AIM--To validate the use of a new mouse monoclonal antibody (1D5) directed against the N-terminal domain (A/B region) of the oestrogen receptor in an immunohistochemical assay (ER-IHA) for paraffin wax embedded tissue. METHODS--Breast cancer specimens were surgically obtained from 119 previously untreated patients. For comparison, oestrogen receptor was measured from cytosol fractions using an established oestrogen receptor enzyme immunoassay (ER-EIA) method. Oestrogen receptor "H-scores" were obtained from the ER-IHA after antigen retrieval using microwave treatment. Where discrepancies occurred between the two methods, further immunohistochemistry was performed using the H222 antibody from the Abbott Laboratories ER-ICA kit. RESULTS--The correlation between the two methods was non- linear, but despite this there was an 86% concordance between ER-EIA and ER-IHA using the 1D5 ER antibody. Fifty four per cent of tumours (64/119) were oestrogen receptor positive and 32% (38/119) were negative by both assays. A mismatch between the ER-EIA and ER-IHA occurred in 17 cases. Seven tumours were IHA positive but EIA+, but five of these were borderline negative by EIA, having values of > 5 and < 10 fmol/mg protein. Ten tumours were IHA negative and EIA+; four of these tumours were completely negative by IHA in the section studied. A further IHA assay, carried out on the 17 tumour mismatches with H222 antibody, showed that three tumours remained substantially discordant. These three tumours were strongly positive with the 1D5 antibody and negative with the H222 antibody. Two of these discordant tumours were of the rare ER negative and PgR positive phenotype and may contain oestrogen receptor that is of biological interest but which lacks the hormone binding epitope. CONCLUSIONS--The concordance between the classic enzyme immunoassay technique and the new immunohistochemical method on paraffin wax embedded sections was good. Moreover, the IHA technique using the 1D5 antibody against the N-terminal was easily reproducible. This technique may allow oestrogen receptor content to be determined in large cohorts of patients in whom archival tumour material is available.