Interobserver reproducibility of immunocytochemical estrogen- and progesterone receptor status assessment in breast cancer.

The interobserver reproducibility in the immunocyto-chemical assessment of estrogen- and progesterone receptor status was investigated in a series of 102 cases of primary breast carcinomas. Immunostaining was performed on 4 microns cryostat sections using estrogen- and progesterone immunocytochemical assays (ER-ICA and PR-ICA) with methylgreen counterstaining. The slides were evaluated independently by two observers. The proportion and staining intensity of ER- or PR-positive tumor cells in areas of invasive tumor growth was subjectively assessed based on the examination of the entire slide. Scoring was performed according to the proposals of Remmele and Reiner. Additionally, the cytosol estrogen- and progesterone receptor content was determined by a standard dextran-coated charcoal (DCC-) assay. Observer 1 interpreted 60 (59%) of the ER-stained specimens and 50 (49%) of the PR-stained specimens as receptor-positive; for observer 2 the respective values were 59 (58%) for ER-stained specimens and 51 (50%) for PR-stained specimens. Positive biochemical receptor status was found in 68 cases with the ER-DCC (67%) and in 47 with the PR-DCC (46%). The interobserver agreement between the two observers on the immunocytochemical receptor status was 89% for the estrogen receptor and 93% for the progesterone receptor 8.8% of the specimens were interpreted differently. Using the Remmele score, the concordance within the group of cases, which had been interpreted as receptor-positive by both observers, was 54% for the estrogen receptor and 57% for the progesterone receptor. The Reiner score showed concordant scorings of 72% for the estrogen receptor and 79% for the progesterone receptor. The present study indicates that complete agreement between scorings of different observers may not be expected, mainly due to differences in the interpretation of the specificity of staining and of the histological structures after immunostaining. The concordance of positive results using the three-graded Reiner score is comparable to that of the three-graded Bloom and Richardson grading system of breast cancer and reflects the limitation of subjective evaluation of morphology in general.