Evaluation Of Secondary Screening Strategies For HPV-Based Cervical Cancer Screening On Self-Samples

Objective:HPV-based cervical cancer screening on self-samples improves the screening coverage,whereas referral of women positive for primary hrHPV screening leads to some unnecessary colposcopies,thus it’s vital to establish an effective secondary triage strategy.The aim of this study was to explore the utility of HPV genotyping,Cobas 4800 HPV Cycle threshold value(Ct-value,reflecting viral load)and p16INK4a immunostaining(p16)in the triage of HPV-positive self-samples,and provide scientific basis for the improvement of population-based cervical cancer screening strategy.Methods:1.The Chinese Multi-site Screening Trial(CHIMUST)was a cervical cancer screening trial led by our center to evaluate the efficiency of HPV screening on self-samples.A general population including 10,000 women were screened via using two HPV tests on self-sampling and physician-sampling,LBC testing was also performed.For any HPV-positive woman,colposcopy-directed biopsy and pathological diagnosis were performed.This study was focused on self-collected samples using the results of Cobas 4800 and LBC.The following analysis was conducted:(1)Absolute risks of each genotype for cervical intraepithelial neoplasia 2/3 or worse(CIN2+/CIN3+)were calculated.Triage of atypical squamous cells of undetermined significance(ASC-US)or worse cytology was used as the comparator,and the diagnostic accuracy of triage strategies derived from HPV 16/18 and different cytological grades was evaluated.(2)The quartiles of Ct-values for Cobas 4800 HPV genotypes were used for grouping and developing screening algorithms.Screening efficacy was analyzed.Diagnostic accuracy for paired comparisons between algorithms was obtained using McNemar’s test.2.A total of 73699 women were screened in Henan Cervical Cancer Screening project via network assisted community model and NGS(Seq)HPV assay on self-samples.HPV-positive women who met any sequential positive result of HPV16/18 or VIA or p16 were referred for colposcopy/biopsy/histologic diagnosis.Based on this cervical screening project,the following studies were conducted:(1)HPV-positive women with adequate data were included for analysis.The accuracy of 35 strategies derived from p16,HPV genotyping and cytology was evaluated.A triage strategy was considered favorable if the negative predictive value(NPV)for CIN3+was≥98%,combined with improvement of sensitivity and specificity in comparison to the comparator,being the‘ASC-US+’strategy or the current guideline strategy(HPV 16/18+or ASC-US+);(2)HPV-positive,cytology-negative(HPCN)women were enrolled for analysis.A triage strategy was considered acceptable if the NPV for CIN3+was≥98%,combined with an improvement of sensitivity and specificity for CIN2+/CIN3+in reference to the comparator,being‘HPV16/18+’as triage.Results:1.A total of 10,498 women were included from the CHIMUST,the overall prevalence of hrHPV,HPV16,HPV 18,and other hrHPV detected by Cobas 4800 HPV assay were 13.7%,2.4%,0.8%,and 10.5%,respectively.HPV 16-positive women had the highest absolute risk among various HPV genotypes for CIN2+/CIN3+whether in normal or abnormal cytology(ASC-US or worse)and among all age groups.When compared with the comparator,combining HPV 16 positivity and/or high-grade squamous intraepithelial lesion(HSIL)or worse yielded higher specificity and similar sensitivity for detection of CIN2+/CIN3+,and a decrease in colposcopy referral rate from 3.5%to 2.7%.Positivity for HPV16 and/or(ASC-US or worse),and positivity for(HPV16 and/or HPV18)and/or(ASC-US or worse)achieved favorable sensitivity compared to 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 from 13.7%for HPV alone to 5.0%and 5.6%respectively.2.In the CHIMUST,the Ct-values of Cobas 4800 HPV16 and other hrHPV were inversely correlated with the severity of cervical lesions(P<0.001).Risks for CIN2+/CIN3+were significantly stratified by Ct-values from channels detecting HPV 16 and other hrHPV.‘HPV with HPV16/18 and reflex Ct-values<33.7’(algorithm G)achieved a favorable sensitivity to‘HPV with ASC-US or worse(≥ASC-US)’(81.9%vs 70.1%for CIN2+,P<0.001),a comparable sensitivity to‘HPV with HPV16/18 reflex cytology ≥ASC-US’(81.9%vs.81.3%for CIN2+,P>0.05),and resulted in a slightly lower specificity than the latter two algorithms(92.6%vs.97.4%and 95.4%respectively for CIN2+,P<0.05).However,algorithm G achieved a comparable sensitivity to HPV testing alone for CIN3+,and reduced the colposcopy referral rate from 13.7%for HPV testing alone to 8.4%.3.A total of 3510 HPV-positive women were included into the analysis from Henan project;422(12.0%)CIN2+and 247(7.0%)CIN3+were identified,the rates of p16+ and ASC-US+were 36.3%and 22.2%,respectively.‘p16+’ triage was more sensitive and less specific than‘ASC-US+’triage(P<0.0001).However,when combined p16 with LBC or HPV genotypes,two were superior to the‘ASC-US+’ strategy:p16 scored 3+;HPV16/3 3/5 8/31+&pl6+.Moreover,four strategies were favorable to the recommended guideline strategy:ASC-US+or p16+;LSIL+ or p16+;HPV16+or p16+;HSIL+or p16+or HPV16+.These triage strategies caught equal or more diseases without increasing the number of colposcopies.4.In Henan project,a total of 136(5.0%)CIN2+and 53(1.9%)CIN3+were detected among 2,731 HPCN women enrolled.Five triage strategies were superior to the‘HPV16/18+’ strategy:p16+;HPV16/33+;‘HPV16+or HPV33/58/31/35+&p16+’:‘HPV16/33+or HPV58/31/35+&p16+’;HPV16/18/31/33/45/52/58+&p16+.Moreover,three triage strategies,the‘p16+’,the‘HPV16+or HPV33/58/31/35+&p16+’,and the‘HPV 16/18/31/33/45/52/58+&p16+’,were favorable in sensitivity and/or specificity to the‘HPV16/33+’ strategy.Conclusions1.Triage of HPV-positive women on self-samples could be optimized by combining HPV16 and/or HPV 18 with different cytological diagnosis,which provided tradeoffs in sensitivity for detecting cervical lesions and colposcopy referral rate.2.HPV viral loads reflected by Cobas 4800 HPV Ct-values were associated with the severity of cervical lesions.Other hrHPV Ct-values with an appropriate cut-off of 33.7,combined with HPV16/18,represents a promising triage of HPV-positive self-samples.Moreover,both screening and triage tests can be done via HPV testing at a single visit.3.The utility of p16 immunostaining could optimize the triage of HPV-positive self-samples via adjusting the staining score or combining with LBC or HPV genotyping,which provides a new insight into the precise screening of cervical cancer.Extended genotyping HPV33 could be utilized to optimize the triage of HPCN women,moreover,p16 immunostaining,either alone or combined with extended genotypes,was more effective than HPV genotyping alone in the triage of HPCN women.