DNA ploidy of fibrolamellar hepatocellular carcinoma by image analysis

DNA Ploidy of Fibrolamellar Hepatocellular Carcinoma by Image Analysis GIULIA ORSATTI, MD, PIETRA D. GREENBERG, MD, DONALD B. ROLFES, MD, KAMAL G. ISHAK, MD, PHD, AND FIORENZO PARONETTO, MD Twelve cases of fibrolameLlarhepatocellularcarcinoma (FLC) were evaluated for DNA ploidy by means of image analysisof Feulgenstained tissue sections. All of the tumors showed a nondiploid DNA distribution(six aneuploidand six tetraploid); no diploid patternwas found. The nuclear area of the tumors (53.8 pm* f 18.0; mean ? standarddeviation)was significantlylarger than that of the surrounding noncancerouslivers (33.2 + 4.7; P < .OOOl).These findingssuggest thatDNA contentin fibrolamellarcarcinoma(FIX) is not directly related to the clinicalbehaviorand that other factors may be responsible for the better prognosis of this variantof HCC. HUM PATHOL 25:936-939. This is a US governmentwork. There are no restrictions on its use.

There has been controversy over the prognostic value of the nuclear DNA content in hepatocellular carcinoma (HCC) . Some studies have indicated a correlation between nuclear DNA ploidy pattern and survival,‘” whereas others have shown no relationship between DNA ploidy and prognostic outcome.5” Fibrolamellar carcinoma (FLC) is a unique type of HCC that has been shown to have a favorable biologicai behavior and prognosis compared with other types of HCC. It usually occurs in young patients (mean age, 23.1 to 26.4 years) without cirrhosis or evidence of viral hepatitis.“,” The purpose of this study was to determine if there is a correlation between the nuclear DNA ploidy pattern and the known favorable prognosis of this rare histological variant of HCC.

MATERIALS Twelve

AND METHODS cases of FLC were retrieved from the files of the

Department of Hepatic and Gastrointestinal Pathology of the Armed Forces Institute of Pathology, Washington, DC. cal information was available for only four patients: males and one female ranging in age between 12 and 28 None of the four patients had background cirrhosis or

Clinithree years. a his-

From the Immunopathology Laboratory, Veterans Affairs Medical Center, Bronx, NY; The Lillian and Henry M. Stratton-Hans Pop per Department of Pathology, Mount Sinai Medical Center of the City University, New York, NY; the Holy Cross Hospital, Silver Spring, MD; and the Department of Hepatic and Gastrointestinal Pathology, Armed Forces Institute of Pathology, Washington, DC. Accepted for publication March 22, 1994. Supported in part by the Department of Veterans Affairs, Washington, DC. I$ words: fibrolamellar carcinoma, hepatocellular carcinoma, nuclear DNA content, DNA ploidy. Address correspondence and reprint requests to Giulia Orsatti, MD, Pathology and Laboratory Medicine Service, Veterans Affairs Medical Center, 130 W Kingsbridge Rd, Bronx, NY 10468. This is a US government work. There are no restrictions on its use. 004&8177/94/2509-0015$0.00/0

936

tory of hepatitis. In addition, 10 cases of the usual variety of HCC were retrieved to be used as controls. Formalin-fixed, paraffin-embedded tissues were cut into 5+m sections and stained with hematoxylin-eosin for light microscopic examination. Histologically all the FLCs consisted of large polygonal cells with abundant eosinophilic cytoplasm, large nuclei, and prominent nucleoli separated into cords by lamellar fibrous strands (Fig 1). The control HCCs showed a trabecular or acinar growth pattern; when graded according to Craig et al” they were grade III (eight tumors) and grade II (two tumors). Feulgen staining for DNA analysis was performed using the CAS DNA staining kit (Cell Analysis Systems, Inc., Elmburst, IL), a commercially available stain and reagent system. The 5-,um sections were placed in 5 N HCl for 1 hour and then transferred to a Coplin jar containing the stain solution for 1 hour. After staining the slides were placed in three consecutive rinse solutions: one for 30 seconds, one for 5 minutes, and one for 10 minutes. The slides were washed with deionized water, placed in acid alcohol for 5 minutes, and then dehydrated in alcohol, cleared in xylene, mounted, and coverslipped. The sections were evaluated for DNA content with a CAS 200 Image Analyzer (Cell Analysis Systems, Inc.). Each stain batch had a control slide of external control cells to calibrate the system. Two hundred tumor cells were analyzed on each slide; in addition, 100 cells from noncancerous adjacent areas were evaluated for normal diploid control. The tissue function provided for the US 200 was used. This function offers a correction for section thickness to account for the possibility that some nuclear material was cut off in the sectioning process. The measurements included nuclear size (in squared micrometers), total nuclear optical density (directly proportional to DNA content and expressed in picograms), and DNA index (DI). These were recorded digitally and displayed as histograms. The DNA content of the cell populations was expressed in ploidies in which the peak corresponding to the DNA content of the internal control benign hepatocytes was used as the diploid control and assigned a DI of 1. The diploid range was between DI 0.8 and 1.2; peaks ranging from DI 1.8 to 2.2 and containing more than 20% of the total number of cells were classified as tetraploid, whereas neoplasms showing peaks outside of the above mentioned ranges were classified as aneuploid. Statistical analysis was performed using the Student’s t-test (two-tailed) using a computerized statistical package (GEStat Dinamic Microsystems, Inc., Silver Spring, MD).

