| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, University of Pittsburgh Medical Center, Presbyterian Hospital, 200 Lothrop St, Rm 4660 CHP MT, Pittsburgh, PA 15213-2582 (G.B., M.P.F., A.B.); and Department of Radiology, Spedali Civili, University of Brescia, Italy (L.G.). Received June 20, 2000; revision requested July 24; revision received August 17; accepted August 31. G.B. supported by the Nicholas Green Fulbright Grant. Address correspondence to M.P.F. (e-mail: federle+@pitt.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Imaging and pathologic findings of 21 hemangiomas in 17 patients were retrospectively reviewed. CT of the liver was performed in all patients; MR imaging, in four. Cirrhosis was confirmed histologically in all patients, and the diagnosis of hemangioma was based on histopathologic findings (15 patients, 18 hemangiomas) or strict imaging criteria (two patients, three hemangiomas). Ten patients underwent imaging follow-up. The number, sizes, location, attenuation, pattern of enhancement, exophytic growth, presence of capsular retraction, and size stability were evaluated.
RESULTS: Of the 21 hemangiomas, five were not detected at CT or MR imaging. Twelve (75%) of 16 hemangiomas were subcapsular, two (12%) of 16 demonstrated exophytic growth, 14 (87%) of 16 demonstrated nodular peripheral enhancement, and 16 (100%) of 16 were isoattenuating to blood vessels. At MR imaging, all five hemangiomas demonstrated nodular peripheral enhancement and hyperintensity on T2-weighted images. Seven lesions were smaller at follow-up, and five lesions developed retraction of the hepatic capsule.
CONCLUSION: Even within the cirrhotic liver, larger hemangiomas can usually be diagnosed confidently with CT or MR imaging. With progressive cirrhosis, however, hemangiomas are likely to decrease in size, become more fibrotic, and are difficult to diagnose radiologically and pathologically.
Index terms: Angioma, gastrointestinal tract, 761.3194 • Liver, cirrhosis, 761.794 • Liver neoplasms, CT, 761.12111, 761.12112, 761.12114, 761.12115 • Liver neoplasms, diagnosis, 761.3194 • Liver neoplasms, MR, 761.121411, 761.12143
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In a series of 508 consecutive hepatectomy specimens obtained during liver transplantation, Dodd et al (2) found only nine cavernous hemangiomas at gross pathologic examination, and only three of these were visualized at preoperative nonhelical CT. This incidence of hemangiomas (1.7% at pathologic examination, 0.6% at CT) is lower than the frequency with which hemangiomas are encountered at unselected autopsy series (3), or at CT in noncirrhotic patients.
The goals of this study were to investigate the natural history and diagnosis of cavernous hemangioma in the cirrhotic liver at CT and MR imaging.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The causes of cirrhosis were hepatitis C (n = 4), hepatitis B (n = 3), primary biliary cirrhosis (n = 2), primary sclerosing cholangitis (n = 2), autoimmune hepatitis (n = 2), alcoholism (n = 2), combined alcoholism and hemochromatosis (n = 1), and combined alcoholism with hepatitis C (n = 1).
Cirrhosis was confirmed histologically in all patients by means of percutaneous core-needle biopsy. In addition, the entire liver was available for pathologic examination in 14 patients (orthotopic liver transplantation in 13, autopsy in one). The diagnosis of hemangioma was based on histopathologic findings (15 patients, 18 hemangiomas) or strict imaging criteria (two patients, three hemangiomas), including nodular peripheral enhancement, isoattenuation to blood vessels, high signal intensity on T2-weighted images, and lack of interval increase in size. Altogether, 10 patients underwent serial CT and/or MR imaging follow-up of the focal hepatic lesions for a mean of 27 months (range, 5–86 months).
The pathologic diagnosis of hemangioma was based on the presence of variably sized, endothelial-lined vascular channels. The presence and amount of fibrosis within the hemangioma was also noted, along with any evidence of thrombosis, scarring, or hyalinization (obliteration of vascular channels with a fibrotic scar). The reports of the histopathologic examinations were reviewed by one author (G.B.).
