We examined the gene expression profiles of esophageal squamous cell carcinomas (ESCCs) with respect to degree of invasive depth and lymph node (LN) involvement in a large cohort. We used high-density oligonucleotide microarrays to examine the expression of 22,115 genes in 54 ESCCs and 11 non-cancerous esophageal tissues. We found that 4,155 genes were biologically significant in both ESCC and non-cancerous esophageal tissue by analysis of Present Call (hybridization quality by Affymetrix) throughout all samples. From these genes, we used a supervised learning method to select genes responsible for the development of ESCC. We found that 999 genes were expressed differentially in pT1/pN0 tumors vs. non-cancerous esophageal tissue. In the same manner, 48, 66 and 30 genes were expressed differentially in pT1/pN0 tumors vs. pT1/pN1 tumors, pT1/pN0 tumors vs. pT2-4/pN0 tumors and pT2-4/pN0 tumors vs. pT2-4/pN1 tumors, respectively. Intriguingly, there were no overlaps between the 48 LN metastasis-related genes of pT1 tumors and the 30 LN metastasis-related genes of pT2-4 tumors, suggesting that ESCCs with distinct invasive depths express different genes linked to LN metastasis. Our present results suggest that the degree of invasive depth must be considered when predicting LN metastasis of ESCC from gene expression profiles.

Related Articles