Quantitative determination of c-erbB-2 in human breast tumours: potential prognostic significance of low values

Abstract The purpose of this prospective multicentric study was to quantify the c-erbB-2 protein and investigate its relationship with DNA amplification and with various prognostic parameters of breast cancer. A total of 1062 primary operable human breast tumours were collected from six French anticancer centres. The c-erbB-2 protein was measured using an enzymoimmunoassay using two monoclonal antibodies directed against the extracellular domain of the protein. The results were expressed in arbitrary units/mg membrane protein (AU) after adjustment for the anticancer centre. A significant association was found between the dosage of the protein and DNA amplification ( P  = 0.0001). A value of 200 AU was found to maximise sensibility and specificity and was chosen as a cut-off for overexpression. Significant associations were found between c-erbB-2 values and oestrogen receptor (ER) ( P  = 0.01), progesterone receptor (PgR) ( P  = 0.0001) and histological grading ( P  = 0.01). The extreme high values (above the mean plus one standard deviation, S.D.) were significantly more numerous in ER − ( P  = 10 −16 ), PgR − ( P  = 10 −14 ) and grade III ( P  = 10 −8 ) tumours. The extreme low values (below the mean minus one S.D.) were significantly more numerous in ER − ( P  = 10 −9 ) and PgR − ( P  = 0.02) tumours. This prospective study confirms that high c-erbB-2 protein values are linked to poor prognostic factors and shows for the first time that low values are also linked to hormone receptor negative tumours, suggesting that these low values might also have a negative prognostic significance.