Search

two cohorts by incorporating the resulting coefficients of multivariable Cox regression model into a score that utilizes linear gene expression values. This gene expression classifier was validated in 6 additional publicly available datasets of stage I/II lung SCC (N 1⁄4 358). The classifier identified high-risk patients in multiple large-scale and geographically diverse cohorts (N 1⁄4 570). The results appear to be independent of race and gene expression platform. Canonical pathways associated with this signature encompass proteins involved in signal transduction, tissue remodeling, and cell motility that would broadly lead to cancer cell migration, invasion and proliferation, suggesting that the gene signature identifies molecular subsets of patients with clinical relevance. This gene classifier could be used to guide clinical decisions after surgical resection. Thus, we would advocate that this classifier be incorporated into prospective trials for further evaluation of its clinical effectiveness.