RESULTS The results of the ploidy analysis of the 12 FLCs are shown in Table 1. Six of the tumors (50%) displayed an aneuploid DNA content (Fig 2A) and the remaining six lesions (50%) showed a tetraploid DNA content (Fig

DNA ANALYSIS

OF FIBROIAMELLAR

CARCINOMA

(Orsatti et al)

contrast, other studies, performed by means of microspectrophotometry’ or flow cytometry”x have shown no correlation between DNA ploidy and patient survival. The percentage of diploid and aneuploid liver cancers varied markedly from one series to another. Fujimoto et al’ found an equal percentage of diploid and aneuploid cases, whereas Chen et al” (22% diploid and 78% aneuploid) and McEntee et alR (33% diploid and 67% nondiploid) observed a majority of aneuploid cases. Conversely, Nagasue et al’ found a slight majority of diploid cases (58.2% diploid and 41.7% aneuploid). When the tumors were separated according to size there was a greater percentage of diploid cases among small tumors (less than 3 or 5 cm) and a greater percentage of aneuploid cases among large tumors in two series? Another study showed an almost equal percentage of diploid and aneuploid cases among small tumors (less than 5 cm) but a vast majority of aneuploid cases among large tumors.’ Several parameters, such as histological grade, patient’s age, presence of cirrhosis, and hepatitis B surface antigen (HBsAg) status, have been studied in connection with the ploidy of HCC with variable results.l-X.‘” However, no previous study correlated the ploidy with the specific histological growth patterns of HCC. In the present investigation we have analyzed the nuclear DNA content of the fibrolamellar variant of HCC. This tumor constitutes a distinct entity from the ordinary HCC not only by virtue of its peculiar histological features but also because of its specific clinical presentation and biological behavior. The usual HCC develops in patients 40 to 60 years of age, affects males more commonly than females (3 or 4:1), and is frequently associated with liver cirrhosis. 13.i4The tumor is typically widely disseminated in the liver or multicentric at the time of diagnosis and, therefore, is rarely resectable. The prognosis is generally poor with an overall median length of survival of approximately 4 months.‘“.i4 In contrast, FLC occurs predominantly in children and young adults”.“; almost half of the liver cell carcinomas in patients younger than 35 years of age belong to this type.‘” Fibrolamellar carcinoma shows no sex predilection and no association with

FIGURE 1. Microscopic view of FLC showing large polygonal cells with abundant cytoplasm separated into cords by lamellar fibrous strands. (Hematoxylin-eosin; original magnification x 100.)

2B); no diploid pattern was identified. The coefficient of variation of the diploid control peaks was 16.6 ? 3.2 (mean -t standard deviation [SD]), whereas that of the aneuploid peaks was 22.8 f 7.9. The nuclear area of the tumor cells (in squared micrometers) was 53.8 5 18.0 (mean + SD), whereas that of the adjacent noncancerous liver cells was 33.2 & 4.7 (Table 1). The difference between the two groups was statistically significant (P < .OOOl). DNA analysis of the 10 nonfibrolamellar HCCs showed three diploid, two tetraploid, and five aneuploid tumors (Table 2). The nuclear area of the tumor cells was 54.6 pm’ + 14.0 (mean + SD), and that of the adjacent non-neoplastic liver cells was 39.7 + 5.5. The difference in occurrence of nondiploid neoplasms between fibrolamellar and control HCCs was not statistically significant.

