Prospective Study On The Relationship Between Human Gene Methylation And HPV Infection And The Outcome Of Low-grade Cervical Lesions

ObjectiveThe accuracy verification of BD Onclarity HPV and Roche cobas nucleic acid was conducted to explore the performance of BD Onclarity HPV test in cervical cancer screening in Chinese population.In addition,the relationship between human gene methylation and the pathological grade of cervical cancer and precancerous lesions was investigated,and a model was established to evaluate the effect of the selected methylated gene fragments for cervical cancer screening.In addition,the selected gene fragments will be used to explore the relationship between persistent or transient infection of high-risk human papillomavirus(hr-HPV)and human gene methylation,as well as the relationship between outcome and human gene methylation in people with pathological grade 1 cervical intraepithelial neoplasia(CIN1).To explore a new way to optimize cervical cancer screening and triage strategy in Chinese women.Materials and MethodIn the BD Onclarity and Roche cobas nucleic acid validation study,we selected 944 cervical samples from three units for assays,of which 588 were sequenced.In the study of human gene methylation,we used three groups:(1)80 women enrolled for cervical cancer screening in Changzhi People’s Hospital of Shanxi Province in 2015;(2)from April 2014 to March 2015,a total of 101 women with HPV+cervical cancer screening results were selected from normal women attending cervical cancer screening and patients attending hospital gynaecological clinics at five research centers.(3)A prospective study conducted in Yanting Maternal and Child Health Hospital,Mianyang City,Sichuan Province from 2019 to 2022.Women enrolled in the natural population were initially screened and followed up for a total of four years using HPV DNA testing and cytological testing.Step 1,80 samples were used to screen human methylated gene fragments and establish the model,and 101 samples were used to verify the model and obtain the model and gene fragments with good performance.Step 2,an exploratory study using the genetic fragments described above,using baseline 2019 cytological samples for methylation sequencing,using HPV DNA testing,cytological testing,and pathological testing results from three consecutive years of follow-up,with a view to establishing that baseline cytological samples for methylation testing.According to the methylation situation,the subsequent lesion trajectory can be predicted(that is,the prediction of transient or persistent HPV infection and the prediction of the outcome of CIN1),and early intervention can be carried out in high-risk groups.Results1.In nucleic acid test accuracy verification,results showed very high overall agreement of HPV DNA tests between Onclarity and cobas(κ=0.7755).No difference in≥CIN2 sensitivity was observed between Onclarity and cobas(both 96.5%),while the specificity of Onclarity for ≥CIN2 was higher(16.6%,95%CI:13.7-19.9)than that of cobas(11.5%,95%CI:9.1-14.5).2.Based on cross-sectional population data,a screening triage model for HPV+population was established.47 gene segments were finally selected for the construction of HPV+population triage model,among which the random forest model performed best.The sensitivity and specificity of random forest model were 88.89%and 84.00%,respectively.The sensitivity and specificity of random forest model were 65.33%,92.31%,57.50%,94.36%,and 88.60%respectively.The sensitivity,specificity,positive predictive value,negative predictive value,and accuracy of the cytological test were 93.33%,30.77%,17.68%,96.66%,and 39.37%.In the samples studied,the triage effect of the methylated fragment we screened was similar to that of E6 protein,compared with the results of commonly used for HPV+population triage.3.On the basis of previous studies,an exploratory study was conducted to predict the status of human gene methylation in HPV+population for transient or persistent infection.Results showed that no model performed well in predicting transient or persistent infection in HPV+ populations.Only when predicting transient or persistent infection in HPV16/18 positive population did the random forest model perform well.4.On the basis of previous studies,exploratory research was conducted to predict the maintenance and progress of the improvement of CIN1 population by human gene methylation status.In predicting the reversal maintenance and progression of CIN1 population,the naive Bayes model showed better performance,with the sensitivity of 75.00%,specificity of 88.46%and AUC of 0.817.Conclusion1.Onclarity has good clinical performance in Chinese women.For HPV DNA testing,Onclarity was comparable in sensitivity to cobas,but slightly more specific.Cobas tested for HPV16 and 18 alone,but the other 12 HPV types were tested in the same channel,while Onclarity had nine channels to stratify the other 12 high-risk HPV types.The stratification provided by Onclarity may separate HPV types that are at low risk of progression from those that are at high risk,facilitating the exploration of each HPV type.Onclarity has demonstrated good clinical performance and is an effective tool for cervical cancer screening in China.2.Human methylation as a triage method for HPV+population has certain advantages,and the performance of E6 protein and p16/Ki67 in our 52 modeling samples and 101 validation samples is similar,indicating that the series of methods in this study are expected to be applied to the triage of HPV+population.3.Human gene methylation is expected to be used in predicting the outcome of CIN1 in the future.4.Based on the methylated gene fragments screened earlier,and constructed by random forest model,support vector machine model and naive Bayes model,8 gene fragments with good performance were screened out,including:SOX1-1_group1,GATA4-13_group2,LMX1 A-1_group 1,ADC YAP 1-18_group 1,AJAP1-15_group 1,LHX8-27_group1,MAL-8_group1 and CDH13-9_group.It provides support for the design of subsequent test kits.