Summary We previously reported the production of a panel of murine monoclonal antibodies which recognize glycoproteins abnormally expressed in human breast tumours. Using two of these antibodies, a double antibody radioimmunoassay was designed to quantify levels of these breast tumour marker glycoproteins in serum. Marker levels greater than 28 units were considered abnormal. Using this criterion, 63% and 75% of patients with breast cancer stages I and II, respectively, and 88% of those with metastatic disease were found to have elevated marker levels. Thirteen percent of patients with non-malignant breast disease also had elevated marker levels. Elevated marker levels were also detected in patients with non breast neoplasms. One hundred and eleven women with metastatic disease were followed. Eighty-two percent of those with progressive disease and 73% of those where disease regressed had 20% changes in marker levels. These changes in marker levels preceded by up to 6 months changes in disease state. From these results we conclude that this assay may be useful for monitoring the course of disease in breast cancer patients. al., 1984;Haynes et al., 1985). Hilkens et al. (1986) followed a series of advanced breast cancer patients over a period of one to six months to correlate circulating antigen levels with the clinical course of the disease; the study examined a small sample size.We previously reported the production and characterization of a panel of monoclonal antibodies reactive with components of the EMA complex expressed at high levels in human breast tumours (Major et al., 1987a;Dion et al., 1987). These antibodies react with two high molecular weight glycoproteins expressed in the majority of both primary and metastatic lesions. The panel of antibodies defines six epitopes present on molecules with apparent molecular weights of 300,000 and 280,000 daltons. Immunoblots of sera reveal that identical molecules are present in the circulation of patients with disseminated breast cancer (Major et al., 1987a 2. Then, volumes of 0, 10, 20,..., 90,ul of the above antigen preparation were added to LS normal serum to a final total volume of I00,ul. 3. Finally, 1:4 dilutions of the latter preparation were made using dilution buffer (wash buffer as defined below containing 0.5% bovine albumin). These dilutions were defined as containing 0, 10, 20,..., 90 units (U) of antigen per 50 4u1. Antigen diluted in this manner gave a linear standard curve (Figure 1). Serum from the normal female LS was used in the dilution of antigen standards because her serum gave the lowest signal (LS) in our control panel of normal sera.Patient sera to be assayed were initially diluted 1:4 in buffer; all further dilutions were made in a solution of 3 parts dilution buffer and 1 part LS serum. All incubations were done at 37°C in a humidified chamber. Polyvinylchloride microtitration plates with round bottom wells (Fisher Scientific, Montreal, Canada) were coated with 4 pg per well of antibody MA9 in 50 p1 PBS and allowed to adsorb for 16 ...