To estimate the clinical value of estrogen receptor (ER) beta expression in breast cancer we used an immunohistochemical method to detect the wild-type ERbeta in 88 primary breast cancers. We used a highly specific polyclonal antibody to the carboxyl terminus of wild-type ERbeta. This antibody reacted with neither other variant forms of ERbeta nor any part of ERalpha. Slides were evaluated on a scale representing the estimated proportion and intensity of positive-staining tumor cells. Positive staining could be seen in 52 (59.1%) of 88 breast cancers; 36 (40.9%) were negative. Although there was no correlation between ERbeta staining and age, node status, tumor size, histological grade, or progesterone receptor (PgR)-enzyme immunoassay (EIA) status, we did observe a significant correlation with ERalpha-EIA (Fisher's exact probability test: P=0.0169). Moreover, ERbeta positive cases showed a better prognosis than negative cases in disease-free survival rate (Logrank test: P=0.0662, Breslow-Gehan-Wilcoxson test: P=0.0318). Our data demonstrated the possibility that wild-type ERbeta protein expression could be used as a good prognostic indicator for breast cancer.