The effectiveness of HPV16 and HPV18 genotyping and cytology with different thresholds for the triage of human papillomavirus-based screening on self-collected samples

Background/Objective Human papillomavirus (HPV) genotyping and cytology have been recommended for colposcopy triage, but it is unclear which combinations of high-risk HPV (hrHPV) types and cytology with various thresholds provide clinically useful information for the triage after primary HPV screening on self-collected samples. Method Chinese Multi-site Screening Trial (CHIMUST) database focused on self-collected samples was reviewed using the results of Cobas4800 HPV assay. Absolute risks of each genotype for cervical intraepithelial neoplasia 2 or worse/ 3 or worse (CIN2+/CIN3+) were calculated. Triage of atypical squamous cells of undetermined significance (ASCUS) or worse cytology was used as the comparator, and diagnostic accuracy for paired comparisons between algorithms was obtained using McNemar's test. Results A total of 10, 498 women were included, the overall prevalence of hrHPV, HPV16, HPV18, and Other hrHPV genotypes were 13.7%, 2.4%, 0.8%, and 10.5%, respectively. HPV16-positive women had the highest absolute risk among various genotypes for CIN2+/CIN3+ whether in normal or abnormal cytology (ASCUS or worse) and among all age groups. When compared with the comparator, combining HPV16 positivity and/or high-grade squamous intraepithelial lesion (HSIL) or worse yielded higher specificity (97.7% vs. 97.0%, p<0.0001), similar sensitivity (90.7% vs. 96.3%, p = 0.256) for detection of CIN3+, and a decrease in colposcopy referral rate from 3.5% to 2.7%, similar results were found for CIN2+. Positivity for HPV16 and/or (ASCUS or worse), and positivity for (HPV16 and/or HPV18) and/or (ASCUS or worse) achieved favorable sensitivity compared with the comparator (80.6% and 81.3% vs. 70.1% respectively for CIN2+, p<0.0001; both 96.3% vs. 96.3% for CIN3+, p = 1.000), these algorithms would reduce the colposcopy referral rate to 5.0% and 5.6% respectively, compared with 13.7% of that for HPV alone. Conclusions Triage of HPV-positive women on self-collected samples by combining HPV16 or HPV16/18 genotyping with different thresholds of cytology could provide tradeoffs in sensitivity for detecting cervical lesions and colposcopy referral rates, and tailor management in various circumstances of clinical practice.