Page 153 - The diagnostic work-up of women with postmenopausal bleeding

cancer. However, there is no consensus in systematic reviews or in international guidelines about the sequence in which these methods should be employed in women with PMB.
2. Which different prediction models on the chance of endometrial cancer in women with PMB are available in literature and which model shows the best performance?
In Chapter 3, we systematically reviewed the literature to map the prediction
models available on this subject and to identify their performance in internal
validation. It would be useful to identify women who display a high risk of having
endometrial cancer based not only on the measurement of endometrial thickness
but also on the patient’s characteristics. Many researchers have studied this subject
and have developed prediction models consisting of different predictor variables.
We identified nine prognostic studies, of which we assessed the quality, the different
phases of development and their performance.The nine articles investigated 27
different possible predictor variables. It was found that the most important patient characteristics are age, body mass index, diabetes, frequency of bleeding, use of anticoagulants and hormones. Endometrial thickness, endometrial morphology and
endometrial border were identified as significant ultrasound variables. In the three
articles studying the use of Doppler for predicting endometrial cancer, endometrial
colour score and vascularity index were identified as the most important predictor
variables. None of the detected models completed external validation or impact
analysis. Models including power Doppler showed the best performance in internal
validation, but Doppler is not easily accessible in general gynaecological practice.We
can conclude that the first step in the approach of women with postmenopausal 8 bleeding should be to distinguish between women with low risk versus high risk of
having endometrial carcinoma based on patient’s characteristics.
3. Is a prediction model based on patient characteristics useful in daily practice to differentiate between women with a high and a low risk of endometrial cancer?
External validation of a prediction model is the first step towards implementing the model in clinical practice. In Chapter 4 we externally validated a mathematical model, which is based on patient characteristics with or without TVS. We retrospectively
Summary / samenvatting
151