Establishment and evaluation for nomogram model of rheumatoid arthritis with skeletal system involvement
10.13602/j.cnki.jcls.2024.11.03
- VernacularTitle: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
- Author:
Ziming WANG
1
;
Guorui LIU
;
Huicong ZHOU
;
Xiaoling YAN
;
Wei YU
Author Information
1. 南京大学医学院附属泰康仙林鼓楼医院检验科,南京 210046
- Keywords:
rheumatoid arthritis;
skeletal system involvement;
LASSO-Logistic regression;
nomogram
- From:
Chinese Journal of Clinical Laboratory Science
2024;42(11):809-815
- CountryChina
- Language:Chinese
-
Abstract:
Objective To explore the risk factors of rheumatoid arthritis patients complicated with skeletal system involvement,and es-tablish a prediction model and evaluate its performance.Methods A total of 165 patients,who were initially diagnosed with rheuma-toid 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 evalu-ated with receiver operating curve(ROC)and the calibration curve,and the clinical effectiveness of the model was evaluated by clini-cal 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,IgG4 level,IgG4/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,IgG4/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(χ2=7.562,P=0.477,＞0.05).Both the clin-ical decision curve and clinical impact curve showed its clinical practicability.Conclusion The established simple and practical nomo-gram model exhibited good ability to identify rheumatoid arthritis patients complicated with skeletal system involvement,which may pro-vide reference basis for clinical diagnosis and treatment.
