Hepatocellular carcinoma

An unusual case of fibrolamellar carcinoma of the liver developed 5 years after removal of a hepatocellular adenoma in a 14-year-old girl belonging to a family with Carney syndrome. Both tumors were studied by light and electron microscopy, flow cytometry, and comparative genomic hybridization. The first tumor, removed at the age of 9, was a bulky well-circumscribed liver mass composed of large eosinophilic cells with a focal pseudoglandular pattern but without cytologie atypia or sclerosis. A diagnosis of hepatocellular adenoma was rendered. Five years later, another hepatic tumor was removed from the right lobe. Microscopic examination revealed polygonal cells, each with a large amount of eosinophilic cytoplasm and a round nucleus with a conspicuous nucleolus. These cells were arranged in nests and strands and separated by bands of dense fibrous tissue, leading to a diagnosis of fibrolamellar carcinoma. Comparative genomic hybridization analysis revealed no genetic alteration in the adenoma; however, several chromosomal aberrations (loss of chromosome regions 1 p and 4p and gains of chromosome regions 6q, 13q, and Xq) were detected in the fibrolamellar carcinoma. To our knowledge, this is the first report of an association between hepatocellular adenoma and fibrolamellar carcinoma.

Hepatocellular adenoma (HCA) is the most prevalent benign neoplasm of the liver and is usually found in young women taking oral contraceptives.1 It is rare in childhood, accounting for less than 2% of pediatric liver tumors,1 and may arise spontaneously or in association with inherited conditions such as Fanconi anemia, glycogenose type I, or familial adenomatous polyposis.2

Fibrolamellar carcinoma (FLC) accounts for less than 10% of all hepatocellular carcinomas (HCC) in the general population but for 30% of HCCs in patients younger than 20 years of age. It occurs almost exclusively in adolescents and young adults without gender preference. Little is known about the etiology and pathogenesis of FLC. Unlike the more common types of HCC, there is no underlying chronic disorder, such as cirrhosis, fibrosis, viral infection, or metabolic abnormalities.

Precursor lesions of FLC are not known.1 Some authors have suggested an association between focal nodular hyperplasia3 and FLC, but this association has not yet been confirmed.

Rare cases of association between common types of HCC and HCA have been reported,4 as have infrequent instances of dysplastic foci within an adenoma. However, an association between an FLC and an HCA has not yet been reported in the literature.

Here, we report a case of an FLC in a 14-year-old girl. This tumor was discovered 5 years after removal of an HCA. Histologic, immunohistochemical, and electron microscopic, and comparative genomic hybridization analyses were performed. To our knowledge, this is the first report of such a tumor.

REPORT OF CASE

In June 1994, a 9-year-old girl presented with right upper quadrant abdominal pain. Computed tomography and magnetic resonance imaging scans revealed a 10-cm hepatic mass extending from segments VI and VII. She had no previous known liver disease, and her neonatal period and infancy were normal. She suffers from asthma, has a known allergy to dust, and has had recurrent episodes of parotitis. As yet, no endocrine abnormalities or tumors in other sites have been found.

The patient's mother has Carney syndrome, with lentigines, cardiac myxoma, fibroadenoma of the breast, and a fibroepithelial polyp of the external auditory canal. No increased level of sexual hormones was detected during this woman's pregnancy. The patient's brother has had a nevus removed. Both grandfathers were reported to have developed tumors of unknown location.

The mass was enucleated, and a diagnosis of HCA was made. The patient was re-evaluated on a yearly basis by ultrasound assessment. Her course was uneventful until 14 years of age, when the girl presented with fatigue and a large hepatic mass, again of the right liver lobe. The tumor was surgically resected, and a diagnosis of FLC was made.

PATHOLOGIC FINDINGS

Morphologic and Immunohistochemical Findings

Tissue for histologie examination was fixed in formalin, embedded in paraffin, cut into 5-µm sections, and stained with routine hematoxylin-eosin, periodic acid-Schiff with and without diastase predigestion, Gomori and Novotny stains for reticulin fibers, chromothrope-anilin-blu, and Sirius red.

For immunohistochernical studies, formalin-fixed, paraffin-embedded tissue was stained using the avidin-biotin peroxidase complex technique and heat-induced epitope retrieval buffer. The antibodies used included [alpha]-fetoprotein (1:4000, Dako, Glostrup, Denmark), fibrinogen (1: 200000, Dako), a,-antitrypsin (1:160000, Dako), CD34 (1: 200, Cell Marque Corporation, Austin, Tex), and Ki-67 (clone MIB-1, 1:100, Dako). For electron microscopy, small fragments of both tumors were fixed in 2.5% glutaraldehyde, embedded in epon-araldyte resin, cut into 500-nm sections, and stained with uranyl acetate-lead citrate. The grids were examined with a Zeiss EM-109 electron microscope.

