HISTOMETRIC ANALYSIS OF PREMALIGNANT ENDOMETRIAL LESIONS

Adenocarcinoma of the endometrium is the most common gynecologic malignancy in the United States. It accounts for about 36,000 cases of invasive cancer each year and is among the top five neoplasms affecting women. Hyperplastic abnormalities of the endometrium follow a continuum of severity, with the risk of progression to carcinoma being related to the complexity of the lesion. At this time dilation and curettage with hysteroscopy, remain the major means of diagnosis. The current management of atypical endometrial hyperplasia is limited to hysterectomy. The overall goal of this research is to establish the foundation for the objective characterization of hyperplastic endometrial lining abnormalities associated with neoplasia of the uterine corpus, based on computer histometric/karyometric analysis. We hypothesize that: 1) The orderly progression from normality, to simple hyperplasia, atypical hyperplasia, and endometrial carcinoma in the uterine lining has distinct morphometric characteristics that can be identified using computer-based histometric analysis. 2) Endometrial adenocarcinoma causes distinct identifiable changes in adjacent hyperplastic and otherwise normal endometrial surfaces, which permit its identification prior to hysterectomy. 3) The responsiveness of atypical hyperplasia to a chemopreventive regimen can be objectively quantified using histometric/karyometric analysis. Our specific aims are to: 1) Develop procedures for the automated digital processing of high resolution microscopic images of histopathologic sections of endometrial biopsies and hysterectomy specimens, including the development of a knowledge file for automated scene segmentation and karyometry. 2) Establish an objective, numeric assessment and grading system for endometrial lesions. 3) Define a nuclear signature of preneoplastic endometrial lesions. These will be accomplished within the context of an ongoing two-part study of atypical hyperplasia (GOG Protocol 167). The ubiquity of preneoplastic abnormalities of the endometrial lining of the uterus, and the potential morbidity, time loss and expense associated with their surgical therapy, require the development of novel diagnostic modalities that reliably predict malignant potential. Digital microscopy and nuclear morphometric analysis will extend the utility of the histopathologic analysis of otherwise negative endometrial biopsy material, by allowing the identification of women with synchronous endometrial carcinoma, or at risk for malignant progression of their hyperplasia. This will allow the development and testing of chemopreventive interventions in low risk patients, and facilitate efficient surgical therapy in those women with potentially aggressive disease. Such capabilities could prevent unnecessary surgery, as well as the vast majority of advanced endometrial cancers. The technique of computerized morphometric analysis has tremendous potential to increase our understanding of the biological behavior of endometrial cancer, as well as revolutionize its diagnosis and treatment endometrial cancer. This proposed research project fits within the context, of a highly structured, multi-faceted, Mentored Patient Oriented Research Career Development Plan that focuses on the prevention and control of preneoplastic gynecologic disease. It utilizes the unique resources available at the University of Arizona, to foster an environment in which academic productivity is optimized and the necessary supervision is received to insure a successful career in clinical investigation.