Diagnostic accuracy of available methylation assays in advanced cervical intraepithelial neoplasia from high-risk HPV-positive women: A systematic review and network meta-analysis

Over the past decade, methylation has developed rapidly for the detection of multiple diseases, and several methylation assays have been studied and even applied in clinical practice. This study undertook diagnostic test accuracy (DTA) and network meta-analysis (NMA) to investigate the value of extensively validated methylation assays for use in clinical practice or research for triage of advanced cervical intraepithelial neoplasia (CIN) among high-risk human papillomavirus (HPV)-positive women. PubMed, Web of Science, the Cochrane Library and Scopus were searched for eligible studies. DTA and NMA were conducted using R Version 4.2.0 with a random-effects model. Twenty-eight studies with 16,256 patients were included. The DTA results showed that the pooled sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of all methylation assays for CIN grade 3+ (CIN3+) were 0.708 [95 % confidence interval (CI) 0.676-0.738], 0.780 (95 % CI 0.736-0.819), 0.436 (95 % CI 0.348-0.528) and 0.920 (95 % CI 0.885-0.945), respectively. The diagnostic odds ratio of CIN3+ was 8.828 (95 % CI 7.109-10.962), which was higher compared with that for CIN2+ (6.115, 95 % CI 4.604-8.123). NMA revealed that most methylation assays included in this study performed similarly to cytology. Among the available methylation assays, S5 classifier has a balanced performance in sensitivity and specificity overall. In conclusion, methylation assays are an effective and accurate triage strategy for advanced CIN among high-risk HPV-positive women. S5 classifier seems to be promising due to its triage performance. Cervi-M and GynTect are suitable for application in developing countries due to their superior specificity and PPV. However, more studies are needed to confirm these conclusions.