Ultrasonographic diagnosis of endometrial disorders: correlation with the histological results in 112 patients

The diagnosis of endometrial lesions hitherto was mainly made on the basis of histological examination. The objective of this study was the assessment of the diagnostic accuracy of transvaginal ultrasound investigation by comparison with the usual histological investigation, which implies painful endometrial sampling. 112 perimenopausal women with irregular vaginal blood loss were examined by transvaginal ultrasound. Shortly afterwards endometrial sampling was performed by means of aspiration, conventional curettage or hysteroscopically guided biopsies. The group of 112 patients included 11 women who had previously received hormone substitution therapy (10%). The other patients (without previous hormone substitution therapy) were divided into a premenopausal group (n = 47) and a postmenopausal group (n = 54). In case of an endometrial thickness (single layer) of less than 3 mm in postmenopausal patients, no (pre)malignant lesions were found. In the pre- and postmenopausal group, using this cut-off-level two of the five endometrial hyperplasias were ultrasonographically missed. All fourteen endometrial carcinomas in 112 patients were ultrasonographically detected by an endometrial thickness > or = 4 mm. In the postmenopausal group specificity was 73% (16/22) using a cut-off-level of 3 mm, in the premenopausal patients it was only 36% (11/31). Regular echo-density or a clear alignment between endo- and myometrium hardly had any value in the diagnosis of (pre)malignant lesions of the endometrium. If our patients with an endometrial thickness of less than 3 mm would not have had an endometrial curettage, 38 of 112 (34%) endometrial samplings might have been avoided. According to our view, transvaginal endometrial examination can be of distinct value in the detection of (pre)malignant endometrium.