DNA methylation-based detection and prediction of cervical intraepithelial neoplasia grade 3 and invasive cervical cancer with the WID (TM)-qCIN test
Abstract
Background: Cervical screening using primary human papilloma virus (HPV) testing and cytology is being implemented in several countries. Cytology as triage for colposcopy referral suffers from several shortcomings.
HPV testing overcomes some of these but lacks specificity in women under 30. Here, we aimed to develop and validate an automatable triage test that is highly sensitive and specific independently of age and sample heterogeneity, and predicts progression to CIN3+ in HPV+ patients.
Results: The WID (TM)-qCIN, assessing three regions in human genes DPP6, RALYL, and GSX1, was validated in both a diagnostic (case-control) and predictive setting (nested case-control), in a total of 761 samples. Using a predefined threshold, the sensitivity of the WID (TM)-qCIN test was 100% and 78% to detect invasive cancer and CIN3, respectively.
Sensitivity to detect CIN3+ was 65% and 83% for women = 30 years of age. The specificity was 90%.
Importantly, the WID (TM)-qCIN test identified 52% of >= 30-year-old women with a cytology negative (cyt-) index sample who were diagnosed with CIN3 1-4 years after sample donation. Conclusion: We identified suitable DNAme regions in an epigenome-wide discovery using HPV+ controls and CIN3+ cases and established the WID (TM)-qCIN, a PCR-based DNAme test.
The WID (TM)-qCIN test has a high sensitivity and specificity that may outperform conventional cervical triage tests and can in an objective, cheap, and scalable fashion identify most women with and at risk of (pre-)invasive cervical cancer. However, evaluation was limited to case-control settings and future studies will assess performance and generalisability in a randomised controlled trial.