SS 06-03Host and HPV methylation – adjunctive biomarkers that improve CIN3 diagnosis specificity

13. Molecular markers
A.T. Lorincz 1, N. Vasiljevic 1, A.R. Brentnall 1, C. Reuter 1, R. Banwait 1, R. Warman 1, L. Cadman 1, J. Austin 1, J. Cuzick 1.
1Centre for Cancer Prevention, Wolfson Institute of Preventive Medicine, Queen Mary University of London. (United Kingdom)

Background / Objectives

HPV DNA tests have excellent sensitivity for CIN3+ but a drawback of viral screening is modest to poor specificity. Women infected with high risk (hr) HPV need triage before referral to colposcopy to reduce adverse outcomes and high costs. Current approaches to triage are cytology, HPV16/18 genotyping, and p16/ki-67 staining. Our main goal has been to develop and validate a superior DNA methylation biomarker panel to triage hrHPV infected patients to colposcopy.


Methods

710 women were selected from a study of 6,000 women attending routine screening. Colposcopy follow-up of cytologically abnormal women within a year of triage revealed 38 with CIN2+. We measured DNA methylation of human gene EPB41L3 and selected genomic regions of HPV16, HPV18, HPV31 and HPV33. The methylation results were combined into a classifier, called S5. Mann-Whitney P, logistic regression log-likelihoods, and areas under the curve (AUC) were calculated, sensitivity and specificity were compared by McNemar’s test. 


Results

At the pre-defined cut-off, S5 showed better sensitivity for CIN2+ than HPV16/18 genotyping (74% vs 54%, P=0.04) in hrHPV positive women, and similar specificity (65% vs 71%, P=0.07). When sensitivity was made equal, S5 had a much higher specificity (91%) than genotyping (p<0.0001). The AUC of S5 was 0.78 (0.69-0.88) for CIN2+ and 0.84 (0.72-0.97) for CIN3+ (p<0.001).


Conclusion

The novel S5 classifier was developed in two colposcopy populations, Predictors 1 and 2 and validated in a screening sample from Predictors 3. Methylation measurement appears to be a consistent and valid approach to triage and can be done reflexively on screening specimens. Additional validation studies, including archived specimens from the FOCAL HPV screening trial in British Columbia are underway.  


References