“…While sex and tumor location have been predictors of survival in other oral cavity cancer studies, they were not associated with survival in the univariate models and were thus not included in the multivariate analysis. Additional prognostic factors were added stepwise into the analysis, as they were associated with survival in the literature (Ash et al 2000;Dubner and Heller 1993;Mücke et al 2009;Pandey et al 2009;Patel et al 2008;Rogers et al 2009;Shah and Gil 2009). Only recurrence of tumor was associated with survival in the multivariate analysis controlling for tumor and N stage, UICC stage, extent of mandibulectomy, and age, with a hazard ratio of 1.459 (P = 0.039, 95% CI, 1.020-2.081).…”