| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, Yamanashi Medical University, Nakakoma, Japan (T.I.); the Departments of Radiology (M.P.F.) and Pathology (M.N.), University of Pittsburgh Medical Center, 200 Lothrop St, Pittsburgh, PA 15213; and the Department of Radiology, Spedali Civili Brescia, Italy (L.G.). Received March 22, 1999; revision requested May 3; revision received June 22; accepted July 20. Address reprint requests to M.P.F. (e-mail: federlemp@radserv.arad.upmc.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Multiphasic helical CT was performed in 25 patients with 44 hepatic adenomas. Nonenhanced scans were obtained in all cases, along with hepatic arterial–dominant phase (HAP) and portal venous–dominant phase (PVP) images at 25–28 and 60–70 seconds after intravenous contrast material injection at 3–5 mL/sec. Twelve patients with 24 adenomas also underwent delayed-phase (5–10-minute) CT. Two independent readers retrospectively reviewed each case for the number of detectable lesions in each CT phase, morphologic features of tumors, and degrees of enhancement.
RESULTS: Thirteen patients had solitary adenomas; 12 patients had two or three adenomas. Both observers agreed on the numbers of lesions detected in all cases and in all phases of enhancement. The detection rate for all 44 adenomas per type of examination was as follows: nonenhanced, 86% (38 of 44); HAP, 100% (44 of 44); PVP, 82% (36 of 44), and delayed, 88% (21 of 24). Tumor margins were well defined in 38 adenomas (86%), and the surface was smooth in 42 adenomas (95%). The right hepatic lobe was the only site of adenoma or was a site along with the left lobe in 29 cases (66%). Tumor fat and calcifications were uncommon (three cases [7%] and two cases [5%], respectively). Other than areas of fat, hemorrhage, or necrosis, the adenomas enhanced nearly homogeneously, especially on PVP and delayed-phase scans. Five patients had coexistent hepatic masses, which were focal nodular hyperplasia (n = 3) or hepatocellular carcinoma (n = 2).
CONCLUSION: Hepatic adenomas often have characteristic features at multiphasic CT that may allow their distinction from other hepatic masses.
Index terms: Liver neoplasms, CT, 761.12113, 761.12114, 761.12115, 761.12119 • Liver neoplasms, diagnosis, 761.3192, 761.3198, 761.323, 761.35
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The purpose of our study was to review our experience with a large number of cases of hepatic adenoma in an attempt to define characteristic CT and pathophysiologic characteristics that might lead to more confident diagnosis and management.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
All 25 patients had undergone preoperative multiphasic CT investigation and had histologic proof of hepatic adenoma. Complete tumor resection (all lesions) was performed in 14 patients by means of hepatic lobectomy (n = 12) or orthotopic liver transplantation (n = 2). In the remaining 11 patients, specimens were obtained from at least one hepatic lesion by percutaneous (n = 9) or laparoscopic (n = 2) needle biopsy. Percutaneous biopsy included histologic core-needle and aspiration cytologic analysis in all nine cases. These 11 patients underwent additional CT and clinical follow-up after biopsy for a mean of 25 months (range, 9–57 months).
Of the 25 patients, 21 were women and four were men, with the age at diagnosis ranging from 27 to 80 years (mean age, 41 years). Twelve of the women had a history of oral contraceptive use (range of use, 3 months to 20 years; mean use, 5.5 years). One woman had glycogen-storage disease (glycogenosis). None of the men had a known predisposing condition, such as use of anabolic steroids. For eight of the patients, the hepatic adenoma was an unexpected finding on CT scans obtained for unrelated reasons, while other patients had abdominal symptoms or signs, including pain (acute [n = 3] or chronic [n = 8]), palpable mass (n = 2), or abnormal liver function (n = 4). No patients had clinical evidence of viral hepatitis or cirrhosis.
CT scans were obtained with a commercially available helical CT scanner (HiSpeed Advantage; GE Medical Systems, Milwaukee, Wis). All patients received a monophasic (3–5-mL/sec) injection of either iothalamate meglumine (Conray 60; Mallinckrodt Medical, St Louis, Mo) or ioversol (Optiray 320; Mallinckrodt Medical). A mean of 2 mL per kilogram of body weight (100–150 mL) was injected intravenously by means of a power injector (Medrad, Pittsburgh, Pa).
