Background: Mucinous differentiation of the endometrium can occur in a spectrum of changes ranging from benign (metaplasia) to malignant (adenocarcinomas with mucinous differentiation). Mucinous proliferations with simple architecture are generally considered benign; however, more complex growth patterns have an uncertain biologic behavior, particularly when these changes are focal and/or are encountered in biopsy or curettage material. The disparity between the degree of cytologic atypia and the neoplastic potential makes their interpretation difficult in routine practice. We categorized and prospectively studied a series of these lesions based upon their degree of architectural complexity and correlated them with follow-up curettings and hysterectomies over a period of three years.
Methods: Mucinous proliferations of the endometrium were subdivided into three categories (A, B, or C) based upon increasing degrees of architectural complexity. Type A were mucin-containing epithelial cells, present singly or in small tufts, within architecturally benign glands or involving the endometrial surface. Type B proliferations were more complex, consisting of mucin-containing epithelial cells forming small pseudoglands with rigid, punched out spaces and no supporting stroma Conspicuous cytologic atypia or architectural features such as a filiform growth pattern characterized type C alterations. One hundred two curettings and 36 hysterectomies from 52 patients were reviewed.
Results: Patient's ages ranged from 39 to 71 years (median, 55 yr); 41 patients (80%) were over age 50. Twenty patients (40%) were receiving hormone replacement therapy. Nineteen type A, 17 type B, and 16 type C mucinous endometrial proliferations were analyzed. Excluding those cases in which a conventional coexisting precancerous lesion was also present in the initial endometrial sample, the percentages of endometrial carcinoma following a curettage diagnosis of types A to C, respectively, were 0, 64.7%, and 100%. Carcinomas following type B alterations were all well-differentiated and all were confined to the endometrium or inner third of the myometrium.
Conclusion: Mucinous endometrial proliferations comprise a spectrum subdivisable into biologically meaningful subsets. A high percentage of type B alterations were found to have endometrial adenocarcinoma on follow-up; however, all were well-differentiated and showed either no or minimal invasion. This finding suggests that the absence of cytologic atypia in complex mucinous lesions identifies subsets of lesions at low concurrent risk for deeply invasive cancer. The presentation of type B lesions as predominantly microglandular surface lesions without co-existing atypical hyperplasias suggests that a subset of well-differentiated adenocarcinomas arise via neoplastic alterations in surface epithelium.