Using immunohistochemistry, we examined pS2 expression in 64 samples of endometrial carcinoma, 11 samples of endometrial hyperplasia and 15 samples of normal endometrium, and compared them with clinicopathological data, estrogen receptor (ER) expression and progesterone receptor (PR) expression. Of the 64 samples of endometrial carcinoma, 45 (70%) expressed the pS2 protein. The average age of the patients with pS2-positive carcinomas (54.8 +/- 8.6 years) was significantly lower than that of the patients with pS2-negative carcinomas, and all premenopausal patients were positive for the pS2 protein. Among histological types, pS2 expression was observed in 33 (92%) of the 36 G1 carcinomas, but in none of the 5 nonendometrioid carcinomas. Of the 48 ER-positive carcinomas, 43 (90%) were pS2-positive and 5 were pS2-negative. Of the 40 PR-positive carcinomas, 37 (93%) were positive for pS2. There were significant associations between pS2 expression and ER/PR expression (p < 0.001). Staining of the pS2 protein was also observed in the samples of normal endometrium. We found a progressive increase in immunoreactivity of pS2 protein from normal endometrium to endometrial hyperplasia and still more in well-differentiated carcinoma. All 11 cases of endometrial hyperplasia were strongly positive for pS2. Furthermore, patients with pS2-positive carcinomas had a better survival rate than those with pS2-negative carcinomas (p < 0.05). Our data suggest that pS2 expression is likely correlated with estrogen-related endometrial carcinoma and is possibly involved in early disease progression.