Accuracy of high-risk HPV testing and HPV16/18 genotyping to triage women with LSIL: a pooled analysis of VALGENT studies

Background / Objectives
Genotyping for the most carcinogenic HPV types (HPV16 and HPV18) could identify those women at highest risk requiring colposcopy or more intensive follow-up in women with low-grade squamous lesions (LSIL).
Results
The VALGENT framework is designed to assess the analytical and clinical performance of HPV tests that offer limited to extended genotyping capability. VALGENT is iterative using panels collated in different countries. A pooled analysis was performed, using data from three completed VALGENT panels, to assess the diagnostic accuracy of genotyping for HPV16/18 and for hrHPV (13 or 14 types) to detect prevalent CIN2+ in women with LSIL. Data pooling was performed using a bivariate normal model designed for meta-analysis of diagnostic test accuracy, taking the intrinsic negative correlation between sensitivity and specificity into account.
Conclusion
Twenty HPV tests were evaluated within three VALGENT panels. The pooled sensitivity and specificity of hrHPV in aggregate to detect CIN2+ was 98.1% (95%CI: 95.5 -99.2%) and 23.8% (95%CI: 20.6-27.3%) in women with LSIL, respectively. HPV16/18 genotyping had a sensitivity and specificity for CIN2+ of 56.2% (CI: 51.0-61.3%) and 77.1% (CI: 73.0-80.8%), respectively. HPV16/18 genotyping was substantially more specific (ratio: 2.65, 95%CI: 2.13-3.28) but also less sensitive than testing for hrHPV (ratio: 0.62, 95%CI:0.57-0.68). No significant inter-panel 27/490differences were observed either for the pooled analysis of hrHPV test accuracy or theHPV16/18 genotyping. The average risk of underlying CIN2+ was 39.7% in HPV16/18-positive women with LSIL, 13.1% in women who were HPV16/18-negative but positive for other hrHPV types and 2.1% for hrHPV-negative women.