TABLE 1. Ploidy Analysis of 12 Cases of Fibrolamellar Carcinoma Fibrolamellar Noncancerous Case NO. 1 2 3 4 5 6 7 8 9 10 11 12

DlSCUsslON The significance of the nuclear DNA content in evaluating the outcome of patients with HCC is to date not entirely clear. The results of several large flow cytometric’.‘, and image cytometric” investigations have indicated that the nuclear DNA ploidy pattern may serve as a valuable marker in predicting the malignant potential and prognosis in patients with HCC. In 937

Nuclear

Area ($) 27.2 31.4 37 36.8 37.6 29.1 33.9 36.9 23 35.2 30.8 39.8

Carcinoma

Liver Nuclear Area (p’) 51.2 57.7 57 64.5 60.4 57.2 52.3 77.1 51.3 45.2 49.7 75.9

DI

Interpretation

2.17 2.65 1.96 1.53 1.57

Tetraploid Aneuploid Tetraploid Aneuploid Aneuploid Tetraploid Aneuploid Tetraploid Aneuploid Aneuploid Tctraploid Trtraploid

2.12 1.55 2.05 2.66 1.51 1.85 2.13

HUMAN

A

b

16 DNA

PATHOLOGY

32

Mass

h3

Volume 25. No. 9 (September

614

B

Ploidy Analysis of 10 Cases of Hepatocellular Carcinoma

2.

Case NO. 1 2 3 4 5 6 7 8 9 10

Nuclear

Area ($) 38.7 26.6 41.7 40.1 41.4 45.6 39 35.3 41.5 48

Hepatocellular Carcinoma ~ Nuclear Area (pLp) 67.5 40.9 38.8 78.5 38.1 44.7 53.1 73.7 49.9 61.3

1;

3; DNA

cirrhosis, viral hepatitis, or oral contraceptives.“.“’ Most importantly, this tumor type has a relatively good prognosis with a mean length of survival of 68 months in one series.” Grossly the tumors are well-circumscribed, usually solitary masses with prominent bands of fibrosis. Microscopically the epithelial component consists of deeply eosinophilic cells with abundant granular cytoplasm and prominent nucleoli. Eosinophilic bodies and pale bodies are frequently identified in the tumor cells.!‘~‘O However, the most characteristic feature is the diffuse fibrous stroma, which is composed of fibrillar bands of collagen arranged in a lamellar and fascicular pattern and in delicate bands between nests of tumor cells.“~‘” Ultrastructurally the most striking feature is an abundance of mitochondria that completely fill the cytoplasm of the tumor cells.‘” Our study of the nuclear DNA content in 12 cases of FLC by image cytometry showed aneuploid (50%) or tetraploid (50%) DNA distribution. No diploid pattern was identified among these neoplasms. In addition, the mean nuclear area of the tumor cells was significantly larger than that of the surrounding noncancerous liver cells. This finding supports the increased complement of DNA in these tumors and confirms previous observations of a proportional increase in nuclear size and DNA content in non-neoplastic aging human livers” and in preneoplastic liver lesions.‘”

Noncancerous Liver

b

(Picograms)

FIGURE 2. (A) Distribution of DNA mass. DNA histogram showing an aneuploid showing a tetraploid peak, DI = 1.85 (case no. 11).

TABLE

1994)

DI

Interpretation

2.26 1.88 0.94 2.34 1.11 1.04 I .67 2.72 1.3 1.82

Aneuploid Tetraploid Dipfoid Aneuploid Diploid Diploid Aneuploid Aneuploid Aneuploid Tetraploid

peak, DI =

Mass

d

6:

(Picograms)

1.51(case no. 10). (B> DNA histogram

Analysis of the control HCCs disclosed three diploid tumors and seven nondiploid tumors; the difference in occurrence of nondiploid neoplasms between fibrolamellar and control HCCs was not statistically significant. The high frequency of nondiploid tumors in our series of FLCs strongly contrasts with the reported favorable prognosis of this tumor type and suggests that the DNA content in FLC may not be directly related with the clinical outcome. Regrettably, we do not have sufficient follow-up information in our cases to support this speculation, and additional studies will be necessary to clarify the relationship between ploidy and prognosis in FLC. Whether tetraploidy in FLC, as in other organ systems, I!)-“:! indicates a subgroup with an even more benign course is also unclear. In one series of HCC in which various types of ploidy were studied,’ tetraploid/ polyploid tumors constituted 30% of the cases. However, insufficient data are available with regard to the biological significance of tetraploidy in HCC and further investigations should be performed to analyze the clinical outcome of tetraploid HCC and FLC. Whereas the role of DNA ploidy in FLC is not completely elucidated, we can speculate that factors other than aneuploidy may be responsible for the generally favorable prognosis of FLC. Perhaps the florid desmoplastic response associated with this variant of HCC may account for restricting further tumor growth and spread, thus contributing to the overall better prognosis. Indeed, the growth pattern of these tumors, most often consisting of well-circumscribed rather than diffusely infiltrating masses, is responsible for their high resectability rate!‘.“’ and, therefore, for the longer length of survival of patients with FLC compared with those with the usual form of HCC.

