Evaluation of estrogen (ER) and progesterone (PR) receptor content is now an important procedure in the management of breast cancer patients. Production of monoclonal antibodies to ER and PR has permitted development of an enzyme immunoassay (EIA) and immunocytochemical assay (ICA). This study compared the results of ICA and EIA to evaluate ER and PR in 197 breast cancers using the same monoclonal antibodies. The ICA results were obtained by automated computer-assisted image analysis using CAS 200. The cut-off values adopted were 15 fmol/mg protein for EIA and 10% of the positive neoplastic area of the nuclei for ICA. For statistical analysis, Spearman's correlation coefficient and chi 2 were used. There was good correlation between ICA and EIA for both ER (r = 0.714; p < 0.0001) and PR (r = 0.815; P < 0.0001). Of 197 tumors, 136 (69.04%) were ER-ICA+, and 138 (70.05%) were ER-EIA+; 111 (56.35%) were PR-ICA+, and 115 (58.38%) were PR-EIA+. Results were concordant, positive or negative with both methods, in 175 cases for ER and in 173 cases for PR. ER and PR results were only discordant in 22 and 24 cases, respectively. Concordance of results obtained by the two methods was 88.83% (P < 0.0001) for ER and 87.81% (P < 0.0001) for PR. Correlation of results obtained by EIA and ICA to determine ER and PR was good. The data obtained suggest that ICA with automated image analysis is an effective means for evaluating ER and PR content in human breast cancer, especially when, as happens ever more frequently nowadays, the tumor is too small to perform EIA or when retrospective studies are performed.