Determination of oestrogen receptors: application of the Passing-Bablock linear regression technique for comparison of enzyme immunoassay and radioligand binding assay in 1841 breast cancer tumours.

To test the qualities of two assays in the same laboratory on the same tumours, a single-point dextran-coated charcoal radioligand binding assay (RLA-DCC) and the Abbott enzyme immunoassay (EIA) were used to perform oestrogen receptor determinations on cytosols from 1841 breast cancers over a 2-year period. Statistical analysis of the data was performed by the Passing-Bablock linear regression technique. The final regression curve between EIA (y) and RLA-DCC (x) yielded y = 1.187 x fmol/mg of protein. However, a high variability in this correlation was observed from 1986 to 1988. This variability could be explained by calibration problems in the immunoassay kits and changes in our technical team. The binding assay appears to be more sensitive to the technicians' experience than the immunoassay. Other technical points are discussed, particularly cytosol preparation and KCl presence or absence in the homogenisation buffer. Finally, the Passing-Bablock and the least squares regression procedures are compared. The conditions allowing optimal correlation and routine determination reliability are defined and the correlation variability is discussed.