The prognostic value of immunohistochemical estrogen receptor analysis in paraffin-embedded and frozen sections versus that of steroid-binding assays.

Estrogen receptors (ER) were independently analyzed using dextran-coated charcoal assays (ER-DCC) and immunohistochemical assays in frozen (ER-ICA) and paraffin-embedded tissue (ER-PAR) from 130 human breast cancer specimens drawn from postmenopausal high-risk patients registered in the Danish Breast Cancer Cooperative Group. ER was best detected with the ER-DCC assay followed by the ER-ICA (relative sensitivity 87%) and the ER-PAR assays (relative sensitivity 71%). The semiquantified staining features of the immunohistochemical assays were statistically significantly correlated with each other and with ER-DCC. Analysis of disease-free interval (DFI) and overall survival (OS) showed that all assays allowed statistically significant discrimination between a high risk and a low risk group, although the sensitivity differences tended to be reflected as small differences in clinical discriminatory power. The patient groups were then stratified according to adjuvant treatment [radiotherapy (RT) versus radiotherapy and tamoxifen (RT + TAM)]. The survival advantage was tied primarily to the receptor status itself in the steroid-binding assays, but was linked to both the receptor status and the adjuvant treatment in the immunohistochemical assays. Thus, the relative risks in terms of DFI and OS were of the same relative magnitude in the RT and RT + TAM groups for ER-DCC assays using a cut-off level of 10 fmol/mg cytosol protein, while there were large differences in the relative risks between RT and RT + TAM groups for ER-ICA and ER-PAR assays. We conclude that an ER assay in fresh tissue should be given first priority, but if there is no fresh tissue, an ER assay in paraffin-embedded tissue offers a reasonably good alternative as a prognosticator and an equivalent alternative as a predictor of the response to endocrine treatment.