. Gibney and colleagues (10) showed that hemangiomas diagnosed in adults were of a stable size and appearance and only rarely changed at follow-up imaging.
Occasionally, malignant and other benign lesions appear well marginated and hyperechoic and may thereby be confused with hemangiomas. Lesions known to simulate hemangiomas include metastases from a variety of primary tumors and hepatocellular carcinomas. Benign lesions such as focal fat, liver adenoma, and focal nodular hyperplasia can also occasionally mimic hemangioma.
Although multiple clinical factors will affect decisions about treating a patient with a typical-appearing hemangioma, it is generally accepted that patients with a known extrahepatic malignancy capable of metastasizing to the liver and patients at risk for hepatocellular carcinoma should undergo additional imaging to confirm the diagnosis of hemangioma or to prove the presence of a hepatic malignancy. Commonly, this will entail performance of tagged red blood cell SPECT (if the lesion is > 2 cm). Specificities and positive predictive values approaching 100% have been reported (24). MR imaging using heavy T2 weighting and dynamic enhancement with a gadolinium-based contrast agent is also an excellent technique for helping to diagnose hemangiomas (25), especially if the lesion is less than 2 cm or is adjacent to the heart or major hepatic vessels (which decreases the accuracy and sensitivity of tagged red blood cell studies). Dynamic contrast-enhanced CT is also helpful in many cases (18,26). Both CT and MR imaging can assist in identifying other lesions and in locating primary tumors. Although not frequently performed in the setting of a suspected hemangioma, percutaneous biopsy may allow a definitive diagnosis and has been proved to be a safe procedure (27,28). Because the need for additional imaging in these patients at high risk for hepatic malignancy is generally agreed on, we did not examine the patients further.
On the other hand, it is controversial whether additional imaging is necessary when a suspected hemangioma is identified in a patient without any of the previously mentioned risk factors for malignancy (17). Some authors recommend a confirmatory examination with a different imaging modality if the lesion is large or follow-up US if the lesion is small (21,29). Others recommend CT regardless of lesion size (30–32) or follow-up US at various intervals (10,33,34). Some textbooks present the option of either follow-up US or no further follow-up (16, 19,20). However, none of these references provide objective data to justify any particular approach. For this reason, we decided to examine this group of patients.
Our results show that the chance of mistaking a malignant hepatic lesion for a hemangioma in a low-risk patient is extremely remote. In fact, in our study, only one patient of 213 (0.47%) possibly had a malignant hepatic mass that was misdiagnosed as a hemangioma. We believe that it is just as possible that this patient had a hemangioma and coincidentally developed hepatic metastases that were diagnosed 22 months later. These results provide strong support against further imaging of any type in this group of patients at low risk for hepatic malignancy.
However, it is important to recognize that the history provided to the radiologist may be insufficient to help determine if the patient is at high or low risk for hepatic malignancy. In such cases, it is important to state in the report that the US findings are consistent with a hemangioma and require no further evaluation, provided that the patient has no prior history or current evidence of extrahepatic malignancy or chronic hepatic disease.
For instance, one patient who had a typical-appearing hemangioma had no prior history of malignancy but underwent chest radiography that showed a lung mass on the same day that US showed a hepatic mass. Because there was current evidence of an extrahepatic malignancy (a mass on the chest radiograph), this patient was placed in the high-risk group and was excluded from our study. A CT scan was subsequently obtained, and the hepatic mass was shown to be a metastasis. This illustrates why it is necessary to include in the dictated report a statement about current and prior evidence of malignancy or chronic hepatic disease.
When interpreting the results of this study, one should realize that some factors that might be thought of as risk factors for hepatic malignancy, such as advanced patient age, hepatomegaly, or lesion size or multiplicity, were not considered. Most important, we did not consider abnormal liver function test results as an indication of high risk and therefore included 23 patients with that history in our study group. It could be argued that abnormal liver function test results are a potential indicator of hepatic metastases. Although this is true, we suspect that most patients with hepatic metastases and abnormal liver function test results also have either a history of extrahepatic malignancy or present with current clinical or imaging evidence of extrahepatic malignancy.
It is also possible that abnormal liver function test results may indicate chronic hepatic disease, even when there is no known history of hepatic disease. This is a potential consideration if the patient is being referred by a clinician, such as a dedicated hepatologist, who treats a high percentage of patients with chronic hepatic disease. Only one of the patients in our study with a known history of abnormal liver function test results was referred by a hepatologist. In such a situation, it may be appropriate to include these patients in the high-risk group.
In applying the results of our study to one's own clinical practice, one can choose to exclude patients with abnormal liver function test results or other potential risk factors (such as multiple lesions or large lesions) from one's low-risk group. If that is done, then the high-risk group will enlarge and more patients will require further imaging examination. But the risk in the low-risk group is likely to decrease even further, and the chance of misdiagnosing a hepatic malignancy as a hemangioma will be even less than what we determined in our study.
