Surface epithelial changes (SECs) have been reported to be associated with endometrial carcinoma in about 49% of cases. They have also been seen to be associated with endometrial hyperplasia. Although morphologically benign, these lesions are considered a marker for underlying malignancy. Their true biologic nature, however, is not known. PTEN has been reported to be altered in endometrial carcinomas and endometrial precancers where no morphologic changes are appreciated. The current study aims to investigate PTEN status in the endometrial SECs. Thirty-two cases of endometrial biopsy are divided into 2 groups: group 1 had 18 cases of SEC and underlying endometrial atypical hyperplasia and group 2 had 14 cases of SEC alone. Four-micron sections were cut from each block and stained with antibodies to PTEN. One section was stained with hematoxylin and eosin to confirm the presence of the area of interest. Positive and negative control slides were run with each batch of staining. All 32 cases were followed for a median period of 54 (range, 21 to 84) months. Overall, PTEN alteration (NULL) was found in 12 of the 32 cases of endometrial samples (37.5%) examined. In group 1, 10 of the 18 cases (56%) of atypical endometrial hyperplasia with "SECs" showed PTEN NULL. Of these 10 cases, 6 (60%) cases showed FIGO grade 1 endometrioid carcinoma in subsequent hysterectomy specimens. Two other cases (20%) had atypical endometrial hyperplasia only. In group 2, two of the 14 (14%) cases of SECs alone showed PTEN NULL phenotype. These 2 patients were followed for 26 and 63 months, respectively, and subsequent endometrial biopsies and pap smears were negative. Of the remaining 12 cases that retained PTEN, 9 cases (75%) were negative when followed for a median of 54 months (range, 21 to 84 mo). SECs seem to be a heterogeneous entity. Presence of PTEN NULL phenotype in SECs and associated endometrial hyperplasia does not necessarily predict an increased incidence of endometrioid carcinoma in subsequent follow-up. However, absence of PTEN NULL phenotype in SECs alone may predict a benign follow-up.