Reproducibility of subjective immunohistochemical estrogen- and progesterone receptor determination in human...

To obtain a reliable scoring system for semi-quantitation of estrogen- (ER) and progesterone receptors (PgR) in human endometrial tissue, we investigated the reproducibility of subjective immunohistochemical ER and PgR determination. Specimens of frozen endometrial tissue were stained once (n = 129) or twice (n = 22) using the ER-ICA and PgR-ICA kits from Abbott. The semi-quantitative approach we used included subjective estimates of the overall staining intensity (I) and of the fraction of stained cells (%). Scoring was performed twice by the same observer and once by another observer (n = 87). Intra- and inter-observer agreement were evaluated using Kappa statistics. We found that more comprehensive scorings of I and % could not be agreed upon by the observers. Only simplified estimates of the fraction of cells stained and overall staining intensity were reproducible. Subjective estimates obtained by this method agreed with estimates obtained by counting (n = 38). Simplified H-scores, which were obtained by multiplication of the simplified estimates of % and I, were reproducible, too. In addition, semi-quantitation of ER and PgR by immunohistology was significantly correlated to quantitation by enzyme-immunoassay (n = 39). Thus, it was possible to reproducibly semi-quantitate ER and PgR only by employing a very simple immunohistochemical scoring of ER and PgR.