Study of distinct protein profiles for early diagnosis of NSCLC using LCM and SELDI-TOF-MS

Abstract

No biomarker has been available to detect early lung cancer so far. The aim of this study is to screen biomarker patterns for early diagnosis of non-small cell lung cancer (NSCLC) using laser capture microdissection (LCM) and surface-enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI-TOF-MS). The 3 groups of the interested cells from 13 NSCLC tissues, 11 normal lung tissues (out of the 13 NSCLC patients), and 6 benign lung diseased tissues (BLD) were successfully separated by LCM, respectively, and the homogeneities of each type of the cell populations in the three groups were estimated to be over 95%. One-hundred- and twenty-three M/Z peaks were found in the NSCLCs and normal lungs, and between the two groups the relative intensity of 98 M/Z peaks was significantly different (P < 0.05) using SELDI-TOF-MS. The diagnostic pattern constructed using support vector machine (SVM) including three proteins, M/Z 4282, 3201, and 4252 Da, respectively, showed maximum Youden Index (YI). The pattern was validated by leave-one-out cross validation (LOOCV) and the results showed that the sensitivity was 100.0%, specificity 90.9%, and positive predictive value (PPV) 92.9%. In the NSCLCs and BLDs 188 M/Z peaks were determined and 54 showed statistically difference (P < 0.05). The sensitivity, specificity, and PPV of the diagnostic pattern consisting of two proteins, M/Z 3204 and 3701 Da, were all 100.0%. So, by using LCM we have successfully purified the interested cells and solved the problem of heterogeneity of lung cancer tissue. SELDI protein chip coupled with SVM could effectively screen the differentially expressional protein profiles and eventually establish biomarker patterns with high sensitivity and specificity.