The multiphasic CT protocol included acquisition of nonenhanced hepatic sections, followed by hepatic arterial–dominant phase (HAP) and portal venous–dominant phase (PVP) images in all patients. In addition, 12 patients underwent delayed-phase hepatic imaging 5–10 minutes following intravenous contrast medium injection. HAP images were obtained after a scanning delay of 25–28 seconds, while PVP images were obtained after a scanning delay of 60–70 seconds. The CT section thickness was 5–7 mm, and the table incrementation was 7.5–10 mm/sec, with the pitch adjusted to allow scanning of the complete liver within one breath hold.
The CT scans were reviewed retrospectively and independently by two radiologists (T.I. and L.G.), with knowledge of the diagnosis of hepatic adenoma but without knowledge of the specific number of lesions or clinicopathologic findings in any patient. In cases of interobserver disagreement, final decisions were reached by consensus. Kappa statistics for interobserver agreement were not calculated, because disagreements were considered relatively few and minor, such as the degree of and homogeneity of contrast medium enhancement of the lesions. Both observers agreed on the number of lesions in all cases.
Tumor characteristics that were evaluated included the number, sizes, and locations of the tumors. The nature of the tumor margin (sharply or ill-defined) and surface (smooth or lobulated) was noted. A tumor capsule was judged to be present if a thin, curvilinear structure surrounded all or part of the tumor and clearly demarcated it from the normal part of the liver. The presence and distribution of calcifications were noted, along with evidence of acute hemorrhage (fluid in or around the tumor that was hyperattenuating to liver on nonenhanced images). Necrosis or subacute hemorrhage was deemed evident by a fluid collection intermediate in attenuation (between the attenuation of water and that of blood). The presence of tumor fat (lipid) was evident by tumor components with attenuation below that of water (<0 HU) and/or visibly less than the attenuation of bile or urine on the nonenhanced CT scans.
The attenuation of the tumor was judged relative to that of the surrounding liver on nonenhanced images, as well as on images obtained in each phase of contrast medium enhancement (HAP, PVP, and delayed phase). Because many tumors were heterogeneous on nonenhanced scans, presumably because of fat, hemorrhage, necrosis, and/or calcification, we attempted to judge the enhancement characteristics of the remaining portions of the tumor as homogeneous or heterogeneous.
The exact number and nature of the hepatic tumors cannot be determined because only two patients underwent liver transplantation and none underwent autopsy. The total number of lesions detected with consensus CT interpretation of images obtained in each phase of the multiphasic CT examination was used to judge the detection and characterization of hepatic lesions at each phase of imaging.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
The morphologic features of size, number, surface characteristics, and capsule were confirmed in all 14 patients who had complete tumor excision and pathologic analysis.
Calcifications were seen at CT in only two lesions (5%) and appeared as large, coarse, calcific opacities within areas of old hemorrhage or necrosis, which were confirmed at pathologic study in both cases. There were no cases of calcification seen at pathologic study and missed at CT, but no special attempts were made by the pathologist (M.N.) to detect calcification.
Necrosis or hemorrhage was identified at CT in 11 of the 44 lesions (25%) in 10 patients. In nine of these lesions (six patients), areas of acute hemorrhage were diagnosed and consisted of heterogeneous, hyperattenuating fluid on nonenhanced images (Fig 2). Two of these 10 patients also had CT evidence of subcapsular or intraperitoneal hemorrhage. All of these findings were confirmed at surgery and/or histopathologic analysis.
|
|
In no cases did CT or histopathologic analysis demonstrate a central or eccentric fibrotic scar or radiating bands of fibrosis.
The attenuation of the adenoma relative to the underlying liver varied with the composition of the tumor and that of the liver, as well as the phase of contrast medium enhancement. The attenuation patterns are summarized in Table 1. In patients with a fatty liver (steatosis; n = 3), the adenoma was hyperattenuating to liver at all phases of contrast medium enhancement and on nonenhanced images.
|
|
|
|
|
|
|
|
Delayed-phase imaging, performed in 12 patients (with 24 total, 16 "noncomplicated" adenomas), revealed no tumors beyond those seen on HAP or PVP images, although three tumors that were isoattenuating on PVP images became hypoattenuating on delayed images (Table 2).
In five lesions (11%), dilated, early-draining venous structures were demonstrated on HAP and PVP images (Fig 3), but only one lesion had a dilated feeding artery or vascular structures visible within the adenoma (Fig 2).
The detection rate for all 44 lesions during each phase of imaging was as follows: nonenhanced phase, 86% (38 of 44); HAP, 100% (44 of 44); PVP, 82% (36 of 44); delayed phase, 88% (21 of 24).
