To determine the efficiency of image analysis in immunohistochemical progesterone receptor (PgR) measurement, 94 primary breast carcinoma tissue samples were evaluated for PgR by biochemical dextran-coated charcoal assay (DCC) and an immunohistochemical method. Frozen sections immunostained for PgR with a monoclonal antibody (Abbott PgR-ICA, Chicago, IL) and the peroxidase-antiperoxidase technique were scored semiquantitatively histologic score by microscopy and quantitatively (percentage nuclear area immunopositivity [PNA] using the CAS 200 image analyzer (Cell Analysis Systems, Elmhurst, IL). There was a positive correlation between dextran-coated charcoal assay and both histologic score (r = 0.82) and PNA (r = 0.69). Selected cutoff points of 60 histologic score and 6.5% PNA based on sensitivity/specificity calculations yielded a predictive value of a negative test of 73% and 80%, respectively, and a positive predictive value of 100% for both; ranges of fmol/mg protein PgR correspond to ranges of histologic score and PNA. The use of an image analyzer to measure PNA in PgR-immunostained sections is a viable alternative to dextran-coated charcoal assay, especially when insufficient fresh tissue is available.
Read full abstract