PURPOSE: Oncogene c-erbB2 produces a transmembrane protein similar in structure to the tyrosine kinase family. Overexpression of c-erbB2 is known to lower the survival rate of breast cancer patients. c-erbB2 protein is an important antigen for tumor specific cytotoxic T lymphocytes induction that is dependent on its presentation as stably complexed with HLA-A2. In 1997, Nistico P reported low frequency of c-erbB2 proto-oncogene overexpression in HLA A2 positive breast cancer patients. And then in this study, correlation of HLA-A2 and the c-erbB2 expression was investigated in breast cancer patients. MATERIALS AND METHODS: HLA-A DNA typing by locus-specific generic PCR and by hybridization with sequence-specific oligonucleotide probes (SSOP) was performed on peripheral blood lymphocytes from 52 breast cancer patients (a PCR-SSOP typing method, involving a PCR amplification in conjunction with digoxigenin labelled sequence-specific oligonucleotide probes). To determine c-erbB2 expression, immunohistochemistry from paraffin-embedded tissues in a series of 47 patients with available tissue blocks was performed by use of rabbit anti-human c-erbB2 oncoprotein (DAKO, Glostrup, Denmark). And then we statistically analyzed the relation between the expressions of HLA-A2 and c-erbB2 in breast cancer patients. RESULTS: 29 out of 52 patients (55.8%) were HLA-A2 positive. 23.4% (11out of 47 patients) of breast cancer patients overexpressed c-erbB2. The patients with c-erbB2 overexpression showed lower estrogen receptor positivity compared to those without c-erbB2 overexpression (10.5%, vs 33.3%). HLA-A2 positive patients showed 18.5% (5/27) of overexpression and HLA-A2 negative patients showed 30.0% (6/20) of c-erbB2 overexpression (p=0.283). CONCLUSIONS: We observed no correlation between HLA-A2 and prognostic factors in breast cancer such as tumor size, axillary nodal status. However, our results showed a tendency without statistical significance between HLA-A2 and high frequency of c-erbB2 overexpression. More accumulation of patients will be needed for better conclusions.