Transvaginal ultrasound measurement of the endometrium remains the first line test for investigating postmenopausal bleeding but integration of patient characteristics into testing may further improve diagnostic algorithms.

Abstract

Postmenopausal bleeding (PMB) is a frequent medical problem that accounts for approximately 5% of all gynaecological consultations. In 10% of women with PMB, endometrial cancer is the underlying cause, which makes endometrial cancer the most common malignancy of the female genital tract in developed countries (Moodley & Roberts, J Obstet Gynaecol 2004;24:736–41). Adoption of transvaginal ultrasound (TVS) measurement of endometrial thickness as the first line diagnostic test in PMB algorithms has minimised the need for more invasive endometrial evaluation with endometrial sampling and hysteroscopy. However, the endometrial thickness cut-off to exclude endometrial cancer and the need for further testing remains the subject of debate. Four systematic reviews and meta-analyses have been published on this subject with differing conclusions (Gupta et al. Acta Obstet Gynecol Scand 2002;81:799–816, Smith-Bindman et al. JAMA 1998;280:1510–7, Tabor et al. Obstet Gynecol 2002;99:663–70, Timmermans et al. Obstet Gynaecol 2010;116:160–7]. Some of these meta-analyses suffer from the fact that authors of smaller studies titrated an optimal cut-off, thereby maximising the specificity for a sensitivity of almost 100%, whereas others suffer from partial verification bias because the need for endometrial biopsy was conditional upon the endometrial thickness measurement by TVS. The study by Wong et al. published in this issue of the journal is unique because it describes the largest cohort of women presenting with PMB where invasive assessment of the endometrium has been done in all women independent from the results of the preceding TVS endometrial thickness measurement. Indeed, the optimal threshold that combines an almost perfect sensitivity with an acceptable specificity appears to be 3 mm. At this cut-off point, positive and negative predictive values were 6.5 and 0.26%. Higher cut-off values, e.g. 5 mm, would result in positive predictive values of over 10% but a negative predictive value of 0.35%. Awareness of these trade-offs can inform decision making. The prevalence of endometrial cancer among women with PMB in this cohort from Hong Kong was 3.7%, which is much lower than the previously assumed 10%. Whether this is due to the large unselected population or due to differences in the incidence of endometrial cancer in an Asian population is unclear and requires further study. Although strategies for investigating PMB based upon TVS endometrial thickness thresholds are well established, further refinement is possible by integrating patient characteristics and thereby the individual risk of endometrial cancer, into decisions about the need and type of testing required. Other studies have shown that increasing age, older age at menopause, BMI, nulliparity, diabetes and recurrent bleeding increase the risk of endometrial cancer. These data of Wong et al. can help further refine and validate existing prediction models. In this way, pre- and post-test probabilities of endometrial cancer could be maximised, which in turn will reduce the harm done and wasting of resources arising from unnecessary testing. None declared. Completed disclosure of interests form available to view online as supporting information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.