CT studies were performed in all 17 patients and included the acquisition of contrast-enhanced images in all and nonenhanced images in 15. Helical CT was performed in 15 patients and included the acquisition of nonenhanced, hepatic arterial–dominant phase (HAP), and portal venous–dominant phase (PVP) images. The HAP series was obtained with a scanning delay of 25–35 seconds, and the PVP series was obtained 60–70 seconds following initiation of the intravenous bolus of contrast material. In the two patients who underwent conventional (nonhelical) CT, nonenhanced and enhanced images were obtained, with acquisition of the latter initiated after a 60-second delay. On average, patients received 150 mL of 60% iodinated contrast medium (iothalamate meglumine [Conray 60] or ioversol [Optiray 320]; Mallinckrodt Medical, St Louis, Mo). At conventional CT, the contrast agent was intravenously administered at a rate of 2.5 or 3.0 mL/sec; at helical multiphasic examinations, the rate was 4 or 5 mL/sec, administered by using a power injector (model OP 100; Medrad, Pittsburgh, Pa). Three patients also underwent delayed-phase imaging of the liver at 5–10 minutes after initiation of contrast medium administration. Section thickness was 5–7 mm for conventional and helical imaging. All CT studies were performed with CT 9800, Advantage, HiLight Advantage, HiSpeed Advantage, or LightSpeed scanners (GE Medical Systems, Milwaukee, Wis).
Four patients underwent MR imaging with a T1-weighted spin-echo sequence (repetition time msec/echo time msec, 140–700/12–20). Three patients also underwent T2-weighted spin-echo imaging (4,000–9,230/70–140; echo train length, eight or 16). T1-weighted imaging was repeated in all patients after the intravenous administration of contrast material. All patients received gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ) at a dose of 0.1 mmol per kilogram of body weight, and two underwent biphasic MR imaging during the hepatic arterial (25–35-second delay) and portal venous (60–70-second delay) phases. One patient underwent delayed imaging 10 minutes after the administration of contrast material. All MR studies were performed with 1.5-T imagers (Signa; GE Medical Systems).
CT and MR images were reviewed retrospectively and jointly by two abdominal radiologists (M.D.F., L.G.) in two separate sessions: first without knowledge of the specific number or locations of hemangiomas and second with full knowledge of the findings at gross pathologic examination and subsequent imaging that established proof of the number, location, and size of hemangiomas.
The two readers attempted to determine the number, size, and location of all hemangiomas. Both agreed on these criteria and findings. A lesion was defined as peripheral or central on the basis of contact with the liver surface. A mass was considered to have nodular peripheral enhancement if CT or MR images showed nodular, cloudlike, and progressive centripetal enhancement during the various phases after the administration of contrast material. Those lesions that demonstrated classic nodular peripheral enhancement and high signal intensity on T2-weighted images were judged as allowing confident imaging diagnosis of hemangiomas. The lesion was considered to have exophytic growth if it produced a focal convex bulge of the liver contour. Capsular retraction was defined by a focal, concave distortion of the hepatic contour. The attenuation of the lesion was judged relative that of the aorta or inferior vena cava on images obtained at each unenhanced and enhanced phase. On T2-weighted images, the intensity of the signal within the lesion was compared with that of cerebrospinal fluid.
In patients who underwent serial imaging, the size stability of the lesions was assessed. None of the patients received treatment, such as chemotherapy or local ablation techniques, prior to transplantation that would likely have influenced the size of a focal hepatic lesion. In the 14 patients in whom total hepatectomy specimens were available, the exact number of the hepatic hemangiomas was determined. Specific and detailed imaging criteria were sought and recorded for each lesion, patient, and study. Finally, we compared our findings from retrospective review of CT and MR images with the diagnosis rendered at the time of the original image interpretation.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The gross pathologic appearance of the hemangiomas was reviewed in the 17 resected specimens. All except one were described as well-circumscribed, spongy, red-to-tan vascular lesions. Variable amounts of necrosis, fibrosis, and thrombosis were noted within and around the hemangiomas. Some hemangiomas were described as hyalinized with extensive fibrosis and obliteration of vascular channels. One lesion was described as being only a fibrotic scar in an area that previously demonstrated the imaging characteristics of a hemangioma, as described later.
Of the 18 hemangiomas with histopathologic proof, 13 were identified at retrospective review of the CT or MR images. The same 13 lesions were identified by both readers at both review sessions (before and after being informed of the lesion findings at pathologic examination). The five hemangiomas not detectable at imaging had a maximum diameter of 2, 10, 10, and 25 mm (one lesion size not specified). The location was subcapsular in three and not specified in two.