REFERENCES 1. Ishizu H: Flow cytometric analysis of the nuclear DNA content of hepatocellular carcinoma. Jpn J Surg 6:662-673, 1989 2. Fujimoto J, Okamoto E. Yamanaka N, et al: Flow cytometric DNA analysis of hepatocellular carcinoma. Cancer 67:939-944, 1991 3. Wenning C, Mengchao W: The biopathologic characteristics of DNA content of hepatocellular carcinomas. Cancer 66:49&501, 1990 4. Chiu JH, Kao HI,, Wu LH, et al: Prediction of relapse or survival after resection in human hepatomas by DNA flow cytometry. J Clin Invest 89:539-545, 1992 938

DNA ANALYSIS

OF FIBROLAMELLAR

5. Ezaki T, Kanematsu T, Okamura T, et al: DNA analysis of hepatocellular carcinoma and clinicopathologic implications, Cancer 61:106109, 1988 6. Chen MF, Hwang TI,, Tsao KC, et al: Flow cytometric DNA analysis of hepatocellular carcinoma: Preliminary report. Surgery 109:45.5458, 1991 7. Nagasue N, Yamanoi A, Takemoto Y, et al: Comparison hetween diploid and aneuploid hepatocellular carcinomas: A flow cytometric study. Br J Surg 79:667-670, 1992 8. McEntee GP, Batts KA, Katzmann JA: Relationship of nuclear DNA content to clinical and pathologic findings in patients with primary hepatic malignancy. Surgery 111:376379, 1992 9. Berman MM, Libbey NP, Foster JH: Hepatocellular carcinoma, polygonal cell type with fibrous stroma. An atypical variant with a favorable prognosis. Cancer 46:1448-1455,198O 10. Craig JR, Peters RL, Edmondson HA, et al: Fibrolamellar carcinoma of the liver: A tumor of adolescents and young adults with distinctive clinicopathologic features, Cancer 46:372-379, 1980 11. Craig JR, Peters RL, Edmondson HA: Tumors of the Liver and Intrahepatic Bile Ducts. AFIP Fascicle, 2nd series, Washington, DC, 1989, pp 153-156 12. Kuo SH, Sheu JC, Chen DS, et al: Cytophotometric measurements of nuclear DNA content in hepatocellular carcinomas. Hepatology 7:330-332. 1987 13. Chlebovski RT, Tong M, Weissman J, et al: Hepatocellular carcinoma. Diagnostic and prognostic features in North American patients. Cancer 53:2701-2706, 1984

CARCtNOMA

(Orsatti et al)

14. Nagasue N, Yukaya H, Hamada T, et al: The natural history of hepatocellular carcinoma. A study of 100 untreated cases. Cancer 54:1461-1465, 1984 15. Farhi DC, Shikes RH, Murari PJ, et al: Hepatocellular carcinoma in young people. Cancer 52:15161525, 1983 16. Farhi DC, Shikes RH, Silverberg SC: Ultrastructure of fibrolamellar oncocytic hepatoma. Cancer 50:702-709, 1982 17. Watanabe T, Tanaka W: Age-related alterations in the size of human hepatocytes: A study of mononuclear and binucleate cells. Virchows Arch B Cell Pathol 39:9-20, 1982 18. Orsatti G, Theise ND, Thung SN, et al: DNA image cytometric analysis of macroregenerative nodules (adenomatous hyperplasia) of the liver: Evidence in support of their preneoplastic nature. Hepatology 17:621-627, 1993 19. Auer GU, Caspersson TO, Wallagen AS: DNA content and survival in mammary carcinoma. Anal Quant Cytol 3:161-165, 1980 20. Coulson PB, ThornthwhiteJT, Woolley TW, et al: Prognostic indicators including DNA histogram type, receptor content, and staging related to human breast cancer patient sun&al. Cancer Res 44:41874196, 1984 21. Quirke P, Dixon MF, Claiden AD, et al: Prognostic significance of DNA aneuploidy and cell proliferation in rectal adenocarcinema. J Pathol 151:285-291, 1987 22. Giaretti W, Danova M, Geido E, et al: Flow cytometric DNA index in the prognosis of colorectal cancer. Cancer 67:1921-1927, 1991

939