Three methods are currently available for defining the estrogen receptor (ER) status in breast carcinomas. These include biochemical ligand binding assay (LBA), immunohistochemical (IC) and mRNA in situ hybridization (ISH) semiquantitative analysis. There is still a debate as to which method should be considered the "gold standard" to define ER status. To address this topic we evaluated the above three methods in a series of 43 breast neoplasms. The results of IC (on fixed sections with ER1D5 immunostaining) and LBA assay showed moderate correlation (p = 0.004), as there were 8 (18.6%) discrepant results. ISH was extremely sensitive, with 92% of positive results. ISH results did not correlate with either IC or LBA. In view of the relative lack of reciprocal correlation among the three methods, it is difficult to define which is the most accurate system to be used to localize ER in breast carcinomas. The easiest and least expensive method probably should be used. The immunohistochemical approach has the advantage of being easily applicable to routinely fixed material, to cytological specimen and is very useful in small lesions identified with mammography only.