54 Chapter 3 diagnostic test and an AUC of 0.5 implicates the absence of diagnostic potential. Both the individual DNA methylation markers and marker panels investigated in the included studies were summarized, irrespective of their diagnostic accuracy. Individual DNA methylation markers achieving an AUC value of ≥ 0.80 were considered highly valuable for the minimally invasive detection of EC and selected from eligible studies. Data extraction Data from selected studies were extracted from the full text by two reviewers (RvdH and JAvT). Collected data were processed using a standardized data registration form reporting the following information: first author and research group, year of publication, journal and belonging impact factor, marker identification methods (e.g. genome-wide screen, a targeted approach or literature analyses), study design (e.g. discovery, test or validation set), study population (i.e. the total number of cases and controls, and tumor subtypes included), patient characteristics (i.e. presented with symptoms or detected during a screening), sample type, DNA methylation markers studied, assay used for DNA methylation detection, outcome measures (i.e. percentage methylated in cases and controls) and marker performance (i.e. AUC value, sensitivity, specificity and belonging 95% confidence interval [CI]). Risk of bias assessment The risk of bias was independently assessed as low, high or unclear according to the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) guidelines by two reviewers (RvdH and BMMW). Scoring was piloted in two independent studies (i.e. not included in this review) to ensure reproducibility. Disagreements between the two reviewers were resolved by discussion with a specialist (RDMS). The risk of bias assessment scores were merely used to determine the quality of selected studies and not to exclude articles from the review. In case a study performed marker discovery as well as marker validation, only the latter was assessed for bias. Furthermore, quality assessment was only focused on the validation of markers analyzed in minimally invasive collected cytological specimens (i.e. cervical scrapes, endometrial brushes, vaginal swabs and vaginal tampons). A figure summarizing the risk of bias scores per study was constructed using Review Manager 5.3 software. RESULTS Search results The literature search and selection process are outlined in Figure 1. A total of 1556 potentially relevant articles were retrieved from PubMed, Embase.com, Web of Science
RkJQdWJsaXNoZXIy MjY0ODMw