Pulmonary tuberculosis (TB) is an infectious disease disturbing status of public health, and accurate diagnosis of TB would effectively help control the disturbance. Our study tried to establish a classification tree model that distinguished active TB from non-TB individuals. We used matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF MS) combined with weak cationic exchange (WCX) magnetic beads to analyse 178 serum samples containing 75 patients with active TB and 103 non-TB individuals (43 patients with common pulmonary diseases and 60 healthy controls). Samples were randomly divided into a training set and a test set. Statistical softwares were applied to construct this model. An amount of 48 differential expressed peaks (P < 0.05) were identified by the training set, and our model was set up by three of them, m/z 7626, 8561 and 8608. This model can discriminate patients with active TB from patients with non-TB with a sensitivity of 98.3% and a specificity of 84.4%. The test set was used to verify the performance, which demonstrated good sensitivity and specificity: 85.7% and 83.3%, respectively. Differential expressed peaks between smear-positive and smear-negative active TB also have been analysed. It came out that m/z 8561 and 8608 not only acted as vital factors in the pathogenesis of active TB but also played an important role in regulating different active TB status. In conclusion, MALDI-TOF MS combined with WCX magnetic beads was a powerful technology for constructing classification tree model, and the model we built could serve as a potential diagnostic tool for active TB.
© 2011 The Authors. Scandinavian Journal of Immunology © 2011 Blackwell Publishing Ltd.