宫颈上皮内瘤变发生风险预测模型的建立与评价

Establishment and evaluation of a risk prediction model for cervical intraepithelial neoplasia

目的:探讨人类乳头瘤病毒(HPV)16、HPV18、HPV52、HPV58型感染对阴道微生物状态及局部免疫状态的影响及建立与评价相关的宫颈上皮内瘤变(CIN)发生风险预测模型。方法:选取2018年6月至2020年6月在广西医科大学附属肿瘤医院妇瘤科行高危型人乳头状瘤病毒(HR-HPV)检测的女性患者226例为研究对象,根据HPV感染结果分为HPV16(+)组(n=88)、HPV18(+)组(n=14)、HPV52(+)组(n=33)、HPV58(+)组(n=20)及HPV(-)组(n=71)。同时根据宫颈活检病理分为慢性炎症组、CIN组。检测各组阴道微生物情况及阴道灌洗液中白细胞介素(IL)-2、IL-10、CD4^(+)、CD8^(+)水平,分析阴道微环境中导致CIN发生的危险因素并建立和评价CIN发生风险预测模型。结果:HPV(+)、IL-2、IL-10、CD4^(+)对CIN发生风险的影响具有统计学意义(P<0.05),纳入CIN发生风险预测因子形成列线图。模型受试者工作特征曲线(ROC)曲线下面积(AUC)为0.920;校准曲线显示预测概率与实际概率有很好的一致性;决策曲线分析则显示模型具有较好临床适用性。结论:基于HPV(+)、IL-2、CD4^(+)、IL-104个预测因素建立的CIN发生风险预测模型具有良好的区分度、准确性和临床适用性,有利于CIN早诊早治。

Objective:To investigate the effects of human papillomavirus(HPV)16/18/52/58 infection on vaginal microecology and local immune status and to establish and evaluate a prediction model for the occurrence of cervical intraepithelial neoplasia(CIN).Methods:A total of 226 female patients who underwent high-risk human papillomavirus(HR-HPV)detection from June 2018 to June 2020 in the Department of Gynecologic Oncology,Guangxi Medical University Cancer Hospital,were selected as the research subjects.According to the results of HPV infection,the 226 females were divided into HPV16(+)group(n=88),HPV18(+)group(n=14),HPV52(+)group(n=33),HPV58(+)group(n=20),and HPV(-)group(n=71).At the same time,according to the pathology of cervical biopsy,the patients were divided into a chronic inflammatory group and a CIN group.For each group,the vaginal microecology and the levels of IL-2,IL-10,CD4^(+),and CD8^(+)in vaginal lavage fluid were detected.Using statistically significant factors in univariate analysis,further multivariate analyses were performed to establish and evaluate the risk prediction model.Results:The effects of HPV(+),IL-2,IL-10,and CD4^(+)on the occurrence of CIN were statistically significant(P<0.05),and these factors were used to construct a nomogram.Based on the established model,the area under the receiver operating characteristic curve was 0.920.The calibration curve showed that the predicted probability is similar to the actual probability.A decision curve analysis indicated the conspicuous clinical applicability of the model.Conclusion:In this study,the CIN risk prediction model established based on four factors[HPV(+),IL-2,CD4^(+),and IL-10]had good differentiation,accuracy,and clinical applicability,which is conducive to the early diagnosis and treatment of CIN.