AUTOMATED IMMUNOHISTOCHEMICAL ASSAY FOR ESTROGEN-RECEPTOR STATUS IN BREAST-CANCER USING MONOCLONAL-ANTIBODY CC4-5 ON THE VENTANA ES

Determination of breast cancer estrogen receptor (ER) status as a pred ictor of tumor response to adjuvant endocrine therapy remains a mainst ay of breast cancer management. Recent second generation anti-ER antib odies and new epitope retrieval methods have produced paraffin-based i mmunohistochemical results that correlate closely with the dextran-coa ted charcoal (DCC) assay and appear to represent a superior method of ER assay. The authors determined the ER status of 103 invasive breast cancers by paraffin-based, automated immunohistochemistry on the Venta na ES using a new monoclonal antibody, CC4-5, and compared the results to those of parallel DCC biochemical analysis and manual immunohistoc hemical analysis using anti-EW monoclonal antibody ER1D5. The specific ity of the CC4-5 antibody for ER protein was confirmed by Western blot analysis. Sixty of 103 cases were positive for ER by CC4-5 automated immunohistochemistry. With a ligand binding assay threshold value of 2 0 fmol/mg protein, there were 50 positive cases by biochemical assay. The biochemical results corresponded to an 88% rate of agreement with automated CC4-5 staining. Analysis of discordant cases revealed that t he majority of CC4-5 immunopositive only cases (8 of 11) were strongly positive, stroma rich tumors, suggesting that corresponding biochemic al measurements were diluted by non representative stromal tissue. The re was only one immunonegative, biochemically positive case (27 fmol/m g protein). Semiquantitation of CC4-5 staining using percent positive tumor cells or weighted average staining intensity (HSCORE) showed mod erate to good correlation with quantitative DCC results (r = 0.64 and 0.62, P < .0001). ER1D5 was not suitable for use on the Ventana ES, mo st likely due to temperature constraints of the instrument. By manual ER1D5 staining, 40 of 79 examined cases were positive corresponding to a 99% rate of agreement with automated CC4-5 staining. Semiquantitati on of ER1D5 staining by percent positive tumor cells and weighted aver age staining intensity (HSCORE) showed excellent correlation with semi quantitation of automated CC4-5 results (r = 0.90 and 0.88, P < .0001) . Automated immunohistochemistry using the Ventana ES and monoclonal a ntibody CC4-5 is a reliable method for determining breast cancer ER st atus. As with other immunohistochemical methods, direct correlation wi th morphology; precludes errors due to tissue sampling, allowing for a ccurate analysis of stroma-rich or partially necrotic tumors and small neoplasms that otherwise would yield insufficient tumor tissue for bi ochemical analysis.