Cancer Stem Cell based molecular predictors of tumor recurrence in Oral squamous cell carcinoma.

OBJECTIVE:This study aimed to identify the cancer stem cell specific biomarkers that can be effective candidate prognosticators of oral squamous cell carcinoma. DESIGN:Microarray-based meta-analysis derived transcriptional profile of head and neck cancers was compared with the Cancer Stem Cell database to arrive at a subset of markers. This subset was further co-related with clinico-pathological parameters, recurrence and survival of oral cancer patients (n = 313) in The Cancer Genome Atlas database and in oral cancer (n = 28) patients. RESULTS:Meta-analysis in combination with database comparison identified a panel of 221 genes specific to head and neck cancers. Correlation of expression levels of these markers in the oral cancer cohort of The Cancer Genome Atlas (n = 313) with treatment outcome identified 54 genes (p < 0.05 or fold change >2) associated with disease recurrence, 8 genes (NQO1, UBE2C, EDNRB, FKBP4, STAT3, HOXA1, RIT1, AURKA) being significant with high fold change. Assessment of the efficacy of the subset (n = 54) as survival predictors identified an additional 4 genes (CDK1, GINS2, PHF5 A, ERBB2) that co-related with poor disease-free survival (p < 0.05). CDK1 showed a significant association with the clinical stage, margin status and with advanced pathological parameters. Initial patient validation indicated that CDK1 and NQO1 significantly co-related with the poor disease-free and overall survival (p < 0.05). CONCLUSION:This panel of oral cancer specific, cancer stem cell associated markers identified in this study, a subset of which was validated, will be of clinical benefit subject to large scale validation studies.