An ultrasound risk-scoring model for prediction of endometrial cancer in post-menopausal women (using IETA terminology)

Objective: Objective was to use terms and definitions described by the International Endometrial Tumor Analysis group (IETA) in evaluating the endometrium on gray scale ultrasound and Doppler imaging for testing the validity of these ultrasound variables in addition to patient clinical variables and creating a risk scoring model for better prediction of endometrial cancer trying to limit invasive procedures for risky cases only and performing a decision algorithm for their management. Study design: A prospective cross-sectional study, 60 patients with post-menopausal Bleeding (PMB) and endometrial thickness ≥5 mm were enrolled. TVS and color Doppler were done. Patient clinical variables and ultrasound endometrial assessment were compared to histopathology to identify the best predictors of endometrial cancer. Univariate, multivariate, and ROC curve analysis were used to design a risk-scoring model then test its validity. Results: The best predictors for endometrial cancer were endometrial thickness ≥12 mm (OR = 1.9), heterogeneous echogenicity of endometrium (OR = 2.7), irregular endometrial midline (OR = 1.9), ill-defined endometrium-myometrium interface (OR = 3.3), Doppler score > 2 (OR = 2.5). These variables were used to perform a risk-scoring model for the prediction of endometrial malignancy with accuracy about 95% and an area under the curve 0.9. At the best cut-off value (score ≥ 10) sensitivity and specificity were 80%, 97.8%, respectively. Conclusions: Terms used by IETA group were clinically valuable and reasonable in assessment of the endometrium. Applying a malignancy model for women with PMB allowed easy their risk classification, better identification, and management.