Abstract: Objective The aims of this study were to compare the different expression between high-risk human papillomavirus load(HR-HPV DNA)and Th1/Th2 type cytokines in local different microenvironments of cervix, and to explore the possibility and significance of predicting cervical cancer. Methods A total of 339 patients with persistent HR-HPV infection were divided into cervical intraepithelial neoplasia(CIN)and cervical cancer.Forty patients with HPV-negative and cytological examination of normal cervix were used as controls.PCR fluorescence assay and double antibody sandwich were used.ELISA assay were used to detect HPV-DNA and Th1 type cytokines including IL-2, IFN-γ, TNF-ɑ and Th2 type cytokines including IL-4, IL-6 and IL-10 from cervical secretion.The ratio of TNF-ɑ/IL-10 was used as an index to measure the immune balance of Th1/Th2.Univariate and multivariate logistic regression were performed to analyze the data and then to screen the predicting indicators of cervical cancer. Results Univariate analysis showed that HR-HPV DNA, IL-2, IL-4, IL-6, IL-10, IFN-γ, TNF-ɑ and TNF-ɑ/IL-10 were significantly different between CIN and cervical cancer groups(P＜0.05), which could be used as a risk factor for predicting cervical cancer.Multivariate logistic regression analysis showed that IL-10, IFN-γ, TNF-ɑ, TNF-ɑ/IL-10 were the influencing factors of cervical cancer.The regression model fitted goodness test was Nagelkerke(R²= 0.982), the pre-judgment rate for CIN was 99.5%, the pre-judgment rate of cervical cancer was 100%, and the total positive rate was 99.7%, suggesting good fitting effect was good and the prediction accuracy was high. Conclusion CINⅠ and CINⅡ have abnormal expression of Th cytokine, but they do not affect Th1/Th2 balance.Th1/Th2 imbalance in CINⅢ stage and Th2 dominant expression promote the occurrence of cervical cancer.Based on the regression model for predicting cervical carcinogenesis, cervical immunity caused inhibitory microenvironment by local immune imbalance is the key link in HR-HPV persistent infection and cervical cancer, and has nothing to do with HR-HPV DNA.

Key words: Local immune microenvironment of cervix, Cervical carcinogenesis, Th1/Th2 type cytokines, High risk human papillomavirus load, Prediction model