类风湿关节炎合并骨骼系统受累列线图模型的构建和评估  

Establishment and evaluation for nomogram model of rheumatoid arthritis with skeletal system involvement

作　　者：王子铭 刘国瑞 周慧聪 严孝岭[1] 虞伟[1] WANG Ziming;LIU Guorui;ZHOU Huicong;YAN Xiaoling;YU Wei(Department of Medical Laboratory,Taikang Xianlin Drum Tower Hospital,Clinical School of Medical College,Nanjing University,Nanjing 210046,Jiangsu;Department of Clinical Laboratory,Jinling Hospital Affiliated to Nanjing University Medical College/General Hospital of Eastern Theater Command,PLA,Nanjing 210002,Jiangsu,China)

机构地区：[1]南京大学医学院附属泰康仙林鼓楼医院检验科,南京210046 [2]南京大学医学院附属金陵医院/东部战区总医院中心实验科,南京210002

出　　处：《临床检验杂志》2024年第11期809-815,共7页Chinese Journal of Clinical Laboratory Science

基　　金：南京市卫生科技发展专项资金(YKK20225);江苏省科技支撑计划-社会发展项目(BE2011786);国家自然科学基金面上项目(81071419)。

摘　　要：目的探讨类风湿关节炎患者合并骨骼系统受累的风险预测因子,并构建和评估预测模型。方法入选2015年9月至2021年3月在泰康仙林鼓楼医院和东部战区总医院病程首次明确诊断为类风湿关节炎的住院患者,共165例纳入研究分析。将合并与无合并骨骼系统受累患者的临床及血清学指标进行比较,基于LASSO-Logistic筛选的预测因子建立列线图模型,用受试者工作特征曲线(ROC曲线)及校正曲线评估模型的区分度和校准度,通过临床决策曲线和临床影响曲线评估模型的临床效用性。结果类风湿关节炎患者中合并有骨骼系统受累99例(60%)。与无合并组患者比较,合并组患者年龄、病程、免疫球蛋白G4(IgG 4)、IgG 4/免疫球蛋白G(IgG)、类风湿因子(RF)、抗链球菌溶血素O(ASO)、血小板压积(PCT)等实验室测量指标差异有统计学意义(P均<0.05);入选年龄、病程、IgG 4/IgG、RF、PCT 5项指标构建类风湿关节炎合并骨骼系统受累的列线图模型,模型C统计量为0.848(95%CI:0.785~0.910),校准曲线显示内部验证预测性能佳,霍斯默-莱梅肖检验拟合优度,χ^(2)=7.562,P=0.477>0.05,且临床决策曲线和临床影响曲线显示具有临床实用性。结论建立简单实用的列线图模型具有较好的识别类风湿关节炎合并骨骼系统受累的能力,可为临床诊疗提供参考依据。Objective To explore the risk factors of rheumatoid arthritis patients complicated with skeletal system involvement,and establish a prediction model and evaluate its performance.Methods A total of 165 patients,who were initially diagnosed with rheumatoid arthritis in Taikang Xianlin Drum Tower Hospital and Eastern Theater General Hospital from September 2015 to March 2021,were included in this study.The clinical and serological parameters of the patients with and without complicated skeletal system involvement were compared.The predictive factors with important significance in LASSO-Logistic analysis were used as the candidate variables,and a nomogram model was established based on the screened predictive factors.The discrimination and calibration of the model were evaluated with receiver operating curve(ROC)and the calibration curve,and the clinical effectiveness of the model was evaluated by clinical decision curve and clinical impact curve.Results The incidence of rheumatoid arthritis patients complicated with skeletal system involvement was 60%(99/165).Compared with the group of non-complicated with skeletal system involvement,statistically significant differences were found in age,course of disease,IgG 4 level,IgG 4/IgG ratio,rheumatoid arthritis(RF),antistreptolysin O(ASO),plateletcrit(PCT)and other parameters of laboratory tests in the complication group(P<0.05).The nomogram model of rheumatoid arthritis complicated with skeletal system involvement was established by five indicators,i.e.,age,course of disease,IgG 4/IgG ratio,RF and PCT.The statistic of the model C was 0.848(95%CI:0.785 to 0.910).Calibration curves showed good predictive performance in internal validation,and the results of Hosmer-lemeshaw test indicated goodness of fit(χ^(2)=7.562,P=0.477,>0.05).Both the clinical decision curve and clinical impact curve showed its clinical practicability.Conclusion The established simple and practical nomogram model exhibited good ability to identify rheumatoid arthritis patients complicated with skeletal

关 键 词：类风湿关节炎 骨骼肌肉系统受累 LASSO-Logistic 列线图 

分 类 号：R446.6[医药卫生—诊断学]