Of the 25 patients with hepatic adenomas, five had coexistent nonadenomatous tumors. Three patients had focal nodular hyperplasia (FNH), and two had HCC. The FNH lesions were 30, 40, and 60 mm in diameter and were uniformly hyperattenuating on HAP images and isoattenuating on PVP and delayed images (Fig 4). Two of the FNH lesions had central scars that were hypoattenuating to the FNH on HAP and PVP images and that were hyperattenuating on delayed images. These findings were thought to be diagnostic of FNH and were confirmed at histopathologic analysis after percutaneous biopsy of two lesions and resection of one lesion.
In the two patients who developed HCC, malignant transformation of the adenoma was first suggested at CT. One patient had no elevation of the 
-fetoprotein level, while the other patient had substantial elevation of the -fetoprotein level (1,567 ng/mL [1,567 µg/L]). HCC was suggested by the presence of tumor heterogeneity that was not easily explained as obvious hemorrhage or fat. One of the HCCs had shown growth and increasing heterogeneity over 7 years of observation following the initial biopsy proof of an adenoma (Fig 5).
|
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Helical CT, especially with multiphasic scanning, allows more accurate detection and characterization of focal hepatic lesions (11–14). The ability to acquire separate series of breath-hold images during the HAP and PVP adds a temporal hemodynamic component to the morphologic depiction of tumors. Nonenhanced and delayed contrast-enhanced images can also provide important diagnostic clues in some cases, such as cavernous hemangiomas, which tend to remain isoattenuating with blood vessels, or fibrotic tumors, which tend to demonstrate delayed persistent enhancement (17–19).
On the basis of helical, multiphasic CT observations and pathophysiology, a group of primary and metastatic hepatic tumors have been classified as hypervascular. The primary hepatic tumors that are often hypervascular include conventional and fibrolamellar HCC (12–14), and the benign lesions include hepatic (hepatocellular) adenoma and FNH (8–10,13). Because management of these various tumors is radically different, confident preoperative diagnosis is essential but has been regarded as very difficult on the basis of published radiologic observations.
Because of the retrospective nature of our study and the fact that the radiologists performing this analysis of morphologic features knew the diagnosis was adenoma, we cannot determine the accuracy of CT in the diagnosis of hepatic adenoma. While our investigation was not designed to test our ability to diagnose and distinguish hepatic adenomas, we believe that these tumors may have common features on multiphasic CT scans that should allow us to improve our diagnostic ability. Because adenomas consist almost entirely of uniform hepatocytes and a variable number of Kupffer cells (1–4), it is not surprising that most of the adenomas in our series were nearly isoattenuating to normal liver on nonenhanced, PVP, and delayed images. Adenomas were almost always hyperattenuating to a fatty liver and were at least transiently hyperattenuating to normal liver during HAP. The lesions enhanced completely or nearly homogeneously, except for areas of necrosis, hemorrhage, or focal fatty degeneration.
Adenomas were sharply marginated (86%), nonlobulated (95%), sometimes encapsulated (25%), and rarely calcified (5%). Hemorrhage was found in 25% of the adenomas (11 of 44) and in 40% of the patients (10 of 25) in our experience. The prevalence of hemorrhage would undoubtedly be influenced by whether CT was being used for screening examinations in a population that was asymptomatic but "at risk" for adenoma, such as patients with type 1 glycogen-storage disease, versus a population with acute abdominal pain.
Lipid or fat deposition within an adenoma was noted in only 7% of lesions in this CT series. This is in contrast to findings of the magnetic resonance (MR) imaging appearance of adenomas. From 35% to 77% of hepatic adenomas have MR imaging characteristics of lipid content, such as hyperintensity on both T1- and T2-weighted images and selective signal intensity dropout on opposed-phase chemical shift MR images (20–22). These findings have been correlated with the variable lipid content of adenomas. Adenomas also appear heterogeneous on MR images in 52%–93% of cases, a finding attributed to both fat and hemorrhage in these lesions (20–22). These apparent discrepancies with our CT findings can be attributed to the improved contrast resolution of MR imaging over CT.
Because conventional HCC also may contain lipid or fat and may resemble adenoma at MR imaging (23), it can be difficult to make this distinction with imaging alone. Other criteria, such as interval growth or elevated serum -fetoprotein levels, favor a diagnosis of HCC. As noted in one of our cases, an adenoma can undergo malignant change to HCC even after years of a stable appearance (5–7). Most cases of HCC occur within a cirrhotic liver, and CT signs of cirrhosis and portal hypertension are usually evident, along with other signs of malignancy, which include portal or hepatic venous invasion, lymphadenopathy, or metastases. Biopsy or even resection may be necessary for diagnosis. Histologic distinction between a hepatic adenoma and a well-differentiated HCC can be challenging, particularly in percutaneous biopsy specimens (5–7).
