Efficiency of Disease and Disease Activity Diagnosis Models of Systemic Lupus Erythematosus Based on Protein Array Analysis.

Abstract

Background Systemic lupus erythematosus (SLE) has become increasingly common in the clinic and requires complicated evidence of both clinical manifestations and laboratory examinations. Additionally, the assessment and monitoring of lupus disease activity are challenging. We hope to find efficient biomarkers and establish diagnostic models of SLE. Materials and Methods We detected and quantified 40 proteins using a quantitative protein array of 76 SLE patients and 21 healthy controls, and differentially expressed proteins were screened out by volcano plot. Logistic regression analysis was used to recognize biomarkers that could be enrolled in the disease diagnosis model and disease activity diagnosis model, and a receiver operating characteristic (ROC) curve was drawn to evaluate the efficiency of the model. A nomogram was depicted for convenient and visualized application of our models in the clinic. Decision curves and clinical impact curves were also plotted to validate our models. Results The protein levels of TNF RII, BLC, TNF RI, MIP-1b, eotaxin, MIG, MCSF, IL-8, MCP-1, and IL-10 showed significant differences between patients with SLE and healthy controls. TNF RII and MIP-1b were included in the SLE diagnosis model with logistic regression analysis, and the value of the area under the ROC curve (AUC) was 0.914 (95% confidence interval (CI), 0.859-0.969). TNF RII, BLC, and MIP-1b were enrolled in the disease activity diagnosis model, and the AUC value was 0.823 (95% CI 0.729-0.916). Both of the models that we established showed high efficiency. Additionally, the three protein biomarkers contained in the disease activity distinguish model provided additional benefit to conventional biomarkers in predicting active lupus. Conclusions The disease diagnosis model and disease activity diagnosis model that we developed based on protein array chip results showed high efficiency in differentiating patients with SLE from healthy controls and recognizing SLE patients with high disease activity, and they have also been validated. This implied that they might greatly benefit clinical decisions and the treatment of SLE.