A detailed histopathologic analysis of 399 primary breast carcinomas was performed, and several morphologic features were correlated with the estrogen receptor (ER) status. In all cases ER status was determined immunocytochemically by estrogen receptor immunocytochemical assay (ER-ICA). In 359 carcinomas ER status was also biochemically determined. Invasive lobular, mucoid, and tubular carcinomas rather than ductal carcinomas were ER-positive more frequently in ER-ICA. Medullary and papillary carcinomas had corresponding lower or higher ER positivity, respectively, by both methods. The correlation of histologic grade and its single factors with ER status was statistically significant by both methods. Lymphocytic reaction to tumor showed a significant inverse relationship to ER status by both methods. A statistically significant higher number of ER-positive carcinomas in ER-ICA and dextran-coated charcoal assay (DCC) occurred when elastic tissue was present. Different associations were found between stromal content, tumor diameter, and ER status in DCC and ER-ICA, respectively.
Cuncer 61:1149-1154,1988.
N RECENT YEARS several investigators demonstrated I correlations between the presence of the estrogen receptor (ER) and various histopathologic factors of primary breast carcinomas. These reports, however, yielded contradictory results. In several studies correlations between ER status and some histologic tumor types, like medullary, mucoid, or invasive lobular carcinomas, have been dem~nstrated.'-~ In some studies a relationship between histologic tumor grade and ER status was r e p ~r t e d . ~. ~. ~' ' In other investigations into these factors no such relationships could be demonstrated. 1 * 8 ~1 2 ~1 3
The discrepancies may, in part, be the result of different methods of biochemical ER assay. Other reasons for the discrepancies may also be differences in evaluation and terminology, and variations of histopathologic features within a tumor.I3
It is now possible to determine ER status by direct immunocytochemical localization on tissue sections with monoclonal antibodies. 14-17 Therefore, this study was performed to investigate whether correlations be-