There were two limitations to this study. Fifty-nine patients were either lost to follow-up or had what we considered to be insufficient follow-up. It is possible that one or more of these patients with a lesion diagnosed as hemangioma could have actually had a hepatic malignancy. Fortunately, this group composed only 22% of our total patients with typical-appearing hemangiomas and with low risk for hepatic malignancy. Although 18 patients in this group were excluded because their clinical follow-up was less than 24 months duration, all had follow-up between 12 and 24 months and none developed clinical evidence of hepatic malignancy in that period. Finally, the group that was excluded and the group that was lost to follow-up were similar to the study group in age; sex; clinical indication for the original US examination; and lesion size, appearance, and multiplicity. Therefore, we believe that it is very unlikely that this group of patients would have had a higher misdiagnosis rate than the much larger group of patients with adequate follow-up.
The other limitation was the lack of histopathologic proof in the majority of patients. In lieu of that, we accepted several forms of imaging or clinical proof. One form consisted of additional evaluation with contrast-enhanced CT, MR imaging, or red blood cell scintigraphy. We and other researchers believe that the typical results of US and of any of these other examinations are highly specific for the diagnosis of hemangioma (16,17).
Another form of proof was long-term stability. If the lesion was reimaged and was of similar size, we accepted 6 months as appropriate for satisfying the condition of long-term stability. This approach also is widely accepted (10,16,17,23). In addition, although we accepted 6 months as an adequate period, most of the patients in this group had lesions with size stability documented at follow-up considerably longer than 6 months (20- or 31-month mean follow-up for those reexamined with US or other techniques, respectively).
Finally, we also accepted clinical evidence of long-term stability, even if the patient was not reexamined with imaging. We assumed that a malignant hepatic neoplasm would manifest clinically within 2 years of its initial detection. Therefore, we included only patients with more than 2 years of follow-up in our statistical evaluation of the misdiagnosis rate. In fact, most of the patients in this group had follow-up much longer than 2 years (46- or 52-month mean follow-up for patients with medical record review or telephone interview, respectively). Therefore, on the basis of the length of our follow-up, we believe that it is very unlikely that we failed to uncover a slow-growing malignant neoplasm.
It is possible that some of the lesions in the group that were proved on the basis of size or clinical stability could have been other benign processes besides hemangiomas. As we found in the group with imaging confirmation, focal fat, focal nodular hyperplasia, and hepatic adenomas can simulate a hemangioma. However, focal fat is of no clinical importance; focal nodular hyperplasia has very little clinical importance apart from rare reports of hemorrhage in larger lesions (35,36); and hepatic adenomas, which have a greater propensity to bleed, are usually discovered and resected at the time of hemorrhage (36,37).
In conclusion, results of this study show that a hepatic lesion with a US appearance typical of hemangioma is extremely unlikely to be a malignancy, provided that the patient has no prior or current evidence of an extrahepatic malignancy or chronic hepatic disease. We believe that this results from both the low pretest probability that these patients will have a malignancy and the low probability that a malignant hepatic lesion will look like a hemangioma. On the basis of these results, we no longer recommend follow-up examinations of any type on our patients who fit into this group.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. H. Song, F. S. Chaudhry, and W. W. Mayo-Smith The Incidental Adrenal Mass on CT: Prevalence of Adrenal Disease in 1,049 Consecutive Adrenal Masses in Patients with No Known Malignancy Am. J. Roentgenol., May 1, 2008; 190(5): 1163 - 1168. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Song, F. S. Chaudhry, and W. W. Mayo-Smith The Incidental Indeterminate Adrenal Mass on CT (> 10 H) in Patients Without Cancer: Is Further Imaging Necessary? Follow-Up of 321 Consecutive Indeterminate Adrenal Masses Am. J. Roentgenol., November 1, 2007; 189(5): 1119 - 1123. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G J Kraus, P Schedlbauer, S Lax, D Zebedin, and F Flueckiger The reverse target sign in liver disease: a potential ultrasound feature in cirrhotic liver nodules characterization Br. J. Radiol., April 1, 2005; 78(928): 355 - 357. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. C. Eberhardt, P. H. Choi, A. M. Bach, S. A. Funt, H. E. Felderman, and L. E. Hann Utility of Sonography for Small Hepatic Lesions Found on Computed Tomography in Patients With Cancer J. Ultrasound Med., April 1, 2003; 22(4): 335 - 343. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Caturelli, M. Pompili, F. Bartolucci, D. A. Siena, M. Sperandeo, A. Andriulli, and M. Bisceglia Hemangioma-like Lesions in Chronic Liver Disease: Diagnostic Evaluation in Patients Radiology, August 1, 2001; 220(2): 337 - 342. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
