Infectious Agents and Cancer journal has recently launched a new collection of papers about “Point-of-Care (POC) for HPV-related genital cancers” putting together some interesting works on the accuracy of HPV tests for screening.
In this article we took the opportunity to reflect on the relations between accuracy measures, prevalence and characteristics of the tested population in the case of HPV-based screening. In screening test evaluation, we look at the clinical accuracy of the test as an intrinsic characteristic of the assay, which interacts with the characteristics of the population, the result being the screening performance. In the case of HPV testing, the clinical accuracy should be conceptualized in two steps, the analytical accuracy of the assay for HPV infection and the biological link between HPV infection and the target disease, i.e. the high-grade cervical intraepithelial neoplasia (hgCIN). This approach highlights that just a few false positive cases result from a lack of analytical specificity while most derive from women who have the infection but it did not progress to hgCIN. In addition, increasing prevalence of hgCIN results in relevant increases of PPV only if due or associated with exposures which increase the progression from infection to hgCIN or the duration of the latter; while an increase due to a higher prevalence of HPV infection would only marginally affect PPV. This approach may help in modelling the performance of HPV-based cervical screening.