Can human papillomavirus DNA testing of self-collected vaginal samples compare with physician-collected cervical samples and cytology for cervical cancer screening in developing countries?

Background: To determine human papillomavirus (HPV) types by polymerase chain reaction (PCR)-reverse line blot assay and examine the concordance between HPV by Hybrid Capture 2 (HC2) and PCR on self-collected vaginal and physician-collected cervical samples and cytology. Methods: This was a cross-sectional study of 546 sexually active women aged ≥30 years with persistent vaginal discharge, intermenstrual or postcoital bleeding or an unhealthy cervix. Participants self-collected vaginal samples (HPV-S) and physicians collected cervical samples for conventional Pap smear and HPV DNA (HPV-P) testing and performed colposcopy, with directed biopsy, if indicated. HPV testing and genotyping was done by HC2 and PCR reverse line blot assay. Concordance between HC2 and PCR results of self- and physician-collected samples was determined using a Kappa statistic (κ) and Chi-square test. Results: Complete data were available for 512 sets with 98% of women providing a satisfactory self-sample. PCR detected oncogenic HPV in 12.3% of self- and 13.0% of physician-collected samples. Overall, there was 93.8% agreement between physician-collected and self-samples (κ = 76.31%, 95% confidence interval [CI]: 64.97–82.29%, p = 0.04)—complete concordance in 473 cases (57 positive, 416 negative), partial concordance in seven pairs and discordance in 32 pairs. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of self-sampling for detection of cervical intraepithelial neoplasia (CIN)2+ disease were 82.5%, 93.6%, 52.4% and 98.4%, respectively; for physician-sampling they were 87.5%, 93.2%, 52.2% and 98.9%, respectively; and for cytology they were 77.5%, 87.3%, 34.1% and 97.9%, respectively. Concordance between HC2 and PCR was 90.9% for self-samples (κ = 63.7%, 95% CI: 55.2–72.2%) and 95.3% for physician-collected samples (κ = 80.4%, 95% CI: 71.8–89.0%). Conclusions: Self-HPV sampling compares favourably with physician-sampling and cytology. A rapid, affordable, HPV self-test kit can be used as the primary method of cervical cancer screening in low-resource situations.