The first tumor, which was excised in 1994, consisted of a well-circumscribed, but unencapsulated tan/brown nodular mass of 10 × 9 cm to 5 × 7 cm, with a narrow margin of uninvolved liver tissue. Several hemorragic foci were observed within the tumor.

Microscopic examination revealed large tumor cells with ample eosinophilic cytoplasm. These cells were aggregated predominantly in a solid fashion, with single or double cell plates and normal reticulin framework (Figure, a through c). A pseudoglandular pattern with bile plugs in centrally located canaliculi was rarely detected (Figure, b). The most striking feature of the tumor was the presence of dilated sinusoids with multiple foci of hemorrhage, imparting a striking peliosis-like aspect to the tumor (Figure, a). Nuclei were uniform and normochromatic, with slightly prominent nucleoli and and no mitotic activity (Figure, b).

The peripheral areas of the tumor contained many hepatocytes with swollen and hydropic cytoplasm; some of these cells had faintly eosinophilic cytoplasmic inclusions but clear-cut pale bodies were not observed (Figure, b). Thin-walled, artery-like vessels lacking corresponding ducts were scattered throughout the tumor. Neither stellate scar nor collagenous stroma was observed. Uninvolved liver showed no evidence of cirrhosis or chronic hepatitis.

The tumor cells were negative for [alpha]-fetoprotein, [alpha]1-antitrypsin, and fibrinogen. Staining with Ki-67 antigen (MIB-1) produced a proliferative index of less than 1% of tumor cells. Immunoreactivity for CD34 was focally positive, mainly restricted to tumor areas surrounding small arteries (Figure, d). Electron microscopic examination revealed abundant mitochondria, sometimes with paracrystalline inclusions, and hyperplastic rough endoplasmic reticulum with dilated cisterns occasionally containing flocculent material (electron micrograph not shown).

Because of the unremarkable cytologie appearance and the lack of any features supporting the diagnosis of HCC, such as mitosis, lymphovascular invasion, or decreased reticulin framework, a diagnosis of liver cell adenoma was rendered.

Macroscopically, the second tumor (excised in 1999) was bulging and nodular, with a maximal diameter of 7.4 cm. The cut surface was brown and tan, and the tumor center was retracted, with a star-shaped scar.

Histologically, this tumor consisted of solid aggregates of large polygonal cells with granular eosinophilic cytoplasm. Several paranuclear cytoplasmic inclusions (pale bodies) were observed (Figure, e). There was some cytoplasmic and canalicular bile formation. Tumor cell nuclei were large and hyperchromatic, and each had a single prominent nucleolus and moderate pleomorphism (Figure, e). Mitotic activity was not observed. Tumor cell aggregates were separated by lamellar strands of dense fibrous tissue (Figure, f and g). Immunohistochemical analysis demonstrated slight positivie reactions for [alpha]-fetoprotein and [alpha]^sub 1^,-antitrypsin in rare hepatocytes. Fibrinogen immunoreactivity was also detected in several pale bodies. Staining with Ki-67 antigen (MIB-1) produced a a proliferative index of 3% of tumor cells.

CD34 diffusely decorated sinusoids, particularly along thickened trabeculae and in periseptal areas (Figure, h). Ultrastructural examination revealed cells characterized by cytoplasm filled with abundant, tightly packed mitochondria, rough endoplasmic reticulum, and large endoplasmic vacuoles, sometimes containing electron-dense material and probably corresponding to the pale bodies observed with light microscopy (electron micrograph not shown).

These morphologic features are characteristic of FLC.

Flow Cytometry

Flow cytometry was performed using fresh tissue from both tumors according to standard protocols. Flow cytometry analysis of the first tumor revealed diploidy, with a DNA index of 1.00 and an S-phase fraction of 1.4% to 3.2%. The second tumor was considered aneuploid, the DNA index was 1.2, and the median S-phase fraction was calculated at 6.8%.

Comparative Genomic Hybridization

Comparative genomic hybridization analysis was carried out as previously described" on formalin-fixed, paraffin-embedded tissue from both tumors. The first tumor, excised when the patient was 9 years of age, showed no chromosomal aberrations. However, the second tumor had a complex pattern of 5 chromosomal aberrations: losses involving chromosomes Ip and 4p and gains involving chromosomes 6q, 13q, and Xq.

