Background: The diagnosis of myasthenia gravis (MG) remains challenging. We performed a proteome-wide search for potential serum protein diagnostic markers for MG using surface-enhanced laser desorption/ionization (SELDI) time-of-flight mass spectrometry (TOFMS).
Methods: Proteomic spectra from 80 MG patients and 80 healthy individuals were generated by SELDI. Samples from 56 MG patients and 56 healthy individuals in the training set were analyzed to set up the decision tree. Samples from 24 MG patients and 24 healthy individuals were used for cross-validation testing.
Results: The SELDI TOFMS analysis generated 101 peaks, representing differentially expressed proteins between 1000 and 20000 Da. Among them, 9 peaks were down-regulated and 30 others were up-regulated in the MG sera compared with the controls. The decision tree used the peak at M4168.94 Da and M1122.57 Da as splitters in the classification process. In the training set, 112 samples were classified as MG or control group, with a sensitivity of 100% and specificity of 89.3%; the 10-fold cross-validated analysis identified the optimal decision tree with the lowest relative cross-validated cost of 0.080. In the test set, the decision tree generated was able to identify 20 of 24 MG patients and 21 of 24 healthy individuals with a sensitivity of 83.3% and a specificity of 87.5%.
Conclusions: SELDI TOFMS is a useful tool for the detection and identification of potential serum biomarkers that can diagnose MG with high sensitivity and specificity.