Development and evaluation of a diagnostic model for autoimmune liver disease based on clinical indicators
10.11855/j.issn.0577-7402.1163.2025.0422
- VernacularTitle:基于临床指标构建自身免疫性肝病诊断模型的价值分析
- Author:
Jia-Cheng JIN
1
;
Jia-Chi YU
;
Rui-Bing LI
;
Mian-Yang LI
Author Information
1. 解放军医学院,北京 100853;解放军总医院第一医学中心检验科,北京 100853
- Keywords:
hepatitis,autoimmune liver disease;
alkaline phosphatase;
gamma-glutamyl transferase
- From:
Medical Journal of Chinese People's Liberation Army
2025;50(8):976-983
- CountryChina
- Language:Chinese
-
Abstract:
Objective To develop and evaluate a diagnostic model for autoimmune liver disease(AILD)using clinical indicators.Methods A total of 181 AILD patients(AILD group)diagnosed in the Department of Gastroenterology or Rheumatology and Immunology,the First Medical Center of Chinese PLA General Hospital,from June 2019 to February 2024,133 patients with hepatitis B virus cirrhosis(HBVC)(HBVC group),and 100 healthy individuals(HC group)were included in the study cohort for retrospective analysis.Clinical data,blood routine tests,and biochemical indicators were compared among the groups.Propensity score matching was used to balance the effects of factors such as gender and age.Multivariate logistic regression analysis was applied to identify characteristic indicators for AILD.Receiver operating characteristic(ROC)curves were used to evaluate the diagnostic efficacy of single and combined indicators,and the optimal diagnostic model for AILD was constructed.The Hosmer-Lemeshow test was performed to assess the model's goodness-of-fit,while ROC analysis was employed to evaluate the predictive validity.Additionally,an independent validation cohort,including 91 AILD patients(AILD validation group),82 HBVC patients(HBVC validation group),and 40 healthy individuals(HC validation group)who visited during the same period,was used to validate the diagnostic model.False-positive and false-negative rates were calculated to assess the diagnostic efficacy and clinical utility of the model.Results The age and the proportion of male patients in AILD and HBVC groups were higher than those in HC group,and the proportion of male patients in AILD group was higher than in HBVC group,the differences were statistically significant(P＜0.05).Prior to matching,compared with HC group,HBVC group and AILD group had significantly lower levels of hemoglobin(HB),red blood cell count(RBC),white blood cell count(WBC),platelet count(PLT),and albumin(ALB)(P＜0.05),but significantly higher levels of alanine aminotransferase(ALT),aspartate aminotransferase(AST),alkaline phosphatase(ALP),gamma-glutamyl transferase(GGT),total bilirubin(TBIL),and direct bilirubin(DBIL)(P＜0.05).Compared with HBVC group,AILD group had significantly higher levels of WBC,PLT,ALB,ALP,and GGT(P＜0.05 or P＜0.001)but significantly lower levels of TBIL and DBIL(P＜0.001).After maching,compared with HC group,AILD group still had significantly lower levels of HB,RBC,WBC,PLT,and ALB,but significantly higher levels of ALT,AST,ALP,GGT,TBIL,and DBIL(P＜0.05).Compared with HBVC group,AILD group had significantly higher levels of HB,RBC,WBC,PLT,ALB,ALT,ALP,and GGT,but significantly lower levels of DBIL(P＜0.05).ROC analysis results showed that AST,GGT,ALP,and DBIL had high predictive values for differentiating AILD from HC(AUC=0.843,0.804,0.802,0.773),while PLT,WBC,GGT,and ALP had high predictive values for differentiating AILD from HBVC(AUC=0.780,0.739,0.729,0.702).For distinguishing AILD from HC,the combined model of AST,ALP,and DBIL(Model e)had an AUC of 0.887,with a sensitivity of 82.2％,and a specificity of 94.7％.For distinguishing AILD from HBVC,the combined model of ALP,GGT,and PLT(Model D)had an AUC of 0.829,with a sensitivity of 78.5％and a specificity of 80.6％.Conclusions The combined diagnostic model based on AST,ALP,and DBIL can effectively differentiate healthy individuals without liver diseases from AILD patients,while the combined diagnostic model based on PLT,ALP,and GGT can effectively distinguish HBVC from AILD patients.
