The aim of the present study was to compare oestrogen receptor (ER) analysis results obtained in cytosols of frozen breast cancer tissue (using biochemical assay) with those obtained in paraffin-embedded tissue (using immunoperoxidase staining with monoclonal antibodies (DAKO-ER, 1D5), and an ER positivity cut-off level of >10% stained nuclei). In 86% (84/98) of the samples the same ER status (28 negative and 56 positive) was obtained with both procedures. In eight cases, the paraffin section was ER positive but the corresponding cytosol sample ER negative, whereas six cases showed the opposite pattern. The ER positive subgroup manifested better outcome after adjuvant treatment than the ER negative subgroup (p = 0.003 (cytosol), and p = 0.004 (paraffin)). As compared with the percentage of stained nuclei, staining intensity yielded no additional information. Although the results of ER analysis of paraffin-embedded material seem promising, it is too early to prefer it to frozen tissue, though this would be useful when no frozen tissue is available.