RESUMEN
BACKGROUND: The development of efficient strategies for managing high-risk human papillomavirus (HR-HPV)-positive women is a major challenge when human papillomavirus-based primary screening is being performed. The objectives of this study were to evaluate the comparative effectiveness of HR-HPV testing based on self-collection (SC) and HR-HPV testing based on collection by a health professional (HP) and to assess the potential usefulness of HR-HPV testing combined with testing with the biomarkers p16/Ki-67, α-mannosidase, and superoxide dismutase 2 (SOD2). METHODS: This was a cross-sectional study of 232 women admitted for colposcopy because of an abnormal Papanicolaou smear. The collected material underwent liquid-based cytology, HR-HPV detection, and immunocytochemical testing (p16/Ki-67, α-mannosidase, and SOD2). The gold standard was the histopathological result; the positive reference was CIN2+. RESULTS: The overall accuracy of HR-HPV testing was 76.6%; the results for the SC group (78.1%) and the HP group (75.2%) were similar. The positive predictive values (HP, 76.5%; SC, 80.0%), the negative predictive values (HP, 66.7%; SC, 64.3%), the positive likelihood values (HP, 1.35; SC, 1.36), and the negative likelihood values (HP, 0.21; SC, 0.19) were also similar. p16/Ki-67 showed higher sensitivity than the other 2 biomarkers: 78.1% versus 45.8% for α-mannosidase and 44.5% for SOD2. The specificities of the biomarkers were equivalent: 71.4% for p16/Ki-67, 77.8% for α-mannosidase, and 71.2% for SOD2. In the HP group, accuracy also leaned more heavily toward the final score (using α-mannosidase and SOD2) without statistical significance (80.8% vs 77.9%). The contrast with the SC group yielded the same level of accuracy. CONCLUSIONS: SC, when associated with testing with biomarkers, is as accurate as collection by HPs in the detection of women at risk for cervical cancer.