Background/aims: The prognosis of hepatocellular carcinoma (HCC) is poor because of frequent intrahepatic metastasis (IM) or multicentric carcinogenesis (MC). This study compared the effectiveness of microarray analysis in the diagnosis of these 2 forms with that of conventional histopathological diagnosis. The aim was to identify IM- or MC-associated genes through delineation of the clonality of multinodular liver cancer.
Methodology: The clonal relationship of 23 tumor foci obtained from 11 surgically resected liver specimens was investigated by genome-wide expression profiling via an in-house cDNA microarray consisting of 4,608 genes.
Results: The gene expression signature of primary HCCs with IM was very similar to that of their corresponding IMs, implying that genes favoring progression of metastasis were initiated in the primary tumors. In comparison, different gene expression was observed in multicentric HCCs. The gene for adrenomedullin, which has been identified as a lead gene in the gene expression signature, was overexpressed in HCCs with IM, as confirmed by real time-PCR and immunohistochemistry.
Conclusions: Analysis of expression profiles by microarray could provide a reliable method of delineating the clonal relationship of multiple nodules of liver cancer and identifying metastasis-associated genes. Adrenomedullin is a factor associated with progression of IM in human HCC.