Comparison of ligand binding assay and enzyme immunoassay of oestrogen receptor in human breast cancer cytosols

Abstract

Nine laboratories cooperating in the Receptor Study Group of the EORTC compared results of ligand binding assay supplemented with Scatchard plot analysis (LBA) and enzyme immunoassay (EIA) for the assessment of oestrogen receptors in 1665 breast cancer cytosols. An excellent agreement was observed between results of Scatchard plot analysis and EIA in each laboratory. Linear correlation coefficients of log-transformed data (log [(ER + 10)/10]) ranged from 0.839 to 0.977 (n = 52-373; P less than 0.001); Spearman's R ranged from 0.797 to 0.972 (P less than 0.001). Orthogonal regression analysis on log-transformed data revealed slopes of 0.794 to 1.141 and intercepts of -0.057 to 0.154 corresponding to -1.2 to 4.3 fmole/mg protein. Both assays compared equally well for pre- as well as postmenopausal patients, which confirms that occupied receptors are not extracted during the preparation of cytosol. The percentage discordance in the classification of tumours as ER positive or negative varied from 4.1 to 13.3 when a cut-off value of 10 fmol/mg protein was used, and from 1.4 to 7.5 at a cut-off level of 20. Considerable variations were observed in the actual receptor levels reported by each laboratory. Since these differences occurred in the results of both methods, they are attributed to differences in tissue handling. It is concluded that the ER-EIA is an excellent alternative to Scatchard plot analysis for the assay of oestrogen receptors. It is recommended, however, that laboratories performing assays on samples obtained from patients who are eligible for entrance in EORTC clinical trials and who wish to use the EIA should validate this assay against the Scatchard plot assay in their own setting.