Objective: To establish a classification model and serum proteomic patterns in non-small cell lung cancer (NSCLC) patients with lymph node metastasis by surface enhanced laser desorption/ionization time of flight mass spectrometry (SELDI-TOF-MS).
Methods: The relative contents of serum proteins of 84 NSCLC patients with different N stages (35 N0 cases, 19 N1 and 30 N2 respectively) were detected by CM10 chip and SELDI-TOF-MS; two decision trees were generated to distinguish lymph nodes metastasis (N0 versus N1 + N2) and mediastinal lymph nodes metastasis (N0 + N1 versus N2) respectively.
Results: The model in which 50 patients were randomly chosen differentiated patients with lymph nodes metastasis from N0 patients with a sensitivity of 96.3%(26/27) and a specificity of 95.7%(22/23) in the training set, a following blind test was taken. Subsequently, compared with 49 patients with lymph node metastasis (N1 + N2), 15 patients with total negative lymph nodes (including lobar, segmental and subsegmental nodes necessarily) were defined as "true" N0 and were chosen to form a better predictive model with a 77.6% (38/49) sensitivity and a 93.3% (14/15) specificity respectively. And 6682.0Da, together with other five proteins, had significant difference between two groups; the result of this model for distinguishing the mediastinal lymph nodes metastasis is more accurate than thoracic CT analyses by Alongi F and many other clinical centers. It had a sensitivity of 80.0% (24/30) and a specificity of 77.8% (42/54) respectively.
Conclusion: SELDI-TOF-MS showed a potential value for predicting lymph nodes metastasis in NSCLC patients. And further studies are required to confirm the models and identify the related proteins.