To validate the HPV viral loads that are reflected by the cycle threshold values of Cobas4800 as the viral load indicators by verifying the consistency of the viral loads per unit (10,000 cells) from the BMRT assay.
The analysis is based on data from the Chinese Multi-Center Screening Trial (CHIMUST). The cases included in the analysis are all positive for physician-collected hrHPV on SeqHPV and/or Cobas4800 or negative for hrHPV but abnormal in cytology (≥LSIL), and some cases selected by nested case-control randomization from those negative for physician-collected hrHPV and cytology. With HPV testing results and relevant Ct values from Cobas4800 available, we tested the entire sample set with the BMRT HPV testing assay and analyzed their agreement with Cobas4800, followed by a comparison of the CtV from Cobas4800 and viral loads (lg) from BMRT by lesion grade.
We included 4,485 women (mean age: 45.4 years) in the study, and 4,290 had complete data. The consistency of genotypes from Cobas4800 and BMRT for hrHPV, HPV-16, HPV-18, and 12-HPV pools was 94.9% (4070/4290, Kappa = 0.827), 99.1% (4251/4290, Kappa = 0.842), 99.6% (4,273/4,290, Kappa = 0.777), and 95.3% (4,089/4,290, Kappa = 0.821), respectively. Further analysis shows that any inconsistency between the two assays is likely among samples with comparatively lower viral loads. When analyzing per lesions of CIN2+ and CIN3+, the CtV from Cobas4800 and VL (lg) from BMRT are highly correlated inversely and follow the linear regression for HPV16 and 12-HPV pool (Pearson's or Spearman's correlation coefficient (r): In CIN3+, r HPV16 = −0.641,
The CtV from Cobas4800 and the viral loads (lg) of per unit cells from the BMRT are well correlated for lesion grading when tested on physician-collected samples. Cobas-CtV is worthy of further study for clinical application.