The Diagnostic Utility of Hepatocyte Antigen, GPC3, and IMP3 in Distinguishing between Hepatocellular Carcinoma and Benign Hepatic Lesions.
Farshid Siadat, Bich N Nguyen, Marcio Gomes, Esmeralda Celia Marginean. University of Ottawa, ON, Canada
Background: Histopathologic distinction between hepatocellular carcinoma (HCC) and benign hepatocellular adenoma (HA) and focal nodular hyperplasia (FNH) can sometimes be challenging, especially in small biopsy samples. Hepatocyte Specific Antigen (HSA) staining has been demonstrated consistently in the vast majority of HCC. More recently, insulin growth factor messenger RNA binding protein 3 (IMP3) expression has been identified in multiple malignant neoplasms including HCC. Glypican-3 (GPC3), a cell surface proteoglycan have been shown to be overexpressed in HCC, but not in HA and FNH. The aim of this study is to determine the usefulness of these markers in the differential diagnosis of hepatocellular mass lesions.
Design: 29 surgical resected or biopsied specimens of well- to moderately- differentiated HCC (25 resections, 5 biopsies), eight HA (all resections) and twenty one FNH (15 resections, 6 biopsies) were obtained from University of Ottawa Medical Center. Immunohistochemistry was performed using HSA (Abcam), IMP3 (Abcam) and GPC3 (Santa Cruz). Cytoplasmic staining was considered positive for HSA and IMP3 and cytoplasmic and/or membranous staining was considered positive for GPC3. The percentage of positively stained tumor cells was recorded and the staining intensity was graded as weak (1+), moderate (2+), or strong (3+).
Results: Strong 2+ to 3+ HSA reactivity was detected in 20 (95.2%) of FNH, all HA, and 26 (89.7%) of HCC. Staining for IMP3 was observed in 20 (95.2%) of FNH and all HA and all HCC. IMP3 showed a stronger (3+) staining at the periphery of the tumors. GPC3 was negative in all FNH and HA whereas 20 (69%) HCC showed strong (3+) reactivity for GPC3. Neither HSA nor IMP3 could differentiate between the 3 lesions. However, GPC3 could distinguish HCC from either HA (p=0.001) or FNH (p<0.001).
Conclusions: IMP3 (Dako) was previously reported as being positive in HCC and negative in HA and non-neoplastic hepatic tissues. However in our study all HA and the vast majority of FNH expressed IMP3.
Our findings suggest that 1) Immunohistochemical detection of GPC3 but not IMP3 is a useful diagnostic tool in segregating well differentiated HCC from HA or FNH, in particular when limited material from a needle biopsy is evaluated. 2) GPC3 expression in HCC can be focal, and thus, the lack of GPC3 staining does not exclude the diagnosis of HCC. 3) IMP3 may be more frequently expressed in HCC than HSA. 4) Morphology is still the gold standard in separating FNH from HA. Larger studies are needed for further validation of these findings.
Tuesday, March 1, 2011 1:00 PM
Poster Session IV # 73, Tuesday Afternoon