Of the 21 hemangiomas, the original imaging interpretations described the same 16 focal liver lesions, as were seen at our retrospective review, but none of the five small hemangiomas, which were also missed at retrospective review. A confident diagnosis of hemangioma was made at the original CT or MR interpretation for 11 hemangiomas (Fig 1), while three lesions were believed to be indeterminate and in need of further evaluation. Two lesions were interpreted incorrectly as being suggestive of HCC. One patient had both a hemangioma (8-cm diameter) and an HCC (3 cm); both were correctly diagnosed at prospective and retrospective CT (Fig 2).
|
|
|
|
|
|
|
|
|
The evolution of the hemangiomas during an extended period was followed at CT for 10 patients with a total of 10 hemangiomas. Seven hemangiomas became smaller, and five subcapsular lesions developed overlying retraction of the hepatic capsule (Fig 4). In one patient, a lesion with a definite CT characteristic of hemangioma progressively diminished in size during 6 years from a diameter of 3 cm to an indistinct fibrotic scar of few millimeters in width and had complete loss of nodular peripheral enhancement (Fig 5). At explantation, only a fibrous remnant was identified at the site of the prior hemangioma. The size of these seven hemangiomas was 3–10 cm in diameter at the time of diagnosis, and they decreased in size by a mean of 4.8 mm/y (range, 2.7–10.9 mm/y).
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The incidence of hemangioma in the general population varies in published reports from 0.4% (11) to 20% (3), the latter resulting from a thorough prospective search of the liver in an unselected autopsy series. However, Dodd et al (2) found only nine hemangiomas among 508 explanted cirrhotic livers; only three of these were visualized at nonhelical CT, and all were regarded as atypical in appearance. Liver specimens demonstrated extensive fibrosis within and surrounding the hemangiomas. These authors (2) speculated that the hemangiomas may have been obliterated by the process of fibrosis in cirrhosis.
Our experience is similar in some respects, but we believe that we have demonstrated the natural history of hemangiomas within the cirrhotic liver and that optimal CT and MR imaging can accurately depict hemangiomas in many cases. In patients with hemangiomas that are followed up for extended periods, we were able to demonstrate progressive diminution in size and loss of some identifying characteristics, such as nodular peripheral enhancement and isoattenuation to blood vessels. In patients with larger hemangiomas or perhaps less extensive cirrhosis, we were able to recognize reliable CT and MR findings of hemangiomas in the majority of our patients, both prospectively (original interpretation) and retrospectively.
The incidence or prevalence of hemangioma in our large population of patients with advanced cirrhosis is impossible to calculate because our search methods required the pathologist and/or radiologist to make a specific diagnosis of hemangioma in their transcribed interpretations. Anecdotally, we have noted a steady increase in the frequency with which we have diagnosed hemangioma in this setting during the past several years.
We believe that our results are largely attributable to improved CT and MR protocols and equipment. In particular, the ability to study the hemodynamics and physiology of focal hepatic lesions with multiphase CT and MR imaging is critical to the accurate diagnosis of focal lesions in the cirrhotic liver. Hemangiomas are typically isoattenuating to blood vessels at nonenhanced CT, do not enlarge over time (12), and demonstrate peripheral nodular (13) and progressive sustained enhancement that remains isoattenuating to blood vessels. The CT appearance of HCC, while variable, is most commonly hypoattenuating to liver (hyperattenuating to blood) at nonenhanced CT, transiently hyperattenuating at HAP CT, then quickly washing out or unenhancing to be iso- to hypoattenuating at PVP CT, and hypoattenuating to liver at delayed CT (8).
MR imaging is likely to offer some advantages over CT in this challenging setting, although our own experience is somewhat limited. Because blood flow to the cirrhotic liver is diminished, contrast enhancement of the cirrhotic liver at CT is often suboptimal (14). Because MR imaging has greater sensitivity to small differences in contrast enhancement and because several fast MR sequences can be used to track the passage of a small, tight bolus of contrast material, MR images may show the characteristic enhancement patterns of hemangiomas and HCC better than CT images. Additional MR sequences, including heavily T2-weighted sequences, are also useful (15).
We were interested in the observation of capsular retraction developing over the five hemangiomas that regressed. Our findings are contrary to those of Soyer et al (16), who reported that capsular retraction adjacent to a hepatic tumor appeared to be a specific sign for malignant hepatic neoplasms. Vilgrain et al (17) and De Caralt et al (10) have both recently reported similar cases of hemangioma with capsular retraction.
