Predictive value of estrogen receptors evaluated by quantitative immunohistochemical analysis in breast cancer

To establish the reliability and biologic importance of quantitative immunohistochemical assays for estrogen receptors, more than 200 stage I and II breast cancers were studied. Immunohistochemical analysis was performed using antibody ER 1D5 (AMAC Westbrook, ME) in an automatic stainer (Tech Mate, Biotek, Santa Barbara, CA). Nuclear stain was quantitated by image analysis (SAMBA 4000; IPI, Chantilly, VA). Statistical analyses were performed using Kaplan-Meier product limit estimator for quantitated values and Cox regression for risk for death and disease progression. Estrogen receptor values obtained using the dextran-coated charcoal method were available for comparison. Quantitated immunohistochemical values, expressed as mean optical density of stained nuclei, could discriminate groups of patients that had as much as 3 times higher risk for death and 2.5 times higher risk for disease relapse (P = .006 and P = .018, respectively). Patients could be divided into three risk groups, based on their overall and disease-free survival, that were statistically significant (P = .018). The same analysis was performed using biochemically assayed estrogen receptors values showing that at a cutoff value of 20 fmol/mg, estrogen status predicted overall survival, albeit with lesser probability values (P = .032). These results show that quantitative immunohistochemistry of estrogen receptors provides results that have better predictive value than the biochemically procured ones. This approach not only obviates many pitfalls of the methods requiring tissue homogenates but also allows simultaneous correlation with histopathologic tumor features and expression of other relevant markers.