| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology (G.B., M.P.F., L.G.) and Starzl Transplantation Institute (B.I.C.), University of Pittsburgh Medical Center, 200 Lothrop St, Pittsburgh, PA 15213. Received April 10, 2001; revision requested May 17; revision received July 11; accepted July 17. Address correspondence to M.P.F. (e-mail: federle@pitt.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Clinical, pathologic, and imaging findings were retrospectively evaluated in 39 patients with HCC in noncirrhotic liver. Helical multiphasic CT scans obtained with 125 mL of contrast medium at a rate of 4 or 5 mL/sec were reviewed for morphologic features such as tumor size, margins, and hemorrhage and degree of enhancement.
RESULTS: All patients (25 men, 14 women; mean age, 61 years) were U.S. residents; none had an Asian surname. Twenty-four patients (62%) had no identifiable risk factors; 34 (87%) were symptomatic. HCC was proved and cirrhosis excluded with biopsy in all cases. HCC was moderately (n = 32) or well (n = 6) differentiated in 97% of cases and poorly differentiated in one. Serum 
-fetoprotein level was elevated in 26 patients. Large tumors (mean diameter, 12.4 cm) were depicted at CT in all cases. Thirty-two patients had a solitary or dominant mass. At CT, tumor margins were well defined in 21 patients, with a lobulated surface in 33. Calcifications were depicted in 11, hemorrhage in 10, fat in four, dilated intrahepatic bile ducts in 17, and abdominal lymphadenopathy in eight. In 38 patients, tumors were heterogeneous with areas of necrosis. HCC was hypoattenuating on nonenhanced images in 34, heterogeneously hyperattenuating at arterial phase in 38, and hypoattenuating at portal phase in 35 patients.
CONCLUSION: HCC developed in the absence of cirrhosis or known risk factors and typically appeared as a large symptomatic hepatic tumor with clinical, laboratory, and CT features that distinguish it from most other hepatic masses.
Index terms: Liver neoplasms, 761.323 • Liver neoplasms, CT, 761.12114, 761.12115 • Liver neoplasms, diagnosis, 761.12114, 761.12115
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Much less is known about the radiologic characteristics of HCC in the noncirrhotic liver. Moreover, to our knowledge, no study with a series as large as ours has reported the helical CT characteristics of this tumor. The purpose of our investigation was to review the clinical, pathologic, and helical multiphasic CT findings in patients with HCC in a noncirrhotic liver.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
All patients had biopsy analysis of both the tumor and the non–tumor-bearing liver. All biopsy specimens were judged sufficient to diagnose or exclude cirrhosis. Percutaneous core needle biopsy specimens were obtained by using an 18-gauge biopsy device (Maxcore; Bard, Covington, Ga) and usually provided sufficient tissue to analyze 7–10 portal tracts. Pathologic evidence of cirrhosis included evidence of regenerative nodules partitioned by fibrotic septa. The pathology reports were reviewed retrospectively by a single investigator (G.B.) for evidence of fibrosis, inflammation, and tumor, along with the degree of tumor differentiation.
Histologic proof of HCC was by means of resection of the tumor in eight patients and biopsy in 31 (percutaneous core and fine needle biopsy in 30, surgical biopsy in one).
All 39 patients had abdominal helical CT evaluation that included nonenhanced and contrast material–enhanced imaging through the liver, including both hepatic arterial phase and portal venous phase imaging, with delays of 25–30 seconds and 60–70 seconds, respectively, after initiation of the bolus of intravenous contrast material. All patients received intravenous contrast material (ioversol, Optiray 350; Mallinckrodt, St Louis, Mo) administered at a rate of 4 or 5 mL/sec and a volume of 125 mL with a power injector (model OP 100; Medrad, Pittsburgh, Pa). Collimation and reconstruction interval were 5 mm, pitch was 1.5–1.7, and gantry rotation was 2 seconds. All scans were obtained with a HiSpeed Advantage or CTI model (GE Medical Systems, Milwaukee, Wis).
In our institution, all patients with abnormal liver function, known risk factors for HCC or chronic liver disease, or a liver mass of unknown cause are scanned with a multiphasic CT protocol. All patients in this series underwent abdominal ultrasonography (US) before CT, and all had a hepatic mass identified on sonograms.
