To analyze gene expression in oral cancer, we produced a specialized in-house cDNA microarray. The cDNA library was constructed from surgical specimens of oral squamous cell carcinoma (SCC) using an oligo-capping method. cDNA clones (n=4,800) were randomly selected and their 5'-end nucleotide sequences were determined. Overlapping clones were excluded, and 1,423 independent clones were selected and used for microarray production. Compared to the public nucleotide sequence database, 61% of our cDNA clones were full-length. By correlating expression patterns across SCC cell lines, we identified 53 genes (7 up-regulated and 46 down-regulated) that are differentially expressed in SCC cell lines compared to normal mucosa. Semi-quantitative RT-PCR analysis confirmed these findings and validity of our cDNA microarray. Using specimens from SCC patients, we investigated the expression status of the IL-1ra gene, which showed the down-regulated gene by microarray analysis. Gene expression clearly fell in the SCC specimens relative to their references, which indicated that our in-house cDNA microarray system rapidly identified and characterized candidate biomarkers for clinical use.