Molecular classification of tumors with special reference to EGFR mutation in lung cancer
- First Online:
- Cite this article as:
- Yatabe, Y. Cancer Chemother Pharmacol (2006) 58: 17. doi:10.1007/s00280-006-0311-9
- 45 Downloads
Unsupervised hierarchical clustering in expression profiling analyses allows molecular classification of tumors based on the similarity of genome-wide expression patterns. This new molecular-based classification shares the current pathological classification in part, although it also provides additional clues to identify cancers by their biological groups. Herein, we introduce a novel means of molecular classification of lung cancer. When examining the gene expression profiling analyses of lung cancers published so far, the molecular classification differs from the current classification schema, dividing lung cancers into two distinct branches that do not segregate SCLC and NSCLC, as might be expected. One of the branches includes adenocarcinoma alone, which is associated with a normal expression profile, while the other branch includes all four histological subtypes. We further examined the adenocarcinoma subset of the first branch. This subset of adenocarcinomas is characterized by frequent development in females and non-smokers, expression of thyroid transcription factor-1 and surfactant proteins, and specific involvement of epidermal growth factor receptor (EGFR) gene mutation. Furthermore, this subset is highly distinctive not only among lung cancers but also among carcinoma arising in other tissue sites, in terms of EGFR gene mutation and expression profiles. Although further studies are needed to clarify this adenocarcinoma subset, the distinction should be taken into consideration in lung cancer research and clinical strategies for treatment.