基于智能算法子宫内膜异位症辅助诊断模型建立.docVIP

基于智能算法子宫内膜异位症辅助诊断模型建立 　　摘要：目的 基于智能算法建立子宫内膜异位症（endometriosis， EM） 的辅助诊断模型，并评价其检验效能。方法 以确诊的220例子宫内膜异位症患者为观察组，以同期检查的健康女性100例为对照组，对她们的血清取样分析CA125、MIF、AQP1、VEGF、EmAb/TNF-α、IL-6 、IL-8 、PD14、HE4、IL-4、TGF-β、RANTES、MMP-9在血清中含量。利用单因素Logistic回归分析筛选出与子宫内膜异位症的发生有统计学意义的12种血清标记物，以这12种血清标记物水平为自变量，以是否患病为因变量，应用支持向量机算法建立EM的12对1的数学辅助诊断模型，并通过60例测试集评价其效果。结果 成功建立了基于支持向量机的EM辅助诊断模型，支持向量机诊断模型的诊断准确度、敏感性、特异性分别为91.67%、93.33%、90%。结论 支持向量机诊断模型的诊断准确率、敏感性及特异性较高，对EM的诊断有重要的参考价值，可做为临床医生的辅助诊断工具。 　　关键词：子宫内膜异位症；血清标记物；Logistic回归；支持向量机 　　Establishing the Auxiliary Diagnosis Model of the Endometriosis Based on Intelligent Algorithms 　　YUAN Jia-ying1，ZHANG Jian-hua2，LIU Hua3 　　（1.Department of Ultrasonic Diagnosis，Hospital of Henan Province Military Area Command，Zhengzhou 450014，Henan，China； 　　2.School of Electrical Engineering，Zhengzhou University，Zhengzhou 450001，Henan，China； 　　3.Department of Equipment，Xinyang Central Hospital，Xinyang 464000，Henan，China） 　　Abstract：Objective Establishing the auxiliary diagnosis model of the endometriosis（EM） based on intelligent algorithms. Methods 220 patients that have been diagnosed as EM patients were set as observation group， 100 healthy women who accepted the diagnosis at the same time were treated as control group. The content of CA125，MIF，AQP1，VEGF，EmAb/TNF-α，IL-6 ，IL-8，PD14，HE4，IL-4，TGF-β，RANTES，MMP-9 in their serum were analyzed. And then screening out 12 kinds of serum markers that of statistical significance with the occurrence of EM by Logistic single factor regression analysis and treat these 12 kinds of serum markers level as independent variable， whether attacked by EM was assumed as dependent variable. Auxiliary EM diagnostic model of 12 to 1 was built according to the SVM（support vector machine） method， and the effect was evaluated form 60 test set at last. Results A EM auxiliary diagnostic model based on the SVM was built， the accuracy rate， sensibility and specificity were 91.67%，93.33%，90% respectively. Conclusion The accuracy rate， sensibili