Expression Profile–Defined Classification of Lung Adenocarcinoma Shows Close Relationship With Underlying Major Genetic Changes and Clinicopathologic Behaviors

Abstract

Purpose This study was conducted to gain insight into the relationship between expression profiles and underlying genetic changes, which are known to be important for the pathogenesis of lung cancers. Methods Expression profiles of 18,175 unique genes and three major targets for genetic changes, p53, epidermal growth factor receptor (EGFR), and K-ras, were investigated in 149 patients with non–small-cell lung cancer, including 90 patients with adenocarcinoma to determine their relationships with various clinicopathologic features and Gene Ontology (GO) terms. Results This study successfully established a basis for expression profile-defined classification, which can classify adenocarcinomas into two major types, terminal respiratory unit (TRU) type and non–TRU type. Our GO term–based identifier of particular biologic processes, molecular functions, and cellular compartments clearly showed characteristic retention of normal peripheral lung features in TRU type, in sharp contrast to the significant association of non–TRU type with cell cycling and proliferation-related features. While significantly higher frequency of EGFR mutation was observed in TRU type, we found that the presence of EGFR mutations was a significant predictor of shorter postoperative survival for TRU type, independent of disease stage. We were also able to identify a set of genes in vivo with significant upregulation in the presence of EGFR mutations. Conclusion This study has shed light on heterogeneity in lung cancers, especially in adenocarcinomas, by establishing a molecularly, genetically, and clinically relevant, expression profile-defined classification. Future studies using independent patient cohorts are warranted to confirm the prognostic significance of EGFR mutations in TRU-type adenocarcinoma.