自身抗体谱数据挖掘及SLE预测模型的建立

目的 探讨自身抗体在系统性红斑狼疮(systemic lupus erythematosus,SLE)及其他自身免疫性疾病(autoimmune diseases,AID)中的阳性分布趋势和诊断价值.方法 回顾分析2015年1月至2016年9月绵阳市中心医院自身抗体谱(autoantibodies,AAB)检测病例10684例,其中SLE 801例,其他AID1564例,非AID8319例,观察16种自身抗体在不同性别、年龄和疾病中的阳性分布差异,分析自身抗体个案聚类结果与临床诊断的相关性及其对SLE的预测价值.结果 16种自身抗体在女性中的阳性率,除外Jo-1、Scl-70、PM-Scl和PCNA,均高于男性,差异有统计学意义(P<0.001);在≤20岁、21~49岁和≥50岁三个年龄组的阳性率,除外Jo-1、PM-Scl、PCNA和AMA-M2,差异均有统计学意义(P<0.001);在SLE的阳性率,除外Jo-1、CENP B、Scl-70和AMA-M2,均高于其他AID和非AID两个疾病对照组,差异有统计学意义(P<0.001).自身抗体个案聚类与临床诊断相关(r=0.603,Kappa=0.476).ANA(OR=2.96)和ds-DNA(OR=2.31)对SLE预测价值排前2位.Logistic模型预测SLE的Se 81.2%,Sp97.6%,LP73.5%,LN98.4%,Ac 96.4%.结论 自身抗体在不同性别、年龄和疾病中的分布存在差异,自身抗体谱表达模式具有聚类特征,多种自身抗体联合性别、年龄因素建立Logistic预测模型有助于SLE诊断.

Data Mining of Autoantibodies Test Results and Establishmentof SLE Prediction Model

Objective To investigate the positive distribution trends and diagnostic value of autoantibodies in SLE and other AID.Methods A retrospective analysis of autoantibodies test results was performed in 10684 cases from our hospital since January 2015 to September 2016, including 801 cases of SLE patients, 1564 cases of other AID patients and 8319 cases of non AID patients.The study goal was to observe the positive distribution of 16 autoantibodies in different gender, age and disease, also to analyzed the relevance between cluster grouping packet and clinical diagnosis.Then we analyzed the predictive value of autoantibodies for SLE.Results The positive rates of 16 autoantibodies were significantly higher in women than that of male except for Jo-1,Scl-70, PM-Scl and PCNA (P<0.001).Therewere significant differences among the three age groups except for Jo-1,PM-Scl, PCNA and AMA-M2 (P<0.001).They were significantly higher in SLE than that of other AID and non-AID except for Jo-1,CENP B, Scl-70 and AMA-M2 (P<0.001).The sample clustering was related to clinical diagnosis(r=0.603,Kappa=0.476).