Correlation between oestrogen receptor protein expression in infiltrating ductal carcinoma of the breast by immunohistochemistry and cytosol measurements.

Fresh frozen neoplastic tissues from 70 infiltrating ductal breast carcinomas were analysed for cytosolic oestrogen receptor (ER) protein content using a solid phase enzyme immunoassay (EIA) method based on a "sandwich" principle (Abbott ER-EIA monoclonal). Formalin-fixed, paraffin-embedded sections from the same carcinomas were examined for nuclear immunoreactivity against a monoclonal antibody for ER protein (Dako) using the standard avidin-biotin complex immunoperoxidase (IP) method after microwave antigen retrieval. The degree of ER positivity by IP was also scored according to a visual estimation of the percentage of cells expressing immunopositivity and the intensity of staining. Twenty-eight (40%) of the carcinomas were ER-positive by EIA and 34 (48.6%) were positive by IP. Twenty-five (35.7%) were ER-positive and 33 (47.1%) were ER-negative by both methods. Nine (12.9%) were ER-negative by EIA but were positive by IP, this discrepancy being ascribed to sampling inadequacy for EIA. However, 3 (4.3%) tumours were ER-positive by EIA and negative by IP. This discrepancy may be variously due to inadequate antigen retrieval, faulty technique and the possibility that the two methods do not measure identical ER proteins. IP appears to have an advantage over EIA in that it has a higher pick-up rate, does not require fresh tissue and can be applied to archival material. However, to reduce false negative estimations, it may be necessary to run IP staining using more than one ER antibody. Standardisation of the IP method for ER is desirable before this method is to be widely adopted in Malaysian laboratories. Quantitation of ER positivity by IP scoring correlated poorly with actual cytosolic levels. Caution should be exercised in attaching patient management value to visual IP scoring.