宫颈高级别鳞状上皮内病变相关因素分析与预测  

Correlation factors analysis and prediction of high-grade squamous intraepithelial lesion

作　　者：李素玉[1] 郑祥钦[1] 陈瑜沙 戴丽玉[1] 林元[1] 衣欢 潘吴媛 Li Suyu;Zheng Xiangqin;Chen Yusha(Fujian Provincial Maternity and Children's Hospital,the Affiliated Hospital of Fujian Medical University,Fuzhou 350001)

机构地区：[1]福建省妇幼保健院福建医科大学附属医院,福州350001

出　　处：《现代妇产科进展》2021年第12期913-918,共6页Progress in Obstetrics and Gynecology

基　　金：国家自然科学基金青年项目(No:81802604);福建省中青年骨干人才培养项目(2019-ZQN-23);福建省妇幼保健院科研基金项目(妇保院研YCXQ18-17)。

摘　　要：目的研究影响患病宫颈高级别鳞状上皮内病变(HSIL)风险的相关因素并构建患病风险预测模型,为临床实践提供参考。方法选取2019年6月至2021年1月于福建省妇幼保健院诊治的慢性宫颈炎、低级别鳞状上皮内病变(LSIL)、HSIL及宫颈癌病例为研究对象,回顾性收集其阴道镜检查结果(醋酸白色上皮、镶嵌、点状血管等)、HPV检测结果(HPV16、31、33、58型是否阳性)及病理诊断结局等资料,分析罹患HSIL或宫颈癌的风险与阴道镜检查结果及HPV检测结果间的关联性,运用SAS9.4软件的logistic逐步回归分析方法筛选影响患病风险的相关因素并构建患病风险预测模型。结果共纳入研究对象5147例,其中慢性宫颈炎占59.74%(3075/5147),LSIL占23.33%(1201/5147),HSIL占15.00%(772/5147),宫颈癌占1.92%(99/5147)。研究对象患HSIL或宫颈癌的风险与其临床症状,醋酸白色上皮、镶嵌或点状血管,HPV16、31、33或58型阳性情况等8个纳入因素在单因素分析时均存在显著关联(P均<0.001)。研究因素间,仅镶嵌与点状血管间的列联系数值相对较大(0.36)。采用逐步回归法构建多变量logistic模型,则依次纳入点状血管、HPV16型阳性情况、醋酸白色上皮、HPV33型阳性情况、HPV58型阳性情况、镶嵌、HPV31型阳性情况,无剔除变量,该模型对训练样本预测的正确率为71.7%,模型的拟合优度良好。利用该模型可对不同HPV检测结果和(或)阴道镜检查结果患者的患病风险进行预测。结论阴道镜检查结果及HPV检测结果均与患HSIL或宫颈癌风险存在关联性;综合阴道镜检查结果及HPV检测结果构建的患病风险预测模型可为临床实践提供参考;HPV16、31、33、58型的多重感染与宫颈病变严重程度间关系密切。Objective:To explore the risk factors of high-grade squamous intraepithelial lesion(HSIL),and construct model for its prediction,and provide references for clinical applications.Methods:Selected the cases of chronic cervicitis,low-grade squamous intraepithelial lesion(LSIL),HSIL and cervical cancer diagnosed and treated by Maternity and Children's Health Care Hospital of Fujian Province from June 2019 to January 2021 as the research objects and retrospectively collected the results of colposcopy(acetowhite epithelium,mosaic,punctate blood vessels,etc.),HPV test(HPV16,31,33,58 positive)and pathological diagnosis of the cases.Then analyse the correlation between the risk of HSIL or cervical cancer and the results of colposcopy and HPV test,and use the logistic stepwise regression analysis of SAS 9.4 software to screen out the related factors influencing the risk and build the risk prediction model.Results:A total of 5147 patients were included in the study,including 59.74%(3075/5147)chronic cervicitis,23.33%(1201/5147)LSIL,15.00%(772/5147)HSIL and 1.92%(99/5147)cervical cancer.The risk of HSIL or cervical cancer was significantly associated with clinical symptoms,acetowhite epithelium,mosaic or punctate vessels,and HPV16,31,33 or 58 positive in univariate analysis(all P<0.001).Among the research factors,only the contingency coefficient value between mosaic and punctate vessels was relatively large(0.36).The multivariable logistic model was constructed by stepwise regression method,which included the presence or absence of punctate vessels,HPV16 positive,acetowhite epithelium,HPV33 positive,HPV58 positive,mosaic,HPV31 positive in turn,and no eliminated variables.The accuracy of the model for the prediction of training samples was 71.7%,and the goodness of fit of the model was good.The model can be used to predict the risk of patients with different HPV test results and/or colposcopy results.Conclusion:The results of colposcopy and HPV test were correlated with the risk of HSIL or cervical cancer.The risk prediction

关 键 词：宫颈癌 高级别鳞状上皮内病变(HSIL) LOGISTIC回归 预测 

分 类 号：R737.33[医药卫生—肿瘤]