Quantitative Measurement of L1 Human Papillomavirus Type 16 Methylation for the Prediction of Preinvasive and Invasive Cervical Disease

Background Methylation of the human papillomavirus (HPV) DNA has been proposed as a novel biomarker. Here, we correlated the mean methylation level of 12 CpG sites within the L1 gene, to the histological grade of cervical precancer and cancer. We assessed whether HPV L1 gene methylation can predict the presence of high-grade disease at histology in women testing positive for HPV16 genotype. Methods Pyrosequencing was used for DNA methylation quantification and 145 women were recruited. Results We found that the L1 HPV16 mean methylation (±SD) significantly increased with disease severity (cervical intraepithelial neoplasia [CIN] 3, 17.9% [±7.2] vs CIN2, 11.6% [±6.5], P < .001 or vs CIN1, 9.0% [±3.5], P < .001). Mean methylation was a good predictor of CIN3+ cases; the area under the curve was higher for sites 5611 in the prediction of CIN2+ and higher for position 7145 for CIN3+. The evaluation of different methylation thresholds for the prediction of CIN3+ showed that the optimal balance of sensitivity and specificity (75.7% and 77.5%, respectively) and positive and negative predictive values (74.7% and 78.5%, respectively) was achieved for a methylation of 14.0% with overall accuracy of 76.7%. Conclusions Elevated methylation level is associated with increased disease severity and has good ability to discriminate HPV16-positive women that have high-grade disease or worse.