Background: Recently, disagreements remain in increasing evidence about the potential value of circulating cell-free DNA (cfDNA) as a noninvasive diagnostic biomarker for ovarian cancer (OC). Here, this update meta-analysis was performed to further assess the diagnostic performance of circulating cfDNA in discriminating OC from non-cancerous individuals.

Methods: We performed a systemic literature search of PubMed, Embase, Web of Science, Cochrane Library, OVID, Chinese National Knowledge Infrastructure (CNKI) and Wanfang databases to obtain 22 eligible articles including a total of 1125 patients and 1244 controls. The pooled sensitivity, specificity, diagnostic odds ratio (DOR) and area under receiver operating characteristics curves (AUROC) of the included studies for cfDNA in diagnosing OC patients were used to estimate the diagnostic value. The clinical utility of cfDNA was evaluated by Fagan nomogram. Heterogeneity was explored utilizing subgroup analysis and meta-regression.

Results: The pooled sensitivity and specificity were 73% and 90%, the DOR and AUROC were 25.29 and 0.90, respectively. Subgroup analyses and meta-regression, according to patients' region, study design, clinical stage, specimen types, detection indicators, simple size, publication year revealed there were no significant sources of heterogeneity. Additionally, subgroup analyses showed qualitative detection (methylation detection); TNM stage I-IV, publication year 2011-2018, serum-based cfDNA assays exhibited better diagnostic performance as compared to quantitative detection, TNM stage III-IV, publication year 2002-2010; plasma-based cfDNA assays, and more participants and prospective studies manifested superior diagnostic accuracy. The result of sensitivity analysis indicated no study exclusively contributed to the heterogeneity and Deeks' funnel plot suggested no evidence of significant publication bias.

Conclusions: Our meta-analysis found the qualitative detection (methylation); TNM stage I-IV, publication year 2011-2018 were related to more effective diagnostic accuracy for OC. However, serum-based cell-free DNA detection should be cautiously interpreted due to unclear factors. Hence, further large-scale longitudinal studies are required to validate the diagnostic potential of cell-free DNA. The present study provides to accrue knowledge of cell-free DNA levels for future researches.