Endometrial carcinoma and vaginal ultrasound: studies of the diagnostic validity

Abstract

Up to now, there has been no effective diagnostic procedure for early diagnosis of endometrial cancer. During recent years, however, study results on transvaginal sonography (TVS) of the endometrium have been published recommending TVS as a screening method for early diagnosis of endometrial cancer in postmenopausal women. The purpose of this study was to determine the sensitivity and specificity of TVS in measuring endometrial thickness and the resulting performance under screening conditions. Patients with postmenopausal bleeding were included in a prospective blind diagnostic study. Endometrial thickness was measured with TVS before D&C. The histological findings of D&C were used as a control. Sensitivity and specificity were calculated after combining the separately and blindly determined findings. The study comprised 195 patients. Endometrial cancer was diagnosed in 29 of these patients, a prevalence of 15%. The median endometrial thickness was 5.5 mm (0-10 mm) in women with endometrial cancer and 4 mm (0-15 mm) in women without cancer (p > 0.05 Mann-Whitney U-test). For a cut-off > or = 4 mm, the sensitivity is 62% and the specificity is 50%. The sensitivities and specificities for all cut-offs are displayed in a receiver operating characteristic (ROC) curve. The relationship between the positive predictive values and the different prevalences is demonstrated in another curve. Using the results of this and other comparable studies, only very low positive predictive values can be calculated (0.03-1.9%) with prevalences expected under screening conditions (0.02-0.12%). Transvaginal sonography is not a useful method in early diagnosis of endometrial cancer.