Purpose: The poor survival rate of hepatocellular carcinoma (HCC) is in part due to the inability to diagnose patients at an early stage. Therefore, the aim of this study was to search for candidate serum marker for HCC and to test their ability to distinguish a HCC from benign liver disease.
Experimental design: Genome-wide analysis by a microarray in 40 HCC patients was done between HCC and paired nontumor liver tissues. Expression of cystatin B (CSTB) was examined by mRNA expression analysis and immunohistochemistry. The serum CSTB levels were measured using a sandwich ELISA method in four groups, including normal healthy subjects (group 1, n = 52) and patients with noncirrhotic chronic hepatitis (group 2, n = 53), cirrhosis (group 3, n = 43), and HCC (group 4, n = 62).
Results: Microarray and statistical analyses identified 248 genes that were expressed differently between HCC and nontumor liver tissues. One of them, CSTB, was expressed preferentially in the HCCs compared with the nontumor tissues, 36 of 45 specimens (80%) by Northern blot and semiquantitative reverse transcription-PCR analyses. The serum CSTB level was much higher in HCC patients than in those with nonmalignant chronic liver disease (groups 2 and 3; P < 0.0001). The receiver operating characteristic curve indicated 5.34 ng/mL to be the optimal value for CSTB, and the sensitivity and specificity for this CSTB value were 85.5% (95% confidence interval, 74.2-93.1%) and 53.1% (95% confidence interval, 42.7-63.4%), respectively, in distinguishing between patients with HCC and those with nonmalignant chronic liver disease.
Conclusion: CSTB is specifically overexpressed in most HCCs and is also elevated in the serum of a large proportion of HCC patients. CSTB or the combination of CSTB and alpha-fetoprotein may be a useful marker for diagnosing patients with HCC with a high sensitivity.