Tumor-host histopathologic variables, stromal myofibroblasts and risk score, are significantly associated with recurrent disease in tongue cancer.

Margin status, a major prognostic parameter in oral cancer, was analyzed vis-à-vis the histopathologic parameters of risk scores and stromal myofibroblasts. Specimens of tongue carcinoma (n = 50) were submitted to a risk score assignment consisting of the worst pattern of invasion, lymphocytic infiltration, and perineural invasion. Frequency of stromal myofibroblasts (alpha-smooth muscle actin stain) was assessed. A triple immunostaining assay with E-cadherin, Ki-67 and alpha-smooth muscle actin was used to identify carcinoma cells undergoing epithelial-mesenchymal transition. Margins were considered 'clean' if the tumor was >or=5 mm away from them. Patients <or=60 years were considered as 'young'. Kaplan-Meier survival analysis with univariate and Cox multivariate regression model with stepwise forward selection, and Fisher's exact tests were used. Abundant myofibroblasts were found in 27 (54%) cases. Carcinoma cells devoid of E-cadherin but amalgamated with the stromal myofibroblasts were identified in 18 (36%) cases. Local recurrence and overall survival were negatively influenced by abundance of stromal myofibroblasts (P = 0.004 and P = 0.008, respectively). High-risk scores (P = 0.011), positive margins, and 'young' age (P = 0.027, each) had an unfavorable impact on recurrence. Multivariate analysis revealed that only abundance of stromal myofibroblasts had an independent adverse effect on local recurrence (hazard ratio [HR] 4.369; P = 0.014; 95% confidence interval [CI], 1.356-14.074). It seems that abundant stromal myofibroblasts (camouflaging some malignant cells) and high-risk scores have an unfavorable impact on the risk of recurrence in particular in 'young' patients. Therefore, the treatment concept should be adjusted accordingly and target concomitantly the epithelial malignancy and its allied stroma.