Immunohistochemical estrogen receptor assay: quantitation by image analysis.

Abstract

The immunohistochemical estrogen receptor (ER) assay quantitated by image analysis was compared with the biochemical dextran-coated charcoal (DCC) assay in 97 primary breast carcinomas. Frozen sections immunostained for ER with a monoclonal antibody (ER-ICA; Abbott Laboratories, Chicago, IL) and the peroxidase-antiperoxidase technique were scored semiquantitatively by microscopy (HSCORE) and quantitatively [percentage of nuclear area immunopositivity (PNA)] using the CAS 200 image analyzer (Elmhurst, IL). There was a significant and identical correlation between DCC and both HSCORE and PNA. According to sensitivity/specificity calculations, the cut-off points of 10 for HSCORE and 1% for PNA were chosen with positive and negative predictive values of 97.7% and 74.5%, respectively. Low DCC-positive cases between 10 and 50 fmol of ER/mg of protein were immunohistochemically ER positive only in 61.5%. Immunohistochemical results of greater than 150 HSCORE and greater than 15% PNA corresponded to high DCC results of greater than 150 fmol/mg of protein. Immunohistochemical assay quantitated by image analyzer (PNA) is a comparable alternative to biochemical ER, especially when insufficient fresh tissue is available.