We believe that hepatic adenomas can usually be distinguished from fibrolamellar HCC, a malignant tumor that also occurs in young adults without underlying cirrhosis. We recently examined 31 patients with fibrolamellar HCC (14). CT demonstrated these tumors as large, heterogeneous, lobulated masses with large, central or eccentric scars and radiating fibrous septa. Calcifications were present in 68% of tumors, and the areas of hypervascularity were heterogeneous in all cases. In addition to these distinguishing characteristics, abdominal lymphadenopathy was noted in 65% of our patients with fibrolamellar HCC (14).
Adenoma and FNH have several features in common, which include their tendency to occur in young women, and may share certain CT features (8–10). FNH does not undergo malignant degeneration, nor is it likely to bleed (2,4,5,8–10). Therefore, unlike adenoma or other hypervascular masses, it rarely requires therapy (1–10). At multiphasic, helical CT, FNH appears as a homogeneous, markedly hypervascular mass on HAP scans, with a central scar that is hypoattenuating on early, enhanced CT scans and hyperattenuating on delayed CT scans (with or without thin, radiating septa) (Fig 4). The FNH is typically nearly isoattenuating to normal liver on noncontrast, PVP, and delayed images but enhances during the arterial phase to a degree approaching the enhancement of the aorta (8–10,13).
FNH may demonstrate small central and septal arteries, along with early draining veins (8–10,13). Choi and Freeny (13) examined a series of patients with FNH who underwent triphasic helical CT and reported both characteristic and atypical features. It may be that certain CT features they considered atypical of FNH were actually a consequence of the occurrence of FNH in a fatty liver, which results in "persistent hyperdensity [hyperattenuation] on portal venous phase CT [scans]," a capsulelike rim, and visible early draining veins.
Choi and Freeny (13) identified a central scar in 10 of their 12 cases of FNH and noted a central artery within the scar in eight of the 10 cases. We believe that these may constitute the most reliable CT features for distinguishing FNH from hepatic adenoma. In our series, we detected a dilated feeding artery and vessels within the adenoma in only one case and never demonstrated a central scar. We did recognize early draining veins in five lesions (11%) (Fig 3), and Choi and Freeny (13) observed early draining veins in three of their 12 cases of FNH. Rather than regard this as an atypical finding suggestive of malignancy, we believe this reflects the hypervascular nature of these benign lesions, even though malignant hypervascular masses may also demonstrate neovascularity and early draining veins. With similar reasoning, we do not regard as atypical an adenoma that is hyperattenuating to fatty liver, even on nonenhanced or delayed scans, a finding which was frequent in our series.
In distinguishing between FNH and adenoma, MR imaging may be useful (20–22). MR imaging is even more sensitive than CT in depicting the characteristic central scar of FNH. Nonenhanced MR imaging usually demonstrates near isointensity between FNH and normal liver, and dynamic gadolinium-enhanced images show rapid lesion enhancement and a return to background liver signal intensity, which is analogous to triphasic CT findings (20–22). Delayed, persistent enhancement of the central scar is seen in both CT and MR imaging (8–10,20–22) but is seen probably slightly better at MR imaging because of its higher contrast sensitivity. Hepatic adenoma, on the other hand, demonstrates very different MR imaging characteristics, which include heterogeneity and evidence of fat and hemorrhagic content in most cases (20–22). Biopsy or technetium 99m sulfur colloid hepatic scintigraphy may still be useful in some cases (8–10).
The coexistence of FNH and hepatic adenoma in a patient may seem extraordinary because both lesions are relatively rare (24,25) (Fig 4). However, FNH is thought by some to be a reaction to a disturbance of hepatic vascularity (26). As such, FNH is encountered with increased frequency in the setting of other hypervascular lesions, which include cavernous hemangioma and hepatic adenoma (24–27).