Image Analysis
Images were reviewed on a picture archiving and communication system workstation (Impax RS3000 1K review station; AGFA Technical Imaging Systems, Richfield Park, NJ) by three experienced abdominal radiologists (G.B., M.P.F., L.G.). Because we were not attempting to determine the accuracy of CT diagnosis, we accepted consensus interpretation of CT findings.
CT findings were tabulated and included the number, distribution, and bidirectional maximum diameter of the tumors and description of their margins (well defined or ill defined) and surfaces (smooth or lobulated). A tumor capsule was defined as a thin curvilinear border that surrounded at least half the tumor and had a distinct attenuation difference. The overall attenuation of the tumor was defined relative to the liver during the same phase (hepatic arterial or portal venous) of imaging and was judged to be homogeneous or heterogeneous. The presence and distribution of calcifications were recorded. Nonenhancing areas with attenuation similar to that of gallbladder contents were regarded as cystlike, representing necrosis. Hemorrhage was regarded as amorphous fluid with attenuation higher than that of the nonenhanced liver. The presence of tumor fat (ie, lipid) was evident when the tumor components had attenuation coefficients that were lower than those of water and/or less than those of bile or urine on the nonenhanced CT images. Intrahepatic dilated bile ducts were defined as hypoattenuating linear bands parallel to branches of the portal vein, better seen on portal venous phase images. Tumor thrombus was defined as a distention of the hepatic and/or portal vein lumen, by enhancement of the thrombus, or by proximity of tumor and the thrombosed vessel. Encasement was deemed present when the vessel was not distinctly visible, but was surrounded by the tumor mass.
Upper abdominal lymphadenopathy was diagnosed when ovoid or round extravisceral masses were identified that were 2 cm or more in diameter and had attenuation less than or equal to that of skeletal muscle.
Pertinent clinical information that was recorded (G.B.) included the age and sex of the patients, presence of known risk factors (alcohol abuse, viral hepatitis, hemochromatosis, 1-antitrypsin deficiency, primary biliary cirrhosis, primary sclerosing cholangitis) for liver injury, and physical signs and symptoms potentially attributable to the hepatic tumor. The level of serum -fetoprotein was recorded before any treatment. All patients were questioned about alcohol consumption, and all had serum testing for viral hepatitis B and C.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
CT demonstrated an obvious tumor in all 39 patients. The HCC was solitary in 13 patients (Fig 1); 19 patients had a dominant mass with smaller satellite lesions (Fig 2). Seven patients had multiple masses without a dominant lesion (Fig 3). The diameter of the largest mass averaged 12.4 cm (range, 2–23 cm). The dominant mass or largest cluster of HCC was in the right lobe in 21 patients and in the left lobe in 12 and was multifocal in six.
|
|
|
|
|
|
|
|
Calcification was present within tumors in 11 (28%) patients and was peripheral in 10 (Fig 4), central in one; 28 patients had no calcifications within the tumor. Necrotic nonenhancing portions of tumor were noted in 38 (97%) of the 39 patients, and 38 patients had heterogeneous areas within all or most of the tumors. Hyperattenuating areas compatible with hemorrhage were noted within tumors in 10 patients (26%). Low-attenuating areas compatible with a fat component were evident in tumors in four patients (10%) (Fig 4). Dilated intrahepatic bile ducts were present in 17 patients (44%) (Fig 2), indicative of HCC obstruction of more central ducts. Tumor thrombus was present within the portal vein in six patients (Fig 1) and within the hepatic vein in three patients. Encasement or obliteration of the portal vein was noted in two patients and of a hepatic vein in two patients. Two patients had tumor thrombus both in the portal and the hepatic veins, and one patient had tumor thrombus in the portal vein and encasement of the hepatic veins.
|
|
On nonenhanced CT images, the HCC was predominantly hypoattenuating to liver in 34 (87%) patients, hyperattenuating in two (Fig 3), and isoattenuating in three. In 38 (97%) of the 39 patients, the HCC lesions were heterogeneously hyperattenuating on hepatic arterial phase CT images (Fig 1), whereas one was hypoattenuating. However, on portal venous phase CT images, the tumors in 35 patients (90%) became hypoattenuating (Figs 1, 2), and four patients had isoattenuating tumor (Fig 3).
