A nomogram model predicting the risk of high-grade cervical squamous intraepithelial lesions: a population-based study

Abstract

Abstract Purpose: This study aimed to develop a nomogram for the prediction of high-grade cervical squamous intraepithelial lesions (HSILs). Methods: This cross-sectional study included the baseline data of the Shanxi Cervical Intraepithelial Neoplasia (CIN) cohort study, in which 1,249 community-dwelling adults (372 patients with pathologically diagnosed CIN and 877 patients with a normal cervix) from a rural area of Shanxi, China, were recruited. Interviewer-administered questionnaires on sociodemographic characteristics, hygiene habits, marriage and childbearing information, and past medical and family histories of cancer were obtained. Human papillomavirus (HPV) infection types were detected by flow-through hybridization. Logistic regression analysis of independent risk factors for HSILs and a nomogram prediction model were established. Results: Of the 1,249 participants, 460 (36.83%) were infected with high-risk HPV (HR-HPV), and the five most frequent types of HPV were HPV16, HPV52, HPV58, HPV56, and HPV33. After adjusting for demographics and other related variables, HPV16 was associated with an 11.363-fold (95% confidence interval [CI], 6.639–19.449) higher risk of HSILs and HPV58 was associated with a 5.758-fold (95% CI, 2.542–13.045) higher risk of HSILs compared to the uninfected group. HR-HPV infection, younger age at menarche, menopause, and tea drinking were selected as nomogram covariates. The concordance index of the nomogram prediction model was 0.822. Conclusion: The most common highly pathogenic HR-HPV types in the study area are HPV16 and HPV58. An easy-to-use nomogram, with reliable discrimination ability and accuracy, was established to help predict HSILs using the identified significant risk factors. Trial registration: This study was registered by the China Clinical Trials Center (registration number: ChiCTR-ROC-15006479).