Epigenetic aberration, such as dysregulation of DNA mathylation is often associated with human carcinogenesis. To analyze DNA methylation profiles of oral cancers, and to identify the methylation pattern specific to the clinicopathological features, we performed genome-wide methylation analysis using Infinium HumanMethylation27 BeadArray system. We found 2,361 aberrantly methylated CpG sites in 59 oral cancers compared to 3 normal mucosal tissues. By unsupervised clustering analysis, oral cancer was divided into two methylation epigenotypes that are significantly correlated with incidence of lymph node metastasis and differentiation grades. We selected the classifier genes that can predict tumor metastasis in the training set, and demonstrated that they are validated to be predictive in the independent test sample, preoperative biopsy samples. In summary, we identified two methylation epigenotypes correlated with clinical outcome in oral cancer, and candidate classifier genes that would be useful in patient stratification for intensive therapy to overcome the tumor recurrence.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 5000. doi:1538-7445.AM2012-5000