Proteins associated with EGFR-TKIs resistance in patients with non-small cell lung cancer revealed by mass spectrometry

The present study aimed to identify potential serum biomarkers for predicting the clinical outcomes of patients with advanced non-small cell lung cancer (NSCLC) treated with epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs). A total of 61 samples were collected and analyzed using the integrated approach of magnetic bead-based weak cation exchange chromatography and matrix-assisted laser desorption/ionization-time of flight-mass spectrometry. The Zhejiang University Protein Chip Data Analysis system was used to identify the protein spectra of patients that are resistant and sensitive to EGFR-TKIs. Furthermore, a support vector machine was used to construct a predictive model with high accuracy. The model was trained using 46 samples and tested with the remaining 15 samples. In addition, the ExPASy Bioinformatics Resource Portal was used to search potential candidate proteins for peaks in the predictive model. Seven mass/charge (m/z) peaks at 3,264, 9,156, 9,172, 3,964, 9,451, 4,295 and 3,983 Da, were identified as significantly different peaks between the EGFR-TKIs sensitive and resistant groups. A predictive model was generated with three protein peaks at 3,264, 9,451 and 4,295 Da (m/z). This three-peak model was capable of distinguishing EGFR-TKIs resistant patients from sensitive patients with a specificity of 80% and a sensitivity of 80.77%. Furthermore, in a blind test, this model exhibited a high specificity (80%) and a high sensitivity (90%). Apelin, TYRO protein tyrosine kinase-binding protein and big endothelin-1 may be potential candidates for the proteins identified with an m/z of 3,264, 9,451 and 4,295 Da, respectively. The predictive model used in the present study may provide an improved understanding of the pathogenesis of NSCLC, and may provide insights for the development of TKI treatment plans tailored to specific patients.