Accurate assessment of malignancy in oral squamous cell carcinoma is essential to optimize treatment planning. To detect a biomarker related to malignant propensity in gingival squamous cell carcinoma (GSCC), quantitative gene expression analysis of tetraspanin family genes was conducted. In 73 cases of GSCC, total RNA was extracted from carcinoma tissues, and gene expression was analyzed by quantitative real time-PCR. Six tetraspanin family genes (CD9, CD63, CD81, CD82, CD151, NAG-2) were investigated. Housekeeping genes (ACTB and GAPDH), anchor protein genes (JUP and PXN) and an integrin gene (ITGA3) were used as reference genes. Forty-five gene expression ratios were calculated from these 11 gene expression levels and were analyzed with clinical parameters using multivariate statistical methods. According to the results of the logistic regression analysis subjecting cervical lymph node metastasis as a target variable, CD9/ACTB (p = 0.013) or CD9/CD82 (p = 0.013) in addition to tumor size (p = 0.028) were detected as significant factors. In Cox proportional hazards regression analysis, delayed cervical lymph node metastasis (p = 0.039) and tumor cell positive surgical margin (p = 0.032) in addition to CD151/GAPDH (p = 0.024) were detected as significant factors for death outcome. A Kaplan-Meier survival curve presented a significantly lower survival rate of the group with a CD151/GAPDH value of 10 or more (log rank and generalized Wilcoxon tests: p = 0.0003). Results of this study present the usefulness of CD9 and CD151 expression levels as biomarkers for assessment of malignancy in GSCC. They also indicate that detection of residual tumor cells at the surgical margin and the biological malignancy of a tumor interdependently affects prognosis.
Copyright 2008 UICC.