Differentiation of hepatic adenoma from hypervascular metastases may be difficult or impossible on the basis of CT findings alone. A careful search of other organs on the abdominal CT scan is warranted to detect a possible primary tumor (such as a pancreatic islet cell or renal cell cancer), although breast or thyroid carcinoma would probably be more likely in the age group of women most likely to have an hepatic adenoma discovered. Most hypervascular metastases are multiple and will have lesions or portions of lesions that are hypoattenuating to normal liver on nonenhanced, PVP, and delayed images (28,29). MR imaging characteristics are likely to be distinctive as hypervascular metastases, as usually hypointense on T1-weighted images, and as markedly hyperintense on T2-weighted images (30,31). Areas of fat and hemorrhage are rare in hypervascular metastases but are commonly detected with MR imaging in adenomas.
In summary, multiphasic, helical CT can demonstrate findings characteristic of hepatic adenoma. These include the presence of a single mass or multiple masses that may contain areas of fat or hemorrhage but that are otherwise nearly isoattenuating to normal liver on nonenhanced, PVP, and delayed scans. The lesions are moderately hyperattenuating to liver on the HAP images and enhance nearly homogeneously. They are sharply marginated and nonlobulated. Other features such as tumor capsule, early draining veins, and calcification are less common and overlap findings encountered in other hepatic masses. Atypical features such as heterogeneous enhancement may require additional imaging, such as MR imaging, or may require biopsy or even surgical resection in some cases to exclude a malignant neoplasm.
|
Footnotes |
|---|
Author contributions: Guarantors of integrity of entire study, T.I., M.P.F.; study concepts and design, T.I., M.P.F.; definition of intellectual content, T.I., M.P.F.; literature research, T.I., L.G.; clinical studies, T.I., L.G., M.N.; data acquisition and analysis, T.I., L.G.; manuscript preparation and editing, T.I., M.P.F.; manuscript review, M.P.F.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. K. Kim, H.-J. Jang, P. N. Burns, J. Murphy-Lavallee, and S. R. Wilson Focal Nodular Hyperplasia and Hepatic Adenoma: Differentiation with Low-Mechanical-Index Contrast-Enhanced Sonography Am. J. Roentgenol., January 1, 2008; 190(1): 58 - 66. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Grazioli, G. Morana, M. A. Kirchin, and G. Schneider Accurate Differentiation of Focal Nodular Hyperplasia from Hepatic Adenoma at Gadobenate Dimeglumine-enhanced MR Imaging: Prospective Study Radiology, July 1, 2005; 236(1): 166 - 177. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. R. Prasad, H. Wang, H. Rosas, C. O. Menias, V. R. Narra, W. D. Middleton, and J. P. Heiken Fat-containing Lesions of the Liver: Radiologic-Pathologic Correlation RadioGraphics, March 1, 2005; 25(2): 321 - 331. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. C. Faria, R. B. Iyer, A. Rashid, and G. J. Whitman Hepatic Adenoma Am. J. Roentgenol., June 1, 2004; 182(6): 1520 - 1520. [Full Text] [PDF]
|
|
|
|
|
|
P. Attal, V. Vilgrain, G. Brancatelli, V. Paradis, B. Terris, J. Belghiti, B. Taouli, and Y. Menu Telangiectatic Focal Nodular Hyperplasia: US, CT, and MR Imaging Findings with Histopathologic Correlation in 13 Cases Radiology, August 1, 2003; 228(2): 465 - 472. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Grazioli, M. P. Federle, G. Brancatelli, T. Ichikawa, L. Olivetti, and A. Blachar Hepatic Adenomas: Imaging and Pathologic Findings RadioGraphics, July 1, 2001; 21(4): 877 - 892. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. D. Levy and P. R. Ros Invited Commentary RadioGraphics, July 1, 2001; 21(4): 892 - 894. [Full Text] [PDF]
|
|
|
|
|
|
G. Brancatelli, M. P. Federle, L. Grazioli, A. Blachar, M. S. Peterson, and L. Thaete Focal Nodular Hyperplasia: CT Findings with Emphasis on Multiphasic Helical CT in 78 Patients Radiology, April 1, 2001; 219(1): 61 - 68. [Abstract] [Full Text]
|
|
|
|
|
|
L. Grazioli, M. P. Federle, T. Ichikawa, E. Balzano, M. Nalesnik, and J. Madariaga Liver Adenomatosis: Clinical, Histopathologic, and Imaging Findings in 15 Patients Radiology, August 1, 2000; 216(2): 395 - 402. [Abstract] [Full Text]
|
|
|
|
|
|
G. Brancatelli, M. P. Federle, L. Grazioli, and B. I. Carr Hepatocellular Carcinoma in Noncirrhotic Liver: CT, Clinical, and Pathologic Findings in 39 U.S. Residents Radiology, January 1, 2002; 222(1): 89 - 94. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||
