Background

Endometrial polyps (EMPs) are generally considered benign proliferative lesions and are commonly encountered in routine surgical pathology practice. The usual histological pattern of endometrial polyps is characterized by irregular proliferative glands, with a fibrotic stroma containing thick-walled blood vessels [1]. The morphologic diversity of endometrial polyps is reflective of the morphologic spectrum of the background endometrium from which EMPs arise. As such, EMPs may range from atrophic to hyperplastic to carcinomatous. However, opinions differ on whether EMPs are intrinsically a marker for concurrent or subsequent malignancy. Endometrial polyps were identified in 12–34% of uteri containing endometrial carcinoma in two earlier studies [2,3]. In another case-control study examining previous pathology in women diagnosed with endometrial carcinoma, endometrial polyps were twice as likely to be detected than in the control group [4]. Rarely, serous endometrial intraepithelial carcinoma (EIC), the presumptive early form of uterine papillary serous carcinomas, may be identified as very minute foci in EMPs [5,6]. This finding may be interpreted as the EMP homologue of similar changes that are occasionally identified in non-polypoid atrophic endometrium. However, given that nonrandom chromosomal aberrations and monoclonality that have been demonstrated in EMPs [7,8], an alternate interpretation is that molecular and/or cytogenetic alterations inherent to EMPs facilitate a neoplastic transformation. The latter interpretation would imply that endometrial polyps are a risk factor for the development of endometrial tumors. And indeed a possible association between endometrial polyps and endometrial malignancy in postmenopausal women has been suggested couple of decades ago [4]. However, there is no direct evidence for a greater propensity of polypoid endometrium to undergo malignant change as compared to the adjacent normal endometrium, and EMP may simply represent am embodiment of the greater propensity of the host endometrium to develop proliferative/neoplastic changes in general [9]. A recent study designed and conducted to investigate the pathological significance of EMPs and their association with pre-malignant and malignant conditions failed to supply evidence of such association. That study involved a large cohort of patients seen in outpatient hysteroscopy clinic for abnormal uterine bleeding. To determine the magnitude of malignant potential among polyps, the authors compared the pathological findings in polyps with non-polypoid specimens. The comparative analysis established that endometrial hyperplasia was more frequent in endometrial specimens with polyps, but the incidence of frank carcinoma in polypoid and non-polypoid endometrium remained the same. Although not age stratified, the study showed that in abnormal uterine bleeding, hyperplasia presented more frequently in women with EMPs compared to those without polyps, but cancer involvement regardless of the histological pattern was not significantly different [9]. Similar results and failure to establish any association of endometrial polyps and carcinoma were demonstrated in another recent study dealing with endometrial polyp characteristics in menopausal women on hormonal replacement therapy [10].

Most standard pathology texts list endometrial polyps as being most prevalent in perimenopausal women and suggest possible association between polyps and malignant involvement [9,11]. However, there has been no detailed age-based analysis of the incidence and malignant involvement of EMPs. In this report, age-related differences in the incidence of EMP at the time of diagnosis in the practice of a busy academic center is examined, with a detailed analysis of the incidence and histologic subtypes of malignancies associated with EMPs. The objectives of the study are 1) to investigate the age-group in which EMP are most commonly encountered in routine surgical pathology practice 2) to document the age-group in which EMPs are most commonly associated with malignancies 3) To investigate whether the age of incidence of the various carcinoma subtypes in EMPs at the time of diagnosis is congruent with published data on similar malignancies arising in non-polypoid endometrim and 4) To investigate whether the histologic subtype distribution of malignancies associated with EMPs is similar or significantly different from the distribution of malignancies arising from non-polypoid endometrium.

Patients and methods

Case retrieval and pathologic classifications

All cases with a diagnosis of EMP were retrieved from the computerized database of the Pathology Department at Yale-New Haven Hospital for the 10-year-period from 1986 to1995. All cases were further investigated for involvement of endometrial cancers including malignancies without myometrial invasion; histologic subtypes of all malignant tumors were catalogued. Histologic types of endometrial malignancies were characterized according to the WHO classification [12]. All cases were reviewed microscopically and confirmed by a second pathologist. For endometrial cancer with mixed histologic type, the presence of a second component was considered if it involved more than 10% of all available sections containing tumor. The cases of endometrial malignancy involving both EMPs and non-polyp endometrium, were classified into the category of EMPs with cancer involvement.

Patients groups

For comparative purposes, the patients were divided into 5 age groups: 25–35; 36–45; 46–55; 56–65; and >65 years; and each group was further classified based on an association (or lack thereof) with endometrial carcinoma.

A starting point of 25 years of age was selected due to the very low incidence of endometrial polyps in patients below this age. Two patients (ages 18 and 19) were excluded from the study as they did not represent a sufficiently large for statistical analysis group. The proportion of both groups (polyps associated with malignancies (malignant polyps) and polyps not associated with malignancies (benign polyps) were statistically compared for each of the aforementioned age-groups. Subsequently, we merged the younger age groups and preserved the >65 year group, which we referred to as "postmenopausal", since significant differences were observed in this specific subset of patients. Larger age groups were arbitrarily designated as reproductive years (25–45), perimenopause (shortly before or after menopause, 46–65) and postmenopause (>65) and statistically analyzed. The postmenopausal status of all women above the age of 65 was verified and the term "postmenopausal" was occasionally used when referring to this particular age group. Our use of this term, although arbitrary, was important in order to put the emphasis on the fact that any pathomorphological findings in this age group are unlikely to be attributed to changes characteristic of the cycling endometrium.

Statistical analysis

Chi-square test was used to compare the proportion of malignancy associated EMPs between the age groups and in regards to particular histological type of malignancies.

Results Out of all diagnostic and therapeutic procedures performed over this period, a total of 513 EMP were identified. The latter included 304 (59%) endometrial biopsies/ curetting samples, and 209 (41%) hysterectomy specimens. In cases in which endometrial biopsies and hysterectomies both showed presence of EMP, only the hysterectomy specimen was considered. The age of patients ranged from 18–91 years with a median of 54 years. Sixty-six (13%) of 513 EMPs were malignant. The histological subtype distribution of those 66 malignancies included 58 endometrioid (87%), 6 serous (9%), 1 carcinosarcoma, and 1 clear cell carcinoma. No mixed histological type of endometrial cancer was found in our series. The incidence of EMP peaked at age group 46–55 years, which was similar to previous reports. In age group 25–35, only 1 (2.5%) of 40 EMPs was associated with endometrial malignancy. In contrast, 37 (32%) of 115 EMPs were associated with malignancy in the age group >65 years (Figure 1; Figure 2). The frequency of EMPs with malignancy involvement increased with age and reached statistical significance (p < 0.001) in the age group >65 years (Figure 3). The most common histological type of malignancy was endometrioid carcinoma, followed by serous carcinoma. The same statistically significant difference for age group >65 years (p < 0.05) remained when larger age groups, including reproductive (25–45), perimenopausal (46–65) and postmenopausal (>65) patients were compared (Figure 4).