A series of 78 cases of oral squamous cell carcinoma was analysed by immunohistochemistry for expression of survivin, a recent apoptosis inhibitor. All cases were positive for survivin expression and were divided into two groups using a system of scores. Disease-specific survival curves were calculated according to Kaplan-Meier algorithm, and log rank test was used to compare survival curves. Then, Cox regression analysis was applied to determine the single contribution of covariates on survival rate. So, Cox analysis allowed us to detect the variables most associated to survival. Among the studied variables, such as grade of differentiation, tumor size, stage, recurrence of disease, lymph node presence, only stage and recurrence of disease were predictors of outcome; however, when we analyzed the survival without considering recurrence (that was the stronger predictor of death), a stepwise Cox analysis showed that Survivin, stage and grade of differentiation are significantly associated to survival, with a higher value for Survivin. These data suggest that survivin expression may identify cases of oral squamous cell carcinoma with more aggressive and invasive phenotype and, therefore, could influence the decision for the therapy at the time of diagnosis.