Somatic mutations in the epidermal growth factor receptor (EGFR) gene were associated with sensitivity to small molecule tyrosine kinase inhibitors for patients with lung adenocarcinomas. In this research, EGFR mutation status was analyzed by DNA sequencing in 153 lung adenocarcinoma tissues. Of these, 75 samples carried EGFR mutations, including 29 with E19del mutation, 33 with L858R mutation, 7 with T790M mutation, and 6 with multiple mutations. Then, 30 samples including 10 with wild type (wt)-EGFR, 10 with L858R and 10 with E19del mutations were selected for Raman and immunohistochemistry (IHC) analyses. After removing the spectra from normal and non-mutated regions, 441 spectra were found appropriate for Raman analysis: 149 from wt-EGFR, 135 from L858R and 157 from E19del mutations. The Raman peaks at 675, 1107, 1127 and 1582 cm(-1) were significantly increased in wt-EGFR tissues which can be attributed to specific amino acids and DNA. The Raman peaks at 1085, 1175 and 1632 cm(-1) assigned to arginine were slightly increased in L858R tissues. The overall intensity of E19del tissues was weaker than others due to exon 19 deletion that removes residues 746-750 of the expressed protein. Principal component analysis (PCA) and support vector machine (SVM) were applied for final prediction. The PCA/SVM algorithm yielded an overall accuracy of 87.8% for diagnosing L858R or E19del from wt-EGFR tissues. Finally, RS provides a simple, rapid and low-cost procedure based upon the molecular signatures for predicting EGFR mutation status.