The aim was to assess the value of three-dimensional (3D) ultrasound to distinguish between benign and malignant endometrium in women with postmenopausal bleeding (PMB).
Methods: 62 patients with PMB and endometrial thickness >4.5 mm underwent transvaginal 3D ultrasound examination. The volume, vascularization index (VI), flow index (FI), and vascularization-flow index (VFI) were calculated for the endometrium and a 2 mm "shell" surrounding the endometrium using VOCALTM software. Histological diagnosis was obtained. Receiver operating characteristic (ROC) curves were drawn to evaluate an ability of measurements to distinguish between benign and malignant endometrium. Logistic regression analysis was used to predict endometrial malignancy.
Results: There were 49 benign and 13 malignant endometria. Endometrial thickness and volume were larger and flow indices were higher both in the endometrium and in the endometrial shell in malignant endometria than in benign endometria (p<0.05). The best cut-ff value for endometrial thickness to predict malignancy was 11.8 mm (sensitivity 0.85, specificity 0.71, LR+ 3.0, LR- 0.2). Endometrial thickness had an area under the ROC curve of 0.82. The best logistic regression model to predict malignancy contained endometrial thickness (odds ratio 1.2, p<0.009) and VI in the endometrial shell (odds ratio 1.1, p<0.016) as predicting variables. The area under the ROC curve for the best logistic regression model was 0.86, and the best risk cut-off for this model (0.22) had a sensitivity of 0.69, a specificity of 0.86, LR+ 4.8, LR- 0.4. Conclusions: The diagnostic accuracy of 3D ultrasound was only marginally superior to that of endometrial thickness.