Discussion

Our results indicated a strong age dependent association of endometrial polyps and endometrial carcinoma. A linear relationship in the association rate of EMPs with malignancies and increasing age was observed, with the highest association rate identified in the >65 years age group, where 32% of the EMPs were associated with malignancy. Histological evaluation and characterization of the morphological types of carcinoma demonstrated that the vast majority (87%) of endometrial carcinomas associated with EMPs were of endometrioid, followed by the serous type (9%). These relative proportions of both major histological subtypes are in accordance with the well-known distribution of similar subtypes of endometrial carcinoma in non-polypoid endometrium [12] and thus demonstrates that neither histologic subtype is more likely than the other to develop in a polyp as compared to the adjacent endometrium. The majority of the serous carcinomas developed in the oldest age group (>65 years), whereas the majority of the endometrioid carcinomas occurred in the 46–55 age group, followed by the 36–45 age group. These age distributions are in accordance with the general concept of Type I and Type II endometrial carcinogenesis [13] and provide some evidence suggesting that carcinomas developing in EMPs do not necessarily have clinicopathologic differences from carcinoma arising in the background endometrium. It is well established that serous carcinoma may exist as a minute foci in the endometrium devoid of myometrial invasion and still show extrauterine involvement [14-16]. In a study of EMPs with serous carcinoma involvement with no or minimal invasion, Silva et al., [17] reached similar findings: in 6 (37.5%) of 16 cases in that study, there was evidence of extra uterine involvement at presentation. The similarities between the patients who presented with advanced disease and the patients who presented with initial stage disease, suggested that serous carcinoma involving endometrial polyps may represent one aspect of a multicentric disease in which, the entire female genital tract and the abdominal peritoneal surfaces would be at high risk for concurrent or subsequent involvement by serous carcinoma even in the absence of myometrial invasion [17] or the extrauterine disease may represent transtubal metastasis [18,19].

This study also confirms previous findings that EMPs are most prevalent in the perimenopausal age group. The reason( s) for this age-segregation, which has remained remarkably consistent across various studies since the mid-fifties, is unclear. Chavez et al., [20] speculated that with the introduction of new minimally invasive technologies (such as office hysteroscopy and sonohysterograms), the demographics of patients with EMPs will change over time as younger women undergoing evaluation for infertility will have "latent" EMPs discovered. However, when the authors compared the mean ages of women with EMPs in 1990 and 1996, there was no statistically significant difference. In addition, multiple EMPs are more prevalent in the postmenopausal women (26%) as compared with their premenopausal (15%) counterparts with EMPs. These findings suggest that the factor, or the constellation of factors responsible for the above mentioned observation is intrinsic to the endometrial polyps and surrounding endometrium condition depending on the age group. Lower incidence of endometrial polyps in the younger age group may be attributed to a possible spontaneous regression mechanism, which is characteristic of the cycling endometrium in young reproductive age women.

Despite the supportive evidence of no difference in the clinico-pathological features and overall distribution of carcinomas arising from EMPs with those arising from non-polypoid endometrium, our data suggest a strong age dependent association between the presence of EMPs and involvement by endometrial carcinoma. The pathogenesis and mechanisms underlying such association are complex and not well established. Recently published data, however, provided some clue of significant differences in receptor expression, response to stimuli, and apoptosis regulation in EMPs compared to benign non-polypoid endometrium which could potentially elucidate some aspects of the possible malignant potential of EMPs. Estrogen and progesterone act as modulators of endometrial proliferation and differentiation through their receptors. Glandular epithelial expression of estrogen and progesterone receptors in polyps is not significantly different from that of the normal cycling endometrium. However, fewer stromal cells express estrogen and progesterone receptors in polyps which suggests that EMPs may result from a decrease in estrogen and progesterone receptors in the stromal cells [21]. In addition, although EMPs depend partially on estrogen receptors and grow in response to estrogen stimulation, their growth is not entirely dependent on them, this is especially so in postmenopausal women. The presence of c-erbB2 over-expression in endometrial polyps, in association with higher proliferation rates were established in a recent study [22]. This finding could explain the presence of polyps showing signs of proliferation even when the adjacent endometrium is atrophic. Thus, C-erbB2 over-expression in endometrial polyps and not in the adjacent atrophic mucosa may render polyps more sensitive to the combination of high gonadotropins and low estrogen levels, which is characteristic in the postmenopausal women.

Another significant histological finding is the glandular epithelia hyperplasia in C-erbB2 -positive polyps as opposed to rather atrophic architecture in C-erbB2 -negative polyps [22]. These findings indicate that the relationship between the expression of estrogen receptors and cell proliferation in normal endometrium and EMPs differ significantly. The balance between mitotic activity and apoptosis, which regulates normal endometrial development in EMPs also shows significant alterations. Bcl-2 is a proto-oncogene, which prolongs the cell survival by inhibiting apoptosis. Bcl-2 expression has been characterized in normal cycling endometrium. Recent studies have also observed that Bcl-2 is strongly expressed in hyperplastic and malignant endometrium [23]. A localized increase in Bcl-2 expression and consequential decline or cessation of apoptosis may be another mechanism underlying the pathogenesis of endometrial polyps [24]. Elevated Bcl-2 expression results in failure of the polyp tissue to undergo normal cycle dependent sequence of proliferation, differentiation and shedding. These data imply that the relationship between receptor expression, cell proliferation and apoptosis in normal and polypoid endometrium differ significantly. Such differences combined with the nonrandom chromosomal aberrations and monoclonality, suggest that EMPs may provide a suitable microenvironment for the development of malignancy, particularly epithelial cancers. In this aspect, the molecular and/or cytogenetic alterations inherent to EMPs in a postmenopausal background could be viewed as factors facilitating and contributing to the process of malignant transformation. Our results showed a strong association of EMPs in postmenopausal patients with endometrial cancer. It raised the possibility that EMPs in postmenopausal women could represent some intermediate stage in the development of carcinoma. A similar suggestion was proposed in a study evaluating the spectrum of pathological findings in Tamoxifen treated breast cancer patients whom develop polyps and carcinoma significantly more frequently than the general population [25]. Also in favor of this hypothesis were the results provided by Silva et al., who found that 10 (76%) of 13 Tamoxifenrelated endometrial carcinomas were associated with EMPs [26].

One potential limitation of our study is our lack of consideration of the impact of variables such as hypertension, obesity and family history. However, since data regarding such possible confounders were not available to us, we set the goals of our investigation to be primarily focused on age related distribution of coinciding morphologic findings. Although we are aware of the limitations of our study and the introduced analytical bias, drawbacks that certainly pertain to any similar retrospective pathomorphologic study, we feel that we have adequately addressed the proposed investigative tasks according to the initially set scope of the study. By using the database of Yale-New Haven Hospital, we collected and analyzed a significant number of cases over an extensive period of time and thereby our study population constituted an adequate representation of the general population in respect to the morphological parameters we investigated.

In summary, the age distribution, histological subtype distribution, and peak incidence of EMPs was similar to previous reports. In contrast, EMPs in postmenopausal women showed a significantly higher association with malignant tumor involvement. Careful microscopic search for malignancy in patients with multiple risk factors, particularly in postmenopausal women is advised in daily surgical pathology practice.