Studies on the diagnosis of endometrial cancer in women with postmenopausal bleeding

Abstract

This thesis aims to evaluate the diagnostic work-up in postmenopausal women presenting with abnormal vaginal bleeding. In the guideline of the Society of Dutch Gynaecologist and Obstetricians a gynecological examination, including cervical cytology, is followed by transvaginal sonography (TVS). When the endometrial thickness (ET) is 4 mm or less, the patient is reassured. In case of abnormal findings in the cervical cytology, or in case the ET is more than 4 mm or not assessable, endometrial sampling is indicated. When feasible an office endometrial sample is taken, else curettage preferably with hysteroscopy should be performed. We evaluated the guideline with respect to adherence, investigated the yield of cervical cytology and studied the final diagnosis in women with non-diagnostic office endometrial samples. We then focused on the prediction of endometrial cancer in women with postmenopausal bleeding in relation to patient characteristics, and subsequently if a strategy using both patient characteristics and TVS results could optimize the diagnostic process. The Dutch Study of Postmenopausal Bleeding (DUPOMEB) study group collected data on 913 women with postmenopausal bleeding in eight hospitals. Adherence to the guideline was present in more than 70% of the patients. The efficiency of the diagnostic process can however increase through limiting histological examination to women with ET > 4mm, and by relying on endometrium aspiration in these women. Cervical cytology is mandatory particularly to reveal cervical lesions, although a minority of endometrial cancers in women with ET < 5 mm are detected though cervical smear. The presence of normal endometrial cells in the cervical smear was not predictive for endometrial cancer. A non-diagnostic office endometrial sample was concluded on in 66 patients. Further evaluation of the endometrium was performed with hysteroscopy and/or curettage, or (uneventful) follow-up. Since three endometrial cancers and one atypical hyperplasia (6%) were discovered it was concluded that a technically well-performed office endometrial sample that is non-diagnostic does not rule out endometrial cancer, and necessitates further endometrial sampling. The performance of TVS in women with diabetes, hypertension and obesity was studied. In diabetic or obese women the mean endometrium thickness was higher than in women without these risk factors unless (pre)malignancy was found. Most women with diabetes and obesity will therefore need endometrial sampling. Presence or absence of hypertension had no impact on the accuracy of TVS. The efficiency of the diagnostic workup of postmenopausal bleeding can be improved when patient characteristics, i.e. age, the use of anticoagulants, parity and the presence or absence of diabetes, and overweight, are combined with ET. Incorporation in clinical practice might be through an algorithm or a clinical decision model, reducing the number of invasive procedures, without compromising accuracy. In endometrial cancer tumor type and grade, and depth of myometrial infiltration are important to decide upon adjuvant treatment. Preoperative selection of patients at high risk of pelvic lymph node metastases was performed by TVS assessment of myometrium infiltration in combination with tumor grading of the histologic specimen. Correct preoperative selection of low-stage, high risk patients was achieved in 80%.