COMMENT

Here, we describe the unusual finding of an FLC in a 14-year-old girl; this tumor developed 5 years after the patient had been diagnosed and treated for an HCA. To the best of our knowledge, the occurrence of such metachronous tumors is unique. Findings of the first tumor supported the diagnosis of liver cell adenoma, ie, sharply circumscribed tumor, unremarkable cytologie appearance, absence of mitosis and lymphovascular invasion, conserved reticulin framework, and a very low proliferation index (

The second tumor displayed morphologic features characteristic of FLC, such as large polygonal cells with prominent nucleoli, eosinophilic granular cytoplasm, and prominent lamellar fibrous tissue.

After the surgical resection of the second tumor, we reevaluated the histology of the first tumor to exclude an unusual variant of FLC, which could have developed secondary genetic alterations with progression. In focal areas of this first tumor, some cells had cytoplasmic homogenous ground-glass-like inclusions resembling typical pale bodies observed in FLC (Figure, b). However these inclusions, in contrast to the cytoplasmic inclusions of the second tumor, were immunohistochemically negative for fibrinogen and [alpha]1-antitrypsin. Ground-glass-like inclusions may be seen in a variety of liver disorders other than HCC and FLC, including drug induction, hepatitis B virus infection, glycogen storage disease type IV, and endoplasmic reticulum storage diseases.6

The complete lack of fibrosis in the neoplastic parenchyma of the first tumor, despite extensive tissue sampling, rendered a diagnosis of FLC quite unlikely. The presence of lamellar strands of thin collagen fibers is a distinctive feature of FLC and, although not sufficient, is necessary for such a diagnosis.

Immunohistochemical, cytometric, and genetic analyses of both tumors further support the assumption that these 2 tumors are distinct. Focal and weak CD34 expression of sinusoids and a very low proliferation index in the first tumor favored a diagnosis of LCA. However, the diffuse pattern of sinusoidal capillarization detected immunohistochemically with CD34 staining and the higher proliferation index observed in the second tumor support a diagnosis of FLC. Moreover, no genetic abnormalities were discovered in the HCA, in contrast to the complex chromosomal imbalances detected in the FLC 5 years later. This finding is in agreement with the results reported by Wilkens et al.7

LCAs are rare benign tumors, accounting for less than 2% of all pediatric hepatic tumors.1. They may arise spontaneously,8 in association with steroid intake,1 or in inherited conditions such as Fanconi anemia, type I glycogen storage disease, Hurler disease, and familial adenomatous polyposis.2 HCAs are generally thought to develop from hepatocytes,9 but recent data suggest that at least a subgroup of HCAs may originate from hepatic progenitor cells (oval cells).10 The role of oval cells in hepatocarcinogenesis is unclear.10

Malignant transformation of a liver cell adenoma in HCC is a rare phenomenon,4 and distinguishing HCAs from well-differentiated HCCs may be challenging for the pathologist.

Only 0.5% to 1% of all HCCs occur in patients younger than 20 years, and HCCs account for a similar percentage of pediatric solid tumors.11 Among pediatric hepatic tumors, HCC is the most frequent malignant liver tumor after the age of 5 years.11 The fibrolamellar variant of HCC accounts for less than 10% of all HCCs, but this variant accounts for 40% of HCCs in patients younger than 20 years.12 In contrast to the common type of HCC in childhood associated with cirrhosis, chronic hepatitis B, biliary cirrhosis, or metabolic cirrhosis in 20% to 30% of cases, FLCs generally occur without underlying liver disease.3 Compared with the usual type of HCC, FLC seems to have a better prognosis, which could be related to younger age at occurrence or to the absence of underlying cirrhosis.3

Precursor lesions of FLC are not known. Some authors have suggested a link between focal nodular hyperplasia and FLC, suggesting that the latter may represent the malignant counterpart of the former. Further studies, however, have provided little support for such an association.3

Little is known about the molecular pathogenesis of HCC in children in general and of FLC in particular. Some authors have recently found genomic homogeneity in FLC, a feature that could explain its relatively indolent growth and distinct clinical behavior.13 Hany et al14 described the cytogenetics of a pediatric FLC with a triploid karyotype and an abnormal chromosome 1q, possibly with a translocation of chromosome 1 to another unidentified chromosome. Here, we report one of the first comparative genomic hybridization analysis of a pediatric FLC, which revealed a complex pattern of aberrations involving chromosomal losses at 1p and 4p and gains at 6q, 13q, and Xq. These results differ significantly from those reported in other comparative genomic hybridization studies of nonfibrolamellar HCC, in which the most frequent chromosomal aberrations are 1q+, 6q-8p-8q+, and 13q-,5,15 and in the few studies on FLC cytogenetics.14

Although the patient in this report shows no evidence of Carney syndrome, this inherited disease is present in her immediate family (mother). HCAs may occur in a familial fashion, eg, in families with familial adenomatous polyposis2; however, hepatocellular tumors have not been associated with Carney syndrome.

