Document Type : Original Article
Authors
1
Department of Chemistry, College of Science, University of Anbar , Iraq
2
Department of Chemistry, College of Science, University Of Anbar, Iraq.
3
General Directorate of Anbar Education Ministry of Education, Iraq.
4
Department of Chemistry, College of Science, University of Kufa, Iraq.
5
Department of Chemistry, College of Education for Pure Science, University Of Anbar, Iraq
Abstract
Rheumatoid arthritis (RA) is characterized by chronic inflammation of the synovial membrane that leads to the destruction of the joints. Measurements of cytokines have been carried out in many previous works with no decisive result. In the present study, three bone biomarkers (osteopontin, vascular-endothelial growth factor-A (VEGF), and Stromelysin-1 (MMP3)), and three inflammatory biomarkers (colony-stimulating factor (GM-CSF), interferon-γ, and tumor necrosis factor-alpha (TNFα)) are assayed and examined in RA by using artificial neural-network analysis and regression analysis. The study enrolled 112 patients with RA and 58 healthy controls. The biomarkers were measured by the enzyme-linked immunosorbent assay (ELISA) technique. The neural-network analysis showed that the top 3 sensitive predictors for RA are MMP3, TNFα, and osteopontin, followed by VEGF, GM-CSF, and interferon-γ. A significant part of the variance in the disease activity scale (DAS28), rheumatoid factor (RF), C-reactive protein (CRP), and anti-citrullinated protein antibodies (ACPA)) could be explained by interferon-γ, GM-CSF, osteopontin, and MMP3, respectively. The neural network and logistic regression findings showed that RA could predict MMP3, TNFα, and osteopontin with a good area under the curve of 0.938. In conclusion, the neural network analysis showed that MMP3, TNFα, and osteopontin are diagnostic biomarkers for RA disease and correlated with many disease-related characteristics.
Keywords
Main Subjects
)
View on SCiNiTO