Salvage surgery in head and neck cancer: External validation of predictors of disease-specific survival.

OBJECTIVE:Salvage surgery in recurrent head and neck squamous cell carcinoma (HNSCC) is associated with poor outcomes. There is great interest to better identify suitable surgical patients. This study aimed to validate the proposal of Hamoir et al. to use three items to predict the outcomes of salvage surgery. MATERIALS AND METHODS:Single-center analysis of 577 patients undergoing salvage surgery for recurrent HNSCC during the period 1985-2016, with a minimal follow-up of 2 years. Patients were classified according to the prediction modeling proposed by Hamoir et al. This prediction modeling is based on three predictors: the tumor site (larynx vs non-larynx), initial staging (stage I-II vs stage III-IV), and site of recurrence (local or regional vs locoregional). RESULTS:Five-year disease-specific survival after salvage surgery was 54.0% (95% CI: 49.6-58.4%). Applying the prediction modeling tested, a total of 212 patients (36.7%) patients had no predictors of poor prognosis, 185 (32.1%) had one predictor, 146 (25.3%) two predictors, and 34 (5.9%) three predictors. Five-year disease-specific survival for patients with none, one, two and three predictors were 82.2%, 47.2%, 29.5% and 20.2% respectively (P = 0.0001). CONCLUSIONS:We have validated the prediction modeling based on the location of the tumor, initial staging and locoregional recurrence proposed by Hamoir et al. This prediction model is easy to apply and provides good information about the possibilities of success of salvage surgery.