LIU Yuting, DUAN Suchun, LI Hongye, SHAN Shuzhi, ZHAO Xin, JIANG Jing, SHI Bin.
To compare the consistency and feasibility of self-sampling for human papillomavirus (HPV) with physician-sampling as a method for cervical cancer screening. Methods 2 033 women aged 30 to 59 who involved in cervical cancer screening were selected for HPV self-sampling and physician-sampling, respectively, the collected specimens were tested for Cobas 4800 HPV and Seq HPV test. Any HPV positive was referred to colposcopy and cervical biopsy. The consistency of HPV positivity rates between self-sampling and physician-sampling, as well as the accuracy in detecting cervical intraepithelial neoplasia (CIN)2 or higher lesions (CIN2+) and CIN3+, were analyzed using histopathology as the gold standard. Results ① Among the 2 033 women, 249 cases (12.2%) were high-risk HPV positive of any type, There were 29 cases (1.4%) of histopathological CIN2+ and 13 cases (0.6%) of CIN3+. The detection rates for CIN2+ were consistent between Cobas-physician-sampling and self-sampling (28/2 033, 1.4%) as well as Seq-physician-sampling and self-sampling (Kappa=0.84). The detection rates for CIN3+were also consistent between Cobas-physician-sampling and self-sampling (13/2 033, 0.6%) as well as Seq-physician-sampling and self-sampling (Kappa=0.87). ② The sensitivity for detecting CIN2+was similar among Cobas-physician-sampling (96.6%), Cobas-self-sampling (96.6%), Seq-physician-sampling (96.6%), and Seq-self-sampling (100%) (P>0.05). The specificity for detecting CIN2+was not significantly different between Cobas-physician-sampling (92.2%), Seq-physician-sampling (92.6%), and Seq-self-sampling (92.4%) (P>0.05), but Cobas-self-sampling (90.9%) showed statistical significance (P<0.05). The negative predictive value (NPV) for detecting CIN2+was not significantly different among Cobas-physician-sampling (99.9%), Cobas-self-sampling (99.9%), Seq-physician-sampling (99.9%), and Seq-self-sampling (100%) (P>0.05).③ The sensitivity for detecting CIN3+was 100% for all four methods, with no statistically significant differences (P>0.05). The specificity for detecting CIN3+was not significantly different between Cobas-physician-sampling (91.5%), Seq-physician-sampling (91.9%), and Seq-self-sampling (91.7%) (P>0.05), but Cobas-self-sampling (90.2%) showed statistical significance (P<0.05). The NPV for detecting CIN3+was not significantly different among Cobas-physician-sampling (100%), Cobas-self-sampling (100%), Seq-physician-sampling (100%), and Seq-self-sampling (100%) (P>0.05). Conclusions HPV self-sampling demonstrated similar detection rates and sensitivity for detecting high-grade cervical lesions compared to physician-sampling, making it a viable option for population-based cervical cancer screening. The Seq HPV test showed similar sensitivity and NPV to the Cobas 4800 HPV test and can be used for HPV self-sampling.