In 1980, Kempson et al coined the term "endometrial epithelial metaplasias" for the phenomenon of replacement of the surface endometrial glandular epithelium with an epithelium that is normally encountered in another mullerian-derived organ such a fallopian tube or endocervix. The authors described total seven patterns of metaplasia including already well known squamous metaplasia and adding six other types including papillary, tubal, eosinophilic, mucinous, hobnail and clear cell type. Interestingly, all the 89 cases reported by Kempson et al were benign and none of them showed any evidence of invasive adenocarcinoma on hysterectomy specimen . In 1987, Anderson et al reported the association of endometrial metaplasia with endometrial adenocarcinoma. Their 15/40 cases with endometrial carcinoma had associated endometrial metaplasia. The eosinophilic metaplasia was the most common metaplasia associated with endometrial cancer. An important observation made was that carcinoma occurring with metaplasia was always low-grade, occurred in young females with associated hyperplasia and hyperestrogenism. A similar association between endometrial metaplasias and endometrial carcinoma was reported by Kaku T et al.
The metaplastic changes have been described in both benign and malignant endometrial processes. Jacques SM et al segregated the metaplastic changes in carcinomas from the benign endometrium. Jacques SM et al called the metaplastic changes occurring on the surface of the endometrial adenocarcinoma as "surface epithelial changes". They described two distinct types of patterns: papillary syncitial type and microglandular type. Most commonly, endometrial adenocarcinoma showed admixture of these two patterns, however these patterns are also present exclusively on the tumor surface. The SEC showed less atypicality than the underlying carcinoma. However, the nuclear atypia in the SEC is more than the benign endometrial metaplasia. The papillary syncitial type of SEC resembles the papillary syncitial change seen with endometrial breakdown. The microglandular type of SEC mimics the microglandular hyperplasia of the cervix.
The SECs mimicking the serous borderline tumor of ovary have not been reported in the literature. These SECs differ from the two types of SECs described by Jacques SM et al. In our cases, cells only focally formed glandular lumen and the cells were less atypical than the underlying tumor cells. Most of the metaplastic change comprised of thin cords of the eosinophilic cuboidal cells with distinct cell membrane arranged in micropapillae with non-hierarchal branching. In a study of 399 consecutive cases of endometrioid carcinoma, Murray SK et al identified 26 cases of uterine endometrioid carcinoma with small nonvillous papillae (ECSP). In their study, they compared ECSP with uterine serous carcinoma, and therefore did not use the term "micropapillary", to avoid confusion with micropapillary type of serous carcinoma. Murray et al concluded that endometrioid carcinoma with small nonvillous papillae may be confused with serous papillary carcinoma on microscopic examination .
On one hand the molecular changes in two major subtypes of endometrial cancer are well known: the estrogen-related type I shows defects in DNA-mismatch repair, mutations in PTEN, beta-catenin, and k-ras, while type II, nonendometrioid such as papillary serous and clear cell shows aneuploidy and p53 mutations . The mechanism of origin of the SEC is unclear. Jacques SM proposed that SECs occur only under the condition of a larger space into which the epithelium can proliferate. Thus, SEC develops only on the endometrial surface and in the dilated malignant glands involving adenomyosis. There is a strong association of endometrial metaplasia with exogenous estrogen intake . Very few studies have investigated the molecular characteristics of the SECs. Quddus et al observed weak and heterogenous p53 reactivity in the metaplastic endometrium . Interestingly, p53 staining is absent in normal and hyperplastic endometrium while a strong, diffuse p53 staining is diagnostic of uterine papillary serous carcinoma. Quddus et al reported variable PTEN expression in the surface epithelial changes associated with hyperplasia and endometrioid carcinoma .
To our knowledge, SECs mimicking serous borderline tumor of the ovary have not been described in English literature. Our two cases expand the spectrum of the SECs described with the endometrioid carcinoma. One of the cases showed PTEN null phenotype and weak, focal p53 expression by immunohistochemistry, supporting the idea that these SECs originated from the underlying endometrioid carcinoma rather than a denovo phenomenon. Knowledge of this morphological variation in SECs may be important while interpreting the endometrial biopsy/curettage for an endometrial cancer.