The pathology reports indicated that the HCC was interpreted as well differentiated in six patients, moderately differentiated in 32 patients, and poorly differentiated in only one patient. Histopathologic examination of the nonneoplastic liver in 21 patients showed varying degrees of periportal inflammation and scattered fibrosis in 19, often with mild to moderate steatosis. Two specimens were interpreted as normal liver. All 39 patients had biopsy proof of "no cirrhosis," but detailed reports from the biopsies of the nonneoplastic liver were not available in 18 cases.
The average level of serum -fetoprotein was 18,841 µg/L (range, 1–254,150 µg/L). In 13 patients, the level was 20 µg/L or less (normal for our laboratory); in 26 patients (67%) the -fetoprotein was abnormally elevated.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
HCC may be much less familiar to North American radiologists who are likely to see it as a disease with much less prevalence, a different set of clinical circumstances, and different treatment options and prognosis. In North America and Western Europe, alcohol has been the most commonly cited cause for HCC (10), although chronic viral hepatitis has emerged as an important cause. In 1995, Nzeako et al (3) reviewed the experience of the Armed Forces Institute of Pathology. They estimated that 40% of the HCC seen in North America occurred in noncirrhotic livers. Excluding fibrolamellar HCC, they reported on 305 such patients, only 32 of whom (10.5%) had evidence of viral hepatitis or alcoholism. Some form of nonspecific hepatitis was evident histologically in 21% of the patients, leaving many patients with no apparent cause for hepatitis or HCC. Therefore, in a minority of patients, HCC appears to arise de novo in an otherwise normal liver.
In our experience, HCC was diagnosed in approximately 180 patients during the study period, 39 of whom (21.6%) did not have cirrhosis. All but two of our 39 patients had histologic evidence of nonspecific liver injury (inflammation, steatosis, fibrosis), but 64% of them had no identifiable risk factors for cirrhosis or HCC.
Several investigators (3,9,11) have consistently reported that HCC occurring in the noncirrhotic liver has several distinguishing features. Patients are younger and are more likely to present with symptoms, to have a single or dominant mass, and to survive liver resection (12). They also have a better prognosis, with a median survival of 2.7 years and with 25% of patients surviving for at least 5 years (9). As with our cases, HCCs are predominantly moderately to well differentiated. Our patients were somewhat older (mean, 61 years), but 87% were symptomatic. Thirty-two (82%) patients had a single or dominant mass. We did not attempt to determine surgical management or survival in this study.
The radiologic appearance of HCC in the noncirrhotic liver has received little attention. Winston et al (11) described the MR features of 25 patients with HCC in noncirrhotic liver, comparing these with 11 patients with HCC in cirrhotic liver. They noted that the tumors were significantly larger in the noncirrhotic group, consistent with reports from pathologists, including Yamashita et al (13). They noted a "scar" as being present in 50% of the HCC in noncirrhotic livers and noted this as a potential source of diagnostic difficulty in distinguishing such "conventional" HCC from fibrolamellar HCC (11).
The HCCs in noncirrhotic livers that we report were large masses (mean, 12.4 cm), predominantly solitary or dominant with satellite lesions (32 [82%] patients), that were often partially encapsulated (51%), with areas of necrosis (97%) and hemorrhage (26%). Obstruction or invasion of portal or hepatic veins was present in 38% of patients and of the bile ducts in 44%.
Some of these characteristics overlap with those of fibrolamellar HCC, a point emphasized by Winston et al (11). However, we believe that most cases can be differentiated with reasonable certainty on the basis of clinical, laboratory, and imaging findings. Most patients with conventional HCC are male, older (50–70 years), and have positive serum tumor markers such as -fetoprotein and/or des-gamma carboxy prothrombin (PIVKA [protein induced by vitamin K absence]). In patients with HCC and without cirrhosis, CT (or MR imaging) will demonstrate large heterogeneous masses, predominantly in the right hepatic lobe (54% of our patients), with frequent areas of hemorrhage and necrosis, no central scar or calcification, little or no evidence of fibrosis, and frequent invasion of bile ducts and veins. Upper abdominal lymphadenopathy is uncommon, with only 21% (eight patients) in our series, and may be due to reactive hyperplasia (eg, from hepatitis) in some of these patients (four of the eight patients) rather than lymphatic metastases.
