Background: Infections with high-risk human papillomaviruses (Hr-HPV) can cause malignant transformation of the human cervical epithelium. HPV DNA tests generally are very sensitive to detect cervical neoplastic lesions but also identify transient HPV infections. As a consequence, the specificity and positive predictive value are low.
Methods: We analyzed viral load of Hr- and possibly Hr-HPV types more than seven orders of magnitude (on a log10 scale) in 999 consecutive BD-SurePath liquid-based cervical cytology samples from routine cervical screening enriched with atypical squamous cells of undetermined significance (n = 100), low-grade squamous intraepithelial lesions (LSIL; n = 100), and high-grade squamous intraepithelial lesions (HSIL; n = 97) using type-specific multiplex quantitative real-time PCR and the BSGP5+/6+-PCR/MPG assay. In the 36-month follow-up, 79 histologically verified CIN2+ and 797 double-negative cytology cases were identified.
Results: Viral loads in LSIL and HSIL were significantly increased compared with no intraepithelial lesion or malignancy in both the quantitative PCR (qPCR) and BSGP5+/6+-PCR/MPG assay (P < 0.0001). The mean viral loads in LSIL and HSIL were not significantly different. Using a newly determined high viral load cut off for 14 Hr-HPV types, the sensitivity for prevalent CIN3+ remained at 100% for both assays compared with the minimal detection threshold. The specificity (corresponding to double-negative cytology at subsequent screening episodes) increased substantially (qPCR, from 91.1% to 95.7%; BSGP5+/6+-PCR/MPG, from 79.8% to 96.2%).
Conclusions: Compared with DNA positivity alone, high Hr-HPV viral loads could reduce the amount of false positive results detected by the BSGP5+/6+-PCR/MPG and qPCR by 81.4% and 52.1%, respectively.
Impact: Quantitative type-specific HPV DNA assays show high flexibility in defining thresholds that allow optimizing clinical accuracy for cervical cancer precursors.