In this 14-year-old girl from a family with a history of Carney syndrome, an FLC was preceded by an HCA. Comparative genomic hybridization analysis revealed greater genetic instability in the FLC than in the HCA. The development of these 2 lesions in the same person sequentially is unique and to our knowledge has not been previously reported. It is unclear whether the sequential appearance of these 2 tumors was coincidental or whether there is an unknown inherited condition that predisposed this patient to develop both lesions.

We thank Kamal G. Ishak, MD, PhD (Department of Hepatic and Gastrointestinal Pathology, Armed Forces Institute of Pathology, Washington, DC) for reviewing the cases and confirming the diagnoses. We also thank Carole Egenter for her excellent technical assistance.

References

1. Ishak KG. Hepatic lesions caused by anabolic and contraceptive steroids. Semin Liver Dis. 1981;1:116-128.

2. Bala S, Wunsch PH, Ballhausen WC. Childhood hepatocellular adenoma in familial adenomatous polyposis: mutations in adenomatous polyposis coli gene and p53. Gastroenterology. 1997;112:919-922.

3. Craig JR. Fibrolamellar carcinoma: clinical and pathologic features. In: Okuda K, Tabor E, eds. Liver Cancer. New York: Churchill Livingstone; 1997:255-262.

4. Ferrell LD. Hepatocellular carcinoma arising in a focus of multilobular adenoma: a case report. Am J Surg Pathol. 1993;17:525-529.

5. Tornillo L, Carafa V, Richter J, et al. Marked genetic similarities between hepatitis B virus-positive and hepatitis C virus-positive hepatocellularcarcinomas. J Pathol. 2000;192:307-312.

6. Willis EJ. The powerhouse of the cell. Ultrastruct Pathol. 1992;16:iii-vi.

7. Wilkens L, Bredt M, Flemming P, Becker T, Klempnauer J, Kreipe HH. Differentiation of liver cell adenomas from well-differentiated hepatocellular carcinomas by comparative genomic hybridization. J Pathol. 2001;193:476-482.

8. Wheeler DA, Edmondson HA, Reynolds TB. Spontaneous liver cell adenoma in children. Ami Clin Pathol. 1986;85:6-12.

9. Bannasch P, Schroder CH. Tumours and tumour-like lesions of the liver and biliary tract: pathogenesis of primary liver tumours. In: MacSween RNM, Burt AD, Portmann BC, Ishak KG, Scheuer PJ, Anthony PP. eds. Pathology of the Liver. 4th ed. London, England: Churchill Livingstone; 2002:784-785.

10. Libbrecht L, De Vos R, Cassiman D, Desmet V, Aerts R, Roskams T. Hepatic progenitor cells in hepatocellular adenomas. Am J Surg Pathol. 2001;25:1388-1396.

11. Parham DM, Jenkins JJ III. Pathology of selected pediatric embryonal neoplasms. Mod Pathol. 1994;7:501-519.

12. Altmann HW. Hepatic neoformations. Pathol Res Pract. 1994;190:513-577.

13. Sirivatanauksorn Y, Sirivatanauksorn V, Lemoine NR, Williamson RC, Davidson BR. Cenomic homogeneity in fibrolamcllar carcinomas. Cut. 2001,49: 82-86.

14. Hany MA, Betts DR, Schmugge M, et al. A childhood fibrolamellar hepatocellular carcinoma with increased aromatase activity and a near triploid karyotype. Med Pediatr Oncol. 1997;28:136-138.

15. Marchio A, Meddeb M, Pineau P, et al. Recurrent chromosomal abnormalities in hepatocellular carcinoma detected by comparative genomic hybridization. Genes Chromosomes Cancer. 1997;18:59-65.

Luigi Maria Terracciano, MD; Luigi Tornillo, MD; Pierino Avoledo, MD; Dietrich von Schweinitz, MD; Thomas Kuhne, MD; Elisabeth Bruder, MD

Accepted for publication October 1, 2003.

From the Institute of Pathology, University of Basel, Basel, Switzerland (Drs Terracciano, Tornillo, and Bruder); Children's University Hospital, Basel, Switzerland (Drs Avoledo and von Schweinitz); and Children's University Hospital, Zurich, Switzerland (Dr Kuhne).

Corresponding author: Luigi Maria Terracciano, MD, Institute of Pathology, Basel University Hospital, Schonbeinstrasse 40, CH-4003 Basel, Switzerland (e-mail: lterracciano@uhbs.ch).

Copyright College of American Pathologists Feb 2004
Provided by ProQuest Information and Learning Company. All rights Reserved