Summary Fine needle aspirates from 82 patients with breast carcinoma were fixed in methacarn, double embedded in agar or gelatin, and then in paraffin wax. Sequential sections were stained with monoclonal antibodies to the oestrogen receptor-related protein P29 (antibody D5), carcinoembryonic antigen (CEA), epithelial membrane antigen (EMA) and cytokeratin (CAM 5.2). Sixty-one of 82 (74%) aspirates provided sections suitable for immunostaining. Twenty-six (43%) were D5 positive, 23 (38%) CEA positive, 59 (97%) EMA positive, and 54 (89%) CAM 5.2 positive. Twenty-six of these patients were treated with some form of endocrine therapy. Twelve (46%) showed positive staining for D5. Eleven (92%) of the 12 D5-positive patients responded or had static disease, and 8% progressed. Of the 14 D5-negative tumours 43% responded or remained static, and 57% progressed. The difference in response between the D5-positive and the D5-negative tumours was significant (P<0.05, Fisher's exact test). There was no correlation between staining for CEA, EMA or cytokeratin and response to endocrine therapy.Biochemical estimation of the oestrogen receptor (ER) content of breast tumours is a well established predictor both of response to endocrine therapy and of survival (Seibert & Lippman, 1982;Witliff, 1984 Greene, 1984) and related proteins (King et al., 1985) are now available. The mouse monoclonal antibody D5 directed against the oestrogen receptor-related protein, P29, can be used to immunostain paraffin embedded material fixed in alcohol or in methacarn, without enzyme predigestion (King et al., 1985;Cano et al., 1986). The rat monoclonal antibody directed against the nuclear oestrogen receptor protein (ERICA antibody, Abbott Laboratories, Chicago, USA) requires either fresh unfixed tissue or DNAse pretreatment of paraffin embedded sections (Shintaku & Said, 1987).There have been reports of the predictive value of staining with other antibodies. It has been stated that the presence of carcinoembryonic antigen (CEA) (Shousha et al., 1979;Walker, 1980), and the intensity of staining for a material similar to the large glycoprotein, termed epithelial membrane antigen (EMA) (Heyderman et al., 1985), using monoclonal antibodies HMFG1 (Wilkinson et al., 1984) Although smears prepared from fine needle aspirates have been used for the immunocytochemical localisation of oestrogen receptor and related proteins (Flowers et al., 1986;Cavailles et al., 1987;Coombes et al., 1987), these preparations are unsuitable for the comparative demonstration of multiple markers due to the variation in number of cells per slide, and limitations in the number of slides that can be prepared from a single aspirate. An alternative approach has been to expel the aspirates into tissue culture medium, and make cytospin preparations on polylysine coated slides (Hawkins et al., 1988 The aim of this study was to evaluate a quick and easy method suitable for a busy outpatient clinic, which would allow samples to be sent through the post for further processing, with no ...