Abstract： Background Patients with single station mediastinal lymph node (N2) non-small call lung ccancer (NSCLC) have a better prognosis than those with multilevel N2.The molecular factors which are involved in disease progression remain largely unknown.The purpose of this study was to investigate gene expression differences between single station and multilevel N2 NSCLC and to identify the crucial molecular factors which are associated with progress and prognosis of stage N2 NSCLC.Methods Gene expression analysis was performed using Agilent 4x44K Whole Human Genome Oligo Microarray on 10 freshfrozen lymph node tissue samples from single station N2 and paired multilevel N2 NSCLC patients.Real-time reverse transcription (RT)-PCR was used to validate the differential expression of 14 genes selected by cDNA microarray of which four were confirmed.Immunohistochemical staining for these validated genes was performed on formalin-fixed,paraffinembedded tissue samples from 130 cases of stage N2 NSCLC arranged in a high-density tissue microarray.Results We identified a 14 gene expression signature by comparative analysis of gene expression.Expression of these genes strongly differed between single station and multilevel N2 NSCLC.Four genes (ADAM28,MUC4,CLDN1,and IGF2) correlated with the results of microarray and real-time RT-PCR analysis for the gene-expression data in samples from 56 NSCLC patients.Immunohistochemical staining for these genes in samples from 130 cases of stage N2 NSCLC demonstrated the expression of IGF2 and CLDN1 was negatively correlated with overall survival of stage N2 NSCLC.Conclusions Our results suggest that the expression of CLDN1 and IGF2 indicate a poor prognosis in stage N2 NSCLC.Further,CLDN1 and IGF2 may provide potential targeting opportunities in future therapies.