Fibrolamellar HCC, on the other hand, occurs typically as a large, sharply defined mass in the left lobe (65%), with a prominent central scar and radiating bands of fibrosis (71%–95%), with calcifications within the scar (40%–68%); rare hemorrhage, necrosis, or fat; and frequent upper abdominal lymphadenopathy (65%) (14–16). In addition, serum tumor markers are rarely elevated in fibrolamellar HCC. We believe that the "scars" that Winston et al (11) described in conventional HCC in noncirrhotic livers were more likely to represent areas of necrosis or hemorrhage, as they had high signal intensity on T2-weighted images.
HCC in the noncirrhotic liver must also be distinguished from other hypervascular tumors, especially hepatocellular adenoma and metastases. Hypervascular metastases are usually encountered in a patient with a known primary tumor, such as neuroendocrine, thyroid, or renal carcinoma, and these are usually multifocal and smaller at the time of diagnosis than the HCC masses we describe. The seven cases of multifocal HCC in our series would be difficult to distinguish from metastases with CT criteria alone.
It may be difficult or impossible to distinguish HCC from adenoma on the basis of imaging findings alone, but adenomas occur almost always in the setting of hepatocellular stimulation from oral contraceptives, anabolic steroids, or abnormal carbohydrate metabolism (glycogen storage disease or diabetes) (17). Moreover, adenomas may rarely undergo malignant transformation, as in one of our cases. For this reason, and also because large adenomas are likely to hemorrhage, these tumors are usually resected.
The cause of HCC in noncirrhotic livers remains uncertain. Only 36% of our patients had evidence of viral hepatitis or excessive alcohol, which can favor the development of HCC even before cirrhosis develops (8). Liver biopsy specimens were not always subjected to polymerase chain reaction analysis for hepatitis B or C, which could have resulted in underdiagnosis in some of our cases. No cause could be established in the remaining 64% of patients. Other hepatic toxins and various conditions (such as 1-antitrypsin deficiency and hemochromatosis) can also predispose to HCC, but these were not confirmed in our series. Some focal hepatic lesions can also lead to HCC. We had one patient with an adenoma that was proved with biopsy and that remained stable in size for 7 years before undergoing malignant transformation to HCC.
As with all retrospective studies, ours has limitations. Specifically, we did not directly compare the patients with HCC in noncirrhotic liver with a matched group of patients with cirrhosis, although we believe that the spectrum of CT findings in cirrhosis and HCC in this setting are already well established.
In summary, HCC can develop in the absence of cirrhosis and even without identifiable risk factors. In this setting, HCC is more likely to manifest as a symptomatic mass in a middle-aged man with abnormally elevated serum tumor markers. The hepatic mass is likely to be a large solitary or dominant mass with lobulated and possibly encapsulated margins. Necrosis and hypervascularity are prominent features, but central scar, fibrosis, and calcification are not, helping to differentiate HCC from fibrolamellar HCC.
|
FOOTNOTES |
|---|
Abbreviation: HCC = hepatocellular carcinoma
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, M.P.F., B.I.C.; data acquisition, G.B.; data analysis/interpretation, G.B., M.P.F., L.G.; manuscript preparation, all authors; manuscript definition of intellectual content and editing, M.P.F.; manuscript revision/review and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. E. Chung, M.-S. Park, Y. N. Park, H.-J. Lee, J. Y. Seok, J.-S. Yu, and M.-J. Kim Hepatocellular Carcinoma Variants: Radiologic-Pathologic Correlation Am. J. Roentgenol., July 1, 2009; 193(1): W7 - W13. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Takayama, M. Makuuchi, M. Kojiro, G. Y. Lauwers, R. B. Adams, S. R. Wilson, H.-J. Jang, C. Charnsangavej, and B. Taouli Early Hepatocellular Carcinoma: Pathology, Imaging, and Therapy Ann. Surg. Oncol., April 1, 2008; 15(4): 972 - 978. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Ding, W.-P. Wang, B.-J. Huang, R.-X. Wei, N.-A. He, Q. Qi, and C.-L. Li Imaging of Focal Liver Lesions: Low-Mechanical-Index Real-time Ultrasonography With SonoVue J. Ultrasound Med., March 1, 2005; 24(3): 285 - 297. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
