To seek out novel promising biomarkers for predicting lung adenocarcinoma (LUAD) prognosis, we conducted this study. First, 279 upregulated and 37 downregulated differentially expressed genes were obtained from LUAD and para-carcinoma tissues by the Affymetrix GeneChip Human Transcriptome Array. Then, we randomly classified samples of LUAD data set GSE31210 as training and testing sets in a 1:1 ratio. Alcohol dehydrogenase 1C (ADH1C) and secreted phosphoprotein 1 (SPP1) were finally identified correlating with the LUAD survival through least absolute shrinkage and selection operator penalized Cox proportion hazards regression model, and applied to build a 2-gene signature related to prognosis in training set. Univariate and multivariable survival analyses suggested that overall survival (OS) and relapse-free survival (RFS) in the 2-gene signature low-risk group were better than the high-risk group. Kaplan-Meier curves proved that elevated ADH1C expression and reduced SPP1 expression were related to better OS and RFS. Besides, the SPP1 expressed higher in LUAD than para-carcinoma tissues using quantitative reverse transcription polymerase chain reaction assay. Finally, the association between the two genes and clinicopathological parameters in 80 LUAD were analyzed, it is suggested that SPP1 was relevant to epidermal growth factor receptor mutation. These findings indicated that ADH1C and SPP1 might be novel promising biomarkers for predicting LUAD prognosis.
Keywords: alcohol dehydrogenase 1C; biomarker; lung adenocarcinoma; prognosis; secreted phosphoprotein 1.
© 2019 Wiley Periodicals, Inc.