Urine and Vaginal Self?Sampling for HR?HPV and CIN2+/CIN3+ Detection in Cervical Cancer Screening: A Narrative Review of Diagnostic Accuracy, Heterogeneity, and Implementation Considerations

Background: Cervical cancer remains a major global health burden, disproportionately affecting women in low- and middle-income countries where screening coverage is limited. While HPV-based screening is highly effective, clinician collected sampling poses logistical and sociocultural barriers. Self-sampling approaches, including vaginal and urine-based HPV testing, have emerged as potential strategies to expand access. However, uncertainties remain regarding their comparative diagnostic accuracy and optimal implementation conditions. This narrative review synthesizes evidence on the diagnostic accuracy of urine and vaginal self-sampling versus clinician collected samples for detecting cervical HPV infection and cervical intraepithelial neoplasia grade 2 or worse (CIN2+). Methods: A narrative review of diagnostic accuracy studies published up to March 2026 was conducted. PubMed/MEDLINE was systematically searched and supplemented by manual screening of reference lists. Studies were eligible if they evaluated HPV detection using urine and/or self-collected vaginal samples, included a paired clinician collected cervical sample, and reported at least one of the following reference standards: histologically confirmed CIN2+ or CIN3+, histologically confirmed HSIL+, or cervical HR-HPV DNA detection. Data were synthesized separately by endpoint, and then compared across population settings (screening, referral, and high-risk groups), assay type, cut off strategies, and pre-analytical conditions such as first void versus random urine and preservation methods. Results: Across 51 studies, urine-based HR-HPV testing showed moderate-high sensitivity for CIN2+/CIN3+ (about 75% - 80%) but modest specificity (around 50% - 60%), with best performance when first void urine, proper preservation, and PCR-based assays were used. Vaginal self-sampling generally achieved higher sensitivity (often >90%) and better specificity (about 65% - 85%), approaching clinician collected samples when PCR assays and optimized cut offs were applied. For HSIL+, evidence was sparse: one large study reported almost perfect accuracy for urine testing, and no robust HSIL only data were available for vaginal sampling. For HR-HPV detection using cervical HPV as reference, urine reached sensitivities around 75% - 85% and specificities near 90%, while vaginal self-sampling exceeded 90% sensitivity with slightly lower specificity. Conclusion: Self-sampling with vaginal and urine specimens shows clinically meaningful accuracy for both HR-HPV and CIN2+/CIN3+ detection,