To identify the optimal viral load threshold of the in-house AML Laboratory RIATOL qPCR HPV genotyping assay (qPCR) (Antwerp, Belgium) assuring satisfactory accuracy to detect high grade cervical intraepithelial neoplasia (CIN2+).
The clinical accuracy of the qPCR to detect CIN2+ was assessed using a set of cervical samples compiled for the VALGENT-3 project. The VALGENT framework is designed to assess the analytical and clinical performance of HPV tests that offer limited to extended genotyping capability. “VALGENT-3” panel comprised 1,600 samples from Slovenian women aged 20-64 years (1,300 sequential cases from routine screening and 300 “enriched” abnormal samples – 100 HSIL, 100 LSIL and 100 ASC-US).
The VALGENT-3 panel contained 126 specimen of women with CIN2+ (used to assess sensitivity) and 1,167 specimen from women with 2 consecutive negative Pap smears (used to assess specificity). Performance relative to the Hybrid Capture 2 (HC2) was also analyzed as per the non-inferiority criteria defined by Meijer et al. in 2009. The trade-off between sensitivity and specificity with different viral load cutoffs was assessed by ROC curve analysis. The cumulative hrHPV load was defined as the logarithm of the sum of the type-specific loads of the 14 HPV types.
The qPCR had a sensitivity and specificity for CIN2+ of 97.6% (CI: 93.2-99.5%) and 85.1% (CI: 82.9-87.1%) respectively when the lowest analytical cutoff was used. At a cutoff of 1.58 log copies/cell, qPCR had a sensitivity of 96.0% (CI: 91.0-98.7%) and a specificity of 89.5% (87.6-91.2%). At this cutoff, accuracy of the qPCR was non-inferior to the HC2: relative sensitivity of 1.00 [CI: 0.97-1.03 (p<0.001)] and relative specificity of 1.00 [CI: 0.98-1.02 (p<0.001)].
HPV tests that provide viral load measurements (or other quantifiable signals) allow flexibility to optimize accuracy required for use in cervical cancer screening.