Concurrent Endometrial Cancer With Endometrial Hyperplasia or Endometrial Intraepithelial Neoplasia - Focus on Risk Factors Analysis

We retrospectively reviewed the cases of seventy-seven patients who had hysterectomies for endometrial hyperplasia between January 1996 and September 2006 at the Department of Obstetrics and Gynecology, National Taiwan University Hospital. A preoperative pathologic diagnosis of endometrial hyperplasia was obtained by D&C (dilation and curettage) in all patients. Twenty of them were diagnosed as having endometrial carcinoma in their hysterectomy specimens. All of the specimens were reviewed by a gynecologic pathologist.

Depending on the final pathologic reports of hysterectomy, we divided the seventy-seven patients into two groups - the non-endometrial carcinoma group and the endometrial carcinoma group. Fifty-seven of the studied patients were in the non-endometrial carcinoma group and twenty were in the endometrial carcinoma group. As already mentioned above, we investigated them by clinical parameters including age, menopausal status, obstetrical history, medical history of diabetes and hypertension, BMI (body mass index) and preoperative pathology of D&C.

The clinical and pathologic characteristics of the twenty patients diagnosed as having endometrial carcinoma postoperatively were also reviewed. Initial manifestation, histological grading of the carcinoma, the depth of myometrial invasion, with or without adjuvant radiotherapy and recurrence were included. The histological grading of endometrial carcinoma was based on FIGO ( International Federation of Gynecology and Obstetrics ) definitions. The BMIs of the patients with endometrial carcinoma were also analyzed by the receiver operating characteristic curve. The Research and Ethics Committee of National Taiwan University Hospital approved this study.

Statistical analyses were performed with the Independent-Samples T test and the Mann-Whitney test. Subsequently, multivariate analysis of risk factors was also employed with Binary logistic regression model to obtain the adjusted odds ratio (OR) and 95% confidence interval (CI) for selective variables. P<0.05 was defined as significant.