Mucinous carcinoma (MC) of the breast is characterized by abundant extracellular mucin and by variable epithelial cellularity. Since some MCs are extremely hypocellular, we questioned the validity of biochemical (BIO) assays in these tumors. We analyzed paraffin-embedded tissue from 34 cases of MC of the breast for quantitative estrogen receptor (ER) and progesterone receptor (PR) using immunohistochemistry (IHC) on the Cell Analysis System CAS-200. Of the 34 cases, 31 (91%) were positive for ER, whereas 18 (53%) were positive for PR. In 21 cases the quantitative ER and PR were assayed biochemically by a dextran-coated charcoal method. Using the BIO results as the "true" values, the sensitivity of IHC for ER and PR was 100% and 78%, and the specificity was 13% and 64%, respectively. The low specificity of the values obtained by IHC was attributed to the fact that eight cases were "falsely" false positive (negative by BIO and positive by IHC) for ER and/or PR. Review of the histologic patterns of all 21 cases showed that 7 of the 8 falsely false positive cases were significantly hypocellular (epithelial cellularity 5-20%) as compared to the remaining cases (epithelial cellularity 20-75%). We conclude that immunohistochemical analysis of ER/PR status using image analysis in MC is a sensitive method, with the ability to detect receptor content when their concentration might be too low to be demonstrated by the conventional method as a result of sparse cellularity.