Premium
Establishment of a decision tree model for diagnosis of early rheumatoid arthritis by proteomic fingerprinting
Author(s) -
Li Yuhui,
Sun Xiaolin,
Zhang Xuewu,
Liu Yanying,
Yang Yuqin,
Li Ru,
Liu Xu,
Jia Rulin,
Li Zhanguo
Publication year - 2015
Publication title -
international journal of rheumatic diseases
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.795
H-Index - 41
eISSN - 1756-185X
pISSN - 1756-1841
DOI - 10.1111/1756-185x.12595
Subject(s) - medicine , rheumatoid arthritis , training set , diagnostic model , mass spectrometry , gastroenterology , chromatography , artificial intelligence , data mining , computer science , chemistry
Aim The objective of this study was to identify proteomic biomarkers specific for rheumatoid arthritis ( RA ) by matrix‐assisted laser desorption/ionization time‐of‐flight mass spectrometry ( MALDI ‐ TOF ‐ MS ) in combination with weak cationic exchange ( WCX ) magnetic beads. Methods Serum samples from 50 patients with RA and 110 disease controls (50 SLE and 60 SS ) and 51 healthy individuals were analyzed. The samples were randomly divided into a training set or test set to develop a diagnostic model for RA . Results A total of 83 protein peaks were identified to be related with RA , in which four of the peaks with mass‐charge ratio ( m/z ) at 8133.85, 5844.60, 13 541.3 and 14 029.0 were selected to establish a model for diagnosis of RA . This classification model could separate patients with RA from diseased and healthy controls with sensitivity of 84.0% and specificity of 92.5%, and its accuracy was confirmed in the blind testing set with high sensitivity and specificity of 80.0% and 93.3%, respectively. Conclusions This study suggested that potential serum biomarkers for RA diagnosis could be discovered by MALDI ‐ TOF ‐ MS . The classification tree model set up in this study might be used as a novel diagnostic tool for RA .