Investigation of post-menopausal bleeding : which ultrasound features are most helpful?

Abstract

PURPOSE: The aim of this dissertation was to find which ultrasound features of the endometrium could be predictive of endometrial cancer (EC), or alternatively of other benign diagnoses, in women presenting with post-menopausal bleeding (PMB). METHODS: A prospective study was performed on 47 women with PMB and endometrial thickness (ET) >=4mm. The endometrium was evaluated at transvaginal ultrasound (TVUS) with Doppler imaging and gel-infusion sonohysterography (GIS), and the findings were correlated with histological diagnosis obtained at out-patient endometrial sampling (ES) or at hysteroscopy (HY). Univariate and multivariate analysis by logistic regression were performed. RESULTS: EC was diagnosed in 5.8% (4/69) of all women presenting with PMB. ET was found to be significantly predictive of EC (p=0.004). The endometrial outline as seen after GIS was also found to be a significant predictor of EC (p=0.008). These two predictors together with multiple vessel vascular pattern of the endometrium on Doppler imaging (p=0.055) and body-mass-index (BMI) (p=0.078) were incorporated into a logistic regression model for predicting EC. Logistic regression modelling revealed only endometrial outline (p=0.003) and ET (p=0.032) to be significantly predictive of EC, when 4 variables were combined. A non uniform endometrial morphology was predictive of a diagnosis of endometrial polyp (p=0.038). GIS revealed the presence of focal intra-cavity lesions in 63.8% of the women with ET>=4mm. Out-patient ES missed 60% of these focal intra-cavity lesions. CONCLUSION: All US features identified as predictive, could be used in combination to estimate EC risk. GIS is a very useful addition to the TVUS investigation as it allows assessment of endometrial outline which is a strong predictor of EC. US-based and statistically derived risk-scores for EC should be one of a series of tests incorporated into clinical algorithms for management of women with PMB.