Small hemangiomas and fibrotic or hyalinized hemangiomas are likely to remain difficult to diagnose (17–20). Some of the small (<1-cm) hemangiomas that we discovered at gross inspection of resected livers will probably be rarely diagnosed at CT and uncommonly diagnosed at MR imaging. Kim et al (21) recently reported that small (<3-cm) hemangiomas and malignant hepatic lesions can be difficult to distinguish at multiphase helical CT. Bright uniform enhancement of both types of lesions was common, and isoattenuation with blood vessel was judged inconsistently in their multireader study. The combination of nonenhanced and dual-phase enhanced imaging (HAP and PVP) achieved the greatest sensitivity and specificity.
Our study was limited by its retrospective nature and by the fact that our reviewers knew the diagnoses of hemangiomas and cirrhosis. Nevertheless, the prospective interpretation of the images was also accurate; we believe that the improved detection in our retrospective review was partly attributable to the use of strict and well-recognized CT and MR criteria for diagnosing hemangioma.
In summary, even within the cirrhotic liver, larger hemangiomas can usually be diagnosed confidently by using optimized CT and MR techniques. With progressive cirrhosis, however, hemangiomas decrease in size, became more fibrotic, and are more difficult to recognize radiologically and pathologically. Hemangiomas should be diagnosed only if they meet strict CT or MR criteria or if they have done so at prior imaging evaluation. Capsular retraction over a focal hepatic lesion does not negate the diagnosis of hemangioma.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, G.B., M.P.F.; study concepts and design, M.P.F.; literature research, G.B.; clinical studies, G.B., M.P.F., A.B.; data acquisition and analysis/interpretation, all authors; manuscript preparation, all authors; manuscript definition of intellectual content, editing, and revision/review, M.P.F.; manuscript final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. H. Kim, J. M. Lee, J. Y. Lee, J. K. Han, S. K. An, C. J. Han, K. H. Lee, S. S. Hwang, and B. I. Choi Value of Contrast-Enhanced Sonography for the Characterization of Focal Hepatic Lesions in Patients with Diffuse Liver Disease: Receiver Operating Characteristic Analysis Am. J. Roentgenol., April 1, 2005; 184(4): 1077 - 1084. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J A Guthrie Cirrhosis and focal liver lesions: MRI findings Imaging, September 1, 2004; 16(4): 351 - 363. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. H. Bryant, M. J. Blomley, T. Albrecht, P. S. Sidhu, E. L. S. Leen, R. Basilico, J. M. Pilcher, L. H. Bushby, C. W. Hoffmann, C. J. Harvey, et al. Improved Characterization of Liver Lesions with Liver-Phase Uptake of Liver-specific Microbubbles: Prospective Multicenter Study Radiology, September 1, 2004; 232(3): 799 - 809. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. M. Fennessy, K. J. Mortele, T. Kluckert, A. Gogate, S. Ondategui-Parra, P. Ros, and S. G. Silverman Hepatic Capsular Retraction in Metastatic Carcinoma of the Breast Occurring with Increase or Decrease in Size of Subjacent Metastasis Am. J. Roentgenol., March 1, 2004; 182(3): 651 - 655. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Vilgrain, F. Uzan, G. Brancatelli, M. P. Federle, M. Zappa, and Y. Menu Prevalence of Hepatic Hemangioma in Patients with Focal Nodular Hyperplasia: MR Imaging Analysis Radiology, October 1, 2003; 229(1): 75 - 79. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Brancatelli, R. L. Baron, M. S. Peterson, and W. Marsh Helical CT Screening for Hepatocellular Carcinoma in Patients with Cirrhosis: Frequency and Causes of False-Positive Interpretation Am. J. Roentgenol., April 1, 2003; 180(4): 1007 - 1014. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Teefey, C. C. Hildeboldt, F. Dehdashti, B. A. Siegel, M. G. Peters, J. P. Heiken, J. J. Brown, E. G. McFarland, W. D. Middleton, D. M. Balfe, et al. Detection of Primary Hepatic Malignancy in Liver Transplant Candidates: Prospective Comparison of CT, MR Imaging, US, and PET Radiology, February 1, 2003; 226(2): 533 - 542. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.-J. Jang, T. K. Kim, H. K. Lim, S. J. Park, J. S. Sim, H. Y. Kim, and J.-H. Lee Hepatic Hemangioma: Atypical Appearances on CT, MR Imaging, and Sonography Am. J. Roentgenol., January 1, 2003; 180(1): 135 - 141. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
