Incorporating real-world scenarios during initial validation stages of point-of-care human papillomavirus (HPV) screening tests: a scoping review

Background High-risk human papillomavirus (HR-HPV) causes around 95% of cervical cancers. People in low-resource settings are more frequently affected, primarily due to lack of access to screening and vaccinations. Point-of-care screening tests using DNA-based isothermal amplification are being developed to address this, but few are tested in real-world scenarios, a critical step in evaluating their future utility. Objectives This scoping review aimed to summarise and contextualise studies developing point-of-care HR-HPV screening tests evaluated in real-world scenarios, defined as using low-resource sample preparation methods on >30 clinical samples. Methods A scoping review of studies in Web of Science and PubMed identified isothermal amplification screening tests meeting this definition. Their alignment with the WHO’s target product profile (TPP) for point-of-care tests was assessed. Results Only 13 studies—7% of total studies addressing DNA-based HPV isothermal amplification in our search—were tested in real-world scenarios. Atila’s ScreenFire and AmpFire platforms were the only platforms taken to mid-stage validation. Of these 13 studies, 12 used a combination of lysis buffer and heat for the sample preparation (1 used magnetic beads); 5 used samples from low-resource locations (8 used samples from high-resource locations); 3 used at least some vaginal or cervicovaginal samples, as recommended by the WHO (9 used cervical, 1 did not specify); and 1 used samples taken immediately (9 used stored clinical samples, 3 did not specify). Only four studies addressed three or more out of seven criteria in relation to the TPP for HR-HPV point-of-care tests. Conclusion To facilitate agreement with the TPP, we recommend that future studies aim to test their systems with freshly obtained vaginal swabs. The use of cervical swabs and preconcentration steps should be accounted for in analysis. Samples from low-resource populations should be used in mid-stage validation.