207 Summary and general discussion comprehensively compare their performance in different sample types. Methylation markers for endometrial cancer detection were based on both our feasibility study described in Chapter 2 (GHSR, SST, ZIC1) and the systematic review described in Chapter 3 (ADCYAP1, BHLHE22, CDH13, CDO1, GALR1, HAND2). Paired samples were collected from endometrial cancer patients (n=103) and compared with unpaired samples of healthy controls (n=317). Optimal three-marker combinations yielded a high diagnostic performance for endometrial cancer detection in urine (AUC 0.95), cervicovaginal self-samples (AUC 0.94), and clinician-taken scrapes (AUC 0.97). Diagnostic performances remained virtually equal after cross-validation and for earlystage endometrial cancer detection. The outcomes of this study demonstrated that endometrial cancer detection in home-collected samples was excellent and comparable to the diagnostic performance in clinician-taken cervical scrapes. In Chapter 5, we explored the feasibility of ovarian cancer detection by molecular testing in urine, cervicovaginal self-samples, and clinician-taken cervical scrapes. Nine methylation markers were tested, which were selected from previous studies on ovarian cancer detection in cervical scrapes and plasma (C2CD4D, CDO1, NRN1) and cervical and endometrial cancer detection in patient-friendly sample types (GALR1, GHSR, MAL, PRDM14, SST, ZIC1). Paired samples were collected from women diagnosed with a benign (n=25) or malignant (n=29) ovarian mass and compared with unpaired samples of healthy control women (n=110). Increased methylation levels were found when comparing ovarian cancer patients with healthy controls in full void urine (C2CD4D, CDO1, MAL), urine supernatant (MAL), and cervical scrapes (C2CD4D, CDO1). Methylation levels of GHSR also discriminated between benign and malignant ovarian masses in the urine sediment. No elevated methylation signals were found in cervicovaginal self-samples of ovarian cancer patients. We also demonstrated that urine contains ovarian cancer-derived DNA by somatic copy number analysis. Copy number aberrations were detected in 4 out of 23 sequenced urine samples of ovarian cancer patients. This pioneering work encourages further development of urine biomarkers for ovarian cancer detection. The outcomes of Part 1 revealed the value of methylation analysis in patient-friendly sample types for endometrial cancer detection of all stages. Convenient modes of sample collection offer the possibility of at-home collection with high patient acceptability. This approach is clinically useful to screen patient populations at risk for endometrial cancer and to streamline who needs to undergo invasive endometrial tissue sampling. Although promising, the clinical effectiveness of this approach requires further confirmation in additional cohorts, including individuals presenting with postmenopausal bleeding and asymptomatic women at risk for endometrial cancer. The presence of ovarian cancer-derived DNA in the urine provides the first 8
RkJQdWJsaXNoZXIy MjY0ODMw