40 Chapter 2 the preservation of genetic material, when handled correctly (36, 38, 39). This enables delivery to a testing laboratory per mail. Urine consists of a heterogeneous collection of cell components. We evaluated three urine fractions (full void, sediment, and supernatant) to determine the most optimal source of DNA for EC detection by methylation analysis, assuming that the urine supernatant mainly contains cell-free DNA fragments, and the urine sediment largely consists of cellular DNA (19). Despite this supposed varying origin of DNA in the different urine components, DNA methylation analysis showed significantly increased methylation levels of all markers in all urine fractions of EC patients as compared to controls. Different urine fractions showed strong to very strong correlations (r ≥ 0.77 – 0.92). Similar findings have been described for the detection of cervical cancer (22, 23) and bladder cancer (25) in different urine fractions. When comparing the AUC values of all fractions, full void urine shows the highest potential for EC detection. An advantage of using full void urine, instead of urine sediment or urine supernatant, is that this fraction does not require pre-processing of the urine sample. Current routine EC diagnostics are facing several challenges and limitations for which urine could offer a potential solution. Transvaginal sonography remains insufficient in distinguishing benign and malignant endometrial lesions, with a specificity that ranges from 36 to 68% among symptomatic women (40). Apart from its limited specificity, not all endometrial malignancies present with thickened endometrium (41, 42), and the optimal cut-off of endometrial thickness that demands further examination is still under debate (43-45). As a result, many women undergo invasive endometrial tissue sampling. This biopsy procedure can be hampered by conditions that hinder access to the uterus (e.g. cervical stenosis or discomfort) or may yield insufficient tissue for diagnosis (46). Urine testing could not only reduce the need of performing invasive biopsies, but also has potential in screening of asymptomatic women or to triage women presenting with postmenopausal bleeding symptoms. Additionally, accurate DNA methylation marker testing in urine could be useful to monitor women with increased EC risk (e.g. women with Lynch syndrome). Among women at risk of developing EC, serial sampling of urine may offer an alternative for repeated invasive testing. Urine sampling for EC detection may also be valuable in developing countries with limited access to effective screening programs and early detection methods. These encouraging results warrant further research to determine whether DNA methylation testing in urine meets the requirements for consideration as a diagnostic tool applicable to clinical practice in the management of EC. Currently, our sample
RkJQdWJsaXNoZXIy MjY0ODMw