ROC曲线评价红细胞沉降率在14种疾病中的诊断价值  

Effect of ROC curve in evaluation of ESR for diagnosis of 14 diseases

作　　者：朱喜丹[1] 田刚[1] 王南华[1] 

机构地区：[1]泸州医学院附属医院检验科,四川泸州646000

出　　处：《国际检验医学杂志》2013年第3期276-277,共2页International Journal of Laboratory Medicine

摘　　要：目的探讨红细胞沉降率(ESR)的ROC曲线分析在14种疾病中的诊断价值,为临床医生提供合理的数据解释和辅助诊断。方法借助泸州医学院附属医院信息系统(HIS)和实验室信息系统(LIS)平台,回顾性分析了2536份ESR测定结果。采用SPSS统计分析软件对14种常见疾病的ESR结果进行ROC和基于人工神经网络模型(ANN)的ROC曲线分析,以期建立一种较好的疾病辅助诊断和评估模式。结果 ESR指标增高对结核、肾脏疾病、多发性骨髓瘤、类风湿性关节炎、川崎病和糖尿病的诊断具有一定的准确性(ROC曲线下面积均大于0.6,P<0.01)。结论对不同疾病的ESR结果进行信息挖掘并结合患者临床表现进行综合分析,有助于疾病的正确诊断和风险评估。Objective To investigate the diagnostic value of erythrocyte sedimentation rate（ESR） in 14 diseases and to provide reasonable data interpretation as well as auxiliary diagnosis for clinical. Methods Based on the information platform of hospital in formation system（HIS） and experiment information system（IAS） in Luzhou Medical College Affiliated Hospital,2 536 cases of ESR data were reviewed. To acquire a superior auxiliary diagnosis and evaluate model for diseases,the ROC data analysis and artificial natural neural network（ANN） model were used to analyze ESR data in 14 diseases by SPSS statistical analysis software. Results Increased ESR detection results were useful for diagnosis of tubercle disease, kidney disease, and general lymphadenomalosis of bone, rheumarthritis,diabetes mellitus and Kawasaki disease （area under the curve of ROC were more than 0.6, P〈0.01, respec- tively）. Conclusion Combined information extraction of ESR and clinical situation of patients to aggregate analysis was useful for diagnosis of disease and risk evaluation.

关 键 词：红细胞沉降率 ROC曲线 疾病 

分 类 号：R446.1[医药卫生—诊断学]