Clinical performance of hrHPV primary screening using vaginal versus cervical samples to detect high-grade intraepithelial lesions.

High-risk human papillomavirus (hrHPV) testing is now the most recommended primary method for cervical cancer screening worldwide. Clinician-collected cervical sampling continues to be the main sampling method, but hrHPV vaginal self-sampling is an appealing alternative because of its greater acceptability and potentially higher cost-effectiveness. This study aimed to determine if hrHPV vaginal self-sampling is comparable to clinician-collected cervical sampling for detecting histologically confirmed high-grade cervical intraepithelial neoplasia (CIN2/3) as part of a cervical cancer screening program in Mexico. We analyzed data from 5,856 women screened during a hrHPV-based screening study. Clinical performance and diagnostic efficiency metrics were estimated for the two sampling methods for the CIN3 and CIN2+ endpoints, using three triage strategies: HPV16/18 genotyping, HPV16/18/33/58 extended genotyping, and HPV16/18/31/33/58 extended genotyping. hrHPV-positivity was found in 801 (13.7%) cervical and 897 (15.3%) vaginal samples. All women with hrHPV-positive samples were referred to colposcopy, which detected 17 total CIN3 cases before considering retrospective triage strategies. Using the HPV16/18/31/33/58 extended genotyping strategy, 245 women had hrHPV-positive cervical samples and 269 had hrHPV-positive vaginal samples. Ten CIN3 cases were detected each among women with hrHPV-positive cervical samples and among those with hrHPV-positive vaginal samples when using this strategy, with no significant differences in sensitivity and specificity observed. We observe that self- and clinician-collected sampling methods are comparable for detecting CIN3 and CIN2+ regardless of the triage strategy used. These findings can help public health officials to develop more cost-effective cervical cancer screening programs that maximize participation.