Abstract
High-risk human papillomavirus (hrHPV) infection is the primary cause of cervical cancer. However, hrHPV testing lacks specificity in detecting neoplastic changes. This study explored the utility of quantitative methylated HPV DNA markers for precise detection of cervical lesions. Using hybridization capture-based bisulfite sequencing, we analyzed genome-wide HPV methylation patterns. The study included a training cohort of 60 cervical exfoliated cell samples and a validation cohort of 29 samples. Analysis of 112 CpG sites across the HPV genome revealed that genome-wide HPV16 methylation levels correlated with disease progression. Squamous cell carcinoma (SCC) showed 1.4-fold higher methylation levels compared to normal tissue (p = 0.0032). Progressive methylation increases in the E5-α and L2 genes were observed across the spectrum of cervical lesion severity, from normal tissue through high-grade squamous intraepithelial lesion (HSIL) to SCC. Intersection analysis of differentially methylated CpG sites between HSIL vs Normal and SCC vs Normal identified 16 consistently hypermethylated CpG sites in the E5-α, E7, L2, and L1 genes, distinguishing both HSIL and SCC from normal tissue. This pilot study identifies a five-CpG methylation panel (E5-α_3887, E5-α_3941, E7_701, L2_4441, and L2_5128) as promising triage biomarkers for HPV16-positive women, achieving high discriminatory performance (AUC = 0.919 in a validation cohort) for detecting cervical lesions. This genome-wide capture sequencing identified novel HPV16 methylation markers that distinguish cervical lesions from normal tissue, supporting the feasibility of HPV methylation-based triage for HPV-positive women in cervical cancer screening.