Histology differentiation in non-small cell lung cancer using matrix-assisted laser desorption/ionization mass spectrometry.

Abstract

e21043 Background: New developments in the treatment of lung cancer have necessitated the correct histologic differentiation between two major histologic types of NSCLC. Methods: Histology-directed tissue matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS) has been demonstrated to be useful for molecular profiling of common solid tumors. In this approach, mass spectra are obtained from discrete locations on the thin sections of frozen tissues sections. In this study, we evaluated various MALDI MS protocols for the histopathologic classification of NSCLC. Results: When adenocarcinoma and squamous cell carcinoma biopsy samples were compared for protein profile using standard sinapinic acid matrix, the median prediction accuracy of the best histology classifier in test sets was 83.3% in 100 random training-to-test partitions (feature selection P <0.05). When 2,5-dihydroxybenzoic acid/α-cyano-4-hydroxycinnamic acid matrices were used for lipid MALDI MS, the median prediction accuracy of the best histology classifier in test sets was 85.7% in 100 random partitions (feature selection P<0.01). We then validated the discriminatory lipid profile using independent set of 51 NSCLC surgical samples. Eleven discriminatory lipids correctly classified the histology of 80.4% of independent NSCLC surgical tissue samples (41 out of 51) in validation set. Phosphatidylcholine 32:0 (m/z 756.68) was overexpressed in adenocarcinomas. Conclusions: Due to the small amount of tissue required for MALDI analysis, rapid experimental procedure, and high predictive values documented in this study, histology-directed MALDI MS, especially lipid profiling, may be a promising approach to the histologic classification of NSCLC (This work was supported by Converging Research Center Program through the Ministry of Education, Science and Technology of Korea (2010K001121)).