Preoperative radiochemotherapy and radical resection for stages II to IV oral and oropharyngeal cancer: grade of regression as crucial prognostic factor.

The purpose of this study was to assess the prognostic value of histological response to preoperative radiochemotherapy in an established multimodal therapy concept for advanced oral and oropharyngeal cancer. Two hundred and twenty-two patients who underwent preoperative radiochemotherapy (RCT: 50 Gy, mitomycin C and fluorouracil) and radical surgery were retrospectively evaluated. Resected tumours of all patients were histologically analysed and response to RCT was classified in histopathological grades of regression (RG). In a multivariate statistical analysis, RG was compared with established factors regarding their predictive value for overall and disease-specific survival. The 5-year overall survival probability in the different groups of histopathological regression grades were: RG1 (no vital tumour): 73.4%, RG2 (minimal tumour remnants encompassing less than 5%): 72.1%, RG3 (5-50% vital tumour cells): 41.9%, RG4 (more than 50% vital tumour): 37.9%. For disease-specific survival probability no significant differences were found between both groups of "responders" (RG1 and RG2) nor between "non-responders" (RG3 and RG4), whereas responders and non-responders differed significantly from each other (log-rank test; p < 0.001). T-classification, N-classification and disease stage, histological grading, tumour site, age, and sex had less prognostic value than RG in a Cox regression model. In the neoadjuvant multimodal therapy concept, histological response to preoperative RCT is a crucial prognostic factor even when surgical R0-resection is accomplished. Thus, non-responders have to be regarded as high-risk patients for recurrence and may benefit from further therapy.