宫颈鳞状细胞癌预测指标的模型建立与分析  

Establishment and analysis of the predictor model of cervical squamous cell carcinoma

作　　者：周宇涵[1] 樊伍峰[1] 董金菊[1] 习红丽[1] 郭广林[1] 黄玉琴[1] Zhou Yuhan;Fan Wufeng;Dong Jinju;Xi Hongli;Guo Guanglin;Huang Yuqin(Affiliated Xiangyang Hospital, Hubei University o f Medicine y Hubei Xiangyang 441000,China.)

机构地区：[1]湖北医药学院附属襄阳医院,湖北襄阳441000

出　　处：《现代肿瘤医学》2016年第14期2289-2293,共5页Journal of Modern Oncology

摘　　要：目的:通过检测HR-HPV感染、hTERC、c-myc基因扩增和MCM5蛋白在宫颈鳞状细胞癌及不同级别宫颈上皮内瘤变中的水平,筛选宫颈癌相关指标,建立回归模型用以预测宫颈鳞状细胞癌,并评估模型效果。方法:筛选初诊病理确诊宫颈鳞状细胞癌和CIN I、II、III级患者共200例作为研究对象,分别检测HR-HPV感染、hTERC、c-myc基因扩增及MCM5蛋白表达,用Logistic向后逐步回归的方法,筛选宫颈癌相关指标,建立回归模型预测宫颈鳞状细胞癌,并评估模型效果。结果:将HR-HPV负荷量及感染状况、hTERC、c-myc基因扩增和MCM5蛋白表达的检测数据绘制直方图,并进行Logistic向后逐步回归分析,得出hTERC、HR-HPV负荷量、MCM5回归系数分别为0.042、0.061和0.052,P值分别是0.024、0.005、0.005(P<0.05),HR-HPV感染状态和c-myc基因P值是0.856和0.682(P>0.05),被回归方程排除,提示hTERC(X_1)、HR-HPV负荷量(X_2)、MCM5(X_5)与回归方程存在线性关系,即与宫颈鳞状细胞癌发生有关,由此建立回归模型Logit(P)=-66.283+0.042X_1+0.061X_2+0.052X_5。评估模型拟合优度和预测准确度,H-L检验P值=1(P>0.05),模型拟合效果好,Cox-Snell R^2=0.643,Nagelkerke R^2=0.958,模型预测准确度98.5%,模型预测准确性高。结论:由hTERC、HR-HPV负荷量和MCM5蛋白建立的回归模型拟合效果较好,对宫颈鳞状细胞癌的预测准确度较高,hTERC、HR-HPV负荷量和MCM5蛋白联合检测能够用于宫颈鳞状细胞癌的预测评估,对CIN患者的分流管理和预后评估、宫颈鳞癌患者的早期诊断均有较高临床价值。Objective:Detecting the HR - HPV infection,hTERC gene amplification,c - myc gene amplification,MCM5 protein in cervical squamous cell carcinoma and different levels of cervical intraepithelial neoplasia, screening indexes of cervical cancer, establish the regression model to predict cervical squamous cell carcinoma, and evaluate prediction effect of the model. Methods : Screening 200 patients as the research object with cervical squamous cell car-cinoma and CIN I, II, III, respectively detecting HR - HPV infection, hTERC gene amplification, c - myc gene amplifi-cation and MCM5 protein expression. Screening indicators related of cervical cancer,establishing regression model to predict cervical squamous cell carcinoma with logistic backward stepwise regression, and evaluating prediction effect of the model. Results : Drawing histogram of detection data in HR - HPV viral load and infection status , hTERC gene amplification, c - myc gene amplification and MCM5 protein expression, calculating hTERC, HR - HPV viral load,MCM5 regression coefficients were 0. 042,0. 061 and 0. 052 with Logistic backward stepwise regression, P values were 0. 024,0. 005 ,0. 005 (P < 0. 05 ) . HR - HPV infection status and c - myc gene P value were 0. 856 and 0. 682 (P >0. 05) which were ruled out by the regression equation. There was a linear relationship between hTERC(X2 ),HR - HPV viral load ( X2 ) , MCM5 ( X5) with regression equation, that was hTERC ( ) , HR - HPV viral load ( X2 ),MCM5(X5) were associated with cervical squamous cell carcinoma. Thus establish regression model Logit(P)=- 66. 283 + 0. 042X! + 0. 061X2 + 0. 052X5. Evaluating model fitting effect and forecast accuracy,Hosmer - Lemeshow P = 1(P>0. 05). Model fitting effect was good, decision coefficient of the model Cox - Snell R2 = 0. 643 , Nagelkerke R2 = 0. 958,prediction accuracy by 98. 5% . Model prediction accuracy was high. Conclusion: Regression model was established by hTERC gene,HR - HPV viral load and MCM5 protein,fitting effect was good. The prediction accuracy o

关 键 词：宫颈鳞状细胞癌 宫颈上皮内瘤变 预测模型 高危型人乳头瘤病毒 端粒酶基因 C-MYC基因 MCM5蛋白 

分 类 号：R737.33[医药卫生—肿瘤]