Endometrial glandular dysplasia: a newly defined precursor lesion of uterine papillary serous carcinoma. Part I: morphologic features.

Abstract

Dysplastic epithelium frequently bridges the changes between normal epithelium and noninvasive carcinoma. However, such a dysplastic lesion has not been previously described in the development of uterine papillary serous carcinoma (UPSC) or between resting endometrium and serous endometrial intraepithelial carcinoma (EIC), which is composed of indisputably malignant noninvasive cancer cells. In this study, we hypothesize that there is a lesion bridging benign endometrium and serous EIC. Based on current understanding of carcinogenesis in general, the lesion should exhibit dysplastic features that are more atypical than "resting endometrium" but fall short of serous EIC. If the putative dysplastic endometrial lesion exists, it should be highly associated with UPSC rather than uterine endometrioid carcinoma (UEC). We examined the morphologic appearance of the endometrium from 32 uteri with UPSC, 16 with serous EIC, and 60 with UEC. The endometrial dysplastic lesions were identified and their pathologic features were characterized. Immunohistochemical staining with p53 and MIB-1 were performed in all sections containing endometrial dysplastic lesions, serous EICs, and benign areas. In addition, 25 postmenopausal endometrial biopsies including 6 benign resting endometria, 8 dysplastic lesions, and 11 serous EICs were also compared for the level of p53 overexpression and cellular proliferative activity. We found that endometrial dysplastic lesions do exist in the endometrial specimens we speculated and examined. We designate it as endometrial glandular dysplasia (EmGD). EmGD was present in 17 (53%) uteri with UPSC compared with 1 (1.7%) uterus removed for UEC (p = 0.001). EmGD was identified in 12 (75%) of 16 serous EIC uteri. Areas of both EmGD and serous EIC were found in 15 (47%) of the 32 UPSC uteri. Transitions from either EmGD to serous EIC or serous EIC to UPSC were present in 8 (25%) of the UPSC cases. No transitions from EmGD to UPSC were identified in any hysterectomy specimen. EmGD was frequently found in endometrial polyps. There was no statistically significant difference between EmGD in a polyp (48%) and EmGD in nonpolypoid endometrium (52%). The majority of EmGDs were multifocal and involved superficial endometrial glands. However, single glandular involvement and endometrial surface epithelial involvement were also seen. Immunohistochemically, EmGD lesions mostly showed intermediate scores/indices of p53 and MIB-1 in comparison with serous EIC and resting endometrium. EmGD is a morphologically distinct entity, which is commonly and specifically associated with uterine tumors with serous differentiation. EmGD may represent the earliest identifiable morphologic change in the development of UPSC. Characteristics of p53 and MIB-1 immunostains of EmGD may be of diagnostic usage in surgical pathology practice. Recognition of EmGD may provide an opportunity to improve the management of UPSC.