Thesis

85 DNA methylation testing for endometrial cancer detection in patient-friendly sample types patients as compared to samples of healthy control women (Figure 1). The discriminatory power of all markers was verified at a tissue level by comparing methylation levels of endometrial cancer tissue of various histological subtypes with normal endometrium (Mann–Whitney U test; all P < .01; Supplementary Figure 2). Table 1: Characteristics of endometrial cancer patients and controls. n (%) Age: median (IQR) Endometrial cancer 103 (100%) 66 (60-71) Histology Endometrioid carcinoma 52 (51%) Grade 1 20 (19%) Grade 2 17 (17%) Grade 3 15 (15%) Serous carcinoma 29 (28%) Clear cell carcinoma 7 (7%) Uterine carcinosarcoma 10 (10%) Mixed carcinoma* 3 (3%) Uterine sarcoma 2 (2%) FIGO stage IA 51 (50%) IB 21 (21%) II 3 (3%) III 18 (18%) IV 10 (10%) Controls Urine 100 61 (55-78) Self-sample 107 60 (60-60) Scrape 110 60 (55-60) *Patients with endometrial carcinomas of mixed subtypes included one mixed clear cell and endometrioid carcinoma, and two mixed clear cell and serous carcinomas. Correlation of individual markers between sample types was assessed for paired samples from patients with endometrial cancer. DNA methylation levels of six markers (ADCYAP1, BHLHE22, CDO1, GHSR, SST and ZIC1) were moderately to strongly correlated between the different sample types (Spearman correlation; r = .43 – .80). Interestingly, while the remaining three markers (CDH13, GALR1, and HAND2) correlated strongly (Spearman correlation; r = .72 – .77) between urine and self-samples, a poor correlation was observed between urine and cervical scrapes (Spearman correlation; r = .01 – .14) and self-samples and cervical scrapes (Spearman correlation; r = .04 – .09). 4

RkJQdWJsaXNoZXIy MjY0ODMw