Assessment of hormone receptors in breast carcinoma by immunocytochemistry and image analysis. I. Progesterone receptors.

Frozen sections of 30 breast carcinomas were stained for progesterone receptors (PRs) using a rat monoclonal primary antibody and an alkaline phosphatase-antialkaline phosphatase technique. The micro-TICAS image analysis system was used for evaluation of the staining, with the results obtained by image analysis compared with the results of biochemical assays for PR. Strong positive/negative concordance (90%) was observed between the immunohistochemical and biochemical assays. However, the numerical values of the positive cases in the two assays did not correlate well, possibly because the biochemical assay does not take tumor cellularity into account. Three PR distribution patterns, designated A, B and C, were identified by image analysis among the breast tumors. In the type A pattern, tumor cell nuclei were diffusely and uniformly labeled. In type B, both clearly negative as well as distinctly positive cells were present. In type C tumors, a broad range of labeling reactions (from negative to intensely positive) was observed. These results imply (1) that the PR content of human breast carcinoma may be accurately and objectively assessed by the image analysis of immunohistochemically stained frozen sections, (2) that image analysis may provide a more accurate estimate of the cellular content of PR than do biochemical assays and (3) that PR distribution patterns obtained through image analysis permit the consistent appraisal of intratumoral heterogeneity of PR expression, which is potentially of prognostic importance.