Objective  To evaluate the clinical value of different combination strategies of high-risk HPV (hr-HPV) testing and Thinprep cytology test (TCT), a cervical cytology test, for cervical cancer screening, especially for high or higher-grade squamous intraepithelial lesion (HSIL⁺) in Shuangliu District, Chengdu City.

  Methods  The study is a population-based randomized clinical trial. Women aged 35 to 65 years meeting the inclusion criteria were enrolled for the study. At the baseline screening conducted in the first year, the participants were randomly assigned to either cytology test or hr-HPV testing at a ratio of 1∶2. If the paticipants had positive results for the baseline hr-HPV test, they would then undergo either cytology test or colposcopy by random assignment. After 24 months, all participants were called back, and combined screening of cytology test and hr-HPV test were performed. Women who had negative results at baseline screening and who entered and completed the third-year follow-up were selected as the subjects of the study. Based on the aforementioned testing findings, the related data were extracted and four different screening protocols were simulated: 1) combined TCT and hr-HPV screening, with referral for colposcopy when there was positive results for either one of the two; 2) combined TCT and hr-HPV screening, with referral for colposcopy when both tests had positive results at the same time; 3) TCT was done for preliminary screening and those who were found to be positive would then undergo hr-HPV test for triage purpose, with subsequent referral made for colposcopy if the hr-HPV results were positive; 4) hr-HPV was done for preliminary screening and those who were found to be positive would then undergo TCT, with subsequent referral made for colposcopy if TCT results were positive. With the detection of HSIL⁺ on histological examination as the endpoint event, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under curve (AUC) of different combination screening models were calculated.

  Results  A total of 3102 women were screened, and 2967 women were included in the statistical analysis in this study. Among the 2967 women, 979 were randomized to cytology and 1988 to hr-HPV genotyping. For prescreening, the positive rate of the cytology group was 5.6% (55/979), with of HSIL⁺ positive rate being 0.2% (2/979), while the positive rate of the hr-HPV group was 7.5% (149/1988), with HSIL⁺ positive rate being 0.9% (18/1988). After 24 months, 2456 women were called back and were given cervical cytology test and hr-HPV test at the same time. Among them, the positive rate of the cytology group was 3.2% (78/2456), while the positive rate of hr-HPV group was 8.7% (215/2456). The overall positive rate of HSIL⁺ was 0.69%(17/2456). Women with a negative baseline hr-HPV had a lower incidence of HSIL⁺ lesions in the long term. The strategy of cervical cytology screening combined with hr-HPV test for triage purpose is the best method, with a sensitivity of 88.9%, a specificity of 58.3%, a PPV of 44.4%, a NPV of 93.3%, and an AUC of 0.736, P=0.039 (95% CI: 0.555-0.917).

  Conclusion  This randomized clinical trial from Shuangliu District, Chengdu City shows that the sensitivity of hr-HPV testing is better than that of cytology test, and the prevalence of HSIL⁺ in women with negative baseline hr-HPV results is lower than that of women with negative baseline cytology results. The screening program of TCT for prescreening plus subsequent hr-HPV test for triage purpose shows better value for the detection of HSIL⁺.