Agreement Between High-Risk Human Papillomavirus Testing in Paired Self-Collected and Clinician-Collected Samples from Cervical Cancer Screening in Spain

Abstract

Background: Implementing self-sampling (SS) in cervical cancer screening requires comparable results to clinician-collected samples (CCS). Agreement measures are essential for evaluating HPV test performance. Previous studies on non-paired samples have reported higher viral cycle threshold (Ct) values in SS compared to CCS, affecting sensitivity for detecting cervical intraepithelial neoplasia grade 2 or worse (CIN2+). Objectives: We aimed to evaluate the agreement of high-risk (hr)HPV testing results between SS and CCS using paired samples and to explore differences in Ct values. Methods: Women aged 30 to 65 years attending cervical cancer screening in two regions of Spain were invited to participated in this study. For each woman there was: CCS collected during the screening visit using liquid-based cytology and cytobrush, and a SS using a brush at home one month later. A PCR-based assay was used for hrHPV detection. Agreement in hrHPV results among both samples, Ct value differences, and their association with screening outcomes were analyzed. Results: This study included 981 women with paired samples. SS had a higher hrHPV prevalence than CCS (overall ratio of 1.3). Positive agreement for all hrHPV genotypes, HPV16, HPV18, and other hrHPV types were 85%, 91.3%, 66.7%, and 83.3%, respectively. Negative agreement was >95% for all results. Median Ct values was slightly higher in SS than in CSS (32.9 vs. 30.6, p = 0.02). Seven CIN2+ cases HPV positive were detected by both methods. One CIN3 case was missed by SS. Conclusions: This study showed a good agreement between SS and CCS for hrHPV testing in a routine screening in Spain. Despite the slightly higher Ct values for SS, no significant impact on sensitivity could be determined due to the low incidence of CIN2+ cases. Further research on larger paired samples is needed to assess the implications of Ct values on test sensitivity.