Salvage surgery is the mainstay of treatment for recurrences or secondary primary tumors in areas that were irradiated earlier. However, locoregional recurrence remains the main cause of death after surgery. Adjuvant reirradiation dramatically reduces locoregional recurrences but the risk-benefit ratio seems to be advantageous mostly for residual microscopic disease. In contrast, the rate of distant metastasis among reirradiated patients indicates that the local treatment alone is not sufficient. Full-dose exclusive chemo-reirradiation (over 60 Gy) can cure a subset of patients when surgery is not feasible. However, reirradiation is associated with a significant rate of severe toxicity and should, therefore, be compared with chemotherapy in randomized trials. Accrual may be difficult because of selection biases such as tumor volume, small volumes (largest axis less than 3-4 cm) being more likely to be irradiated. In addition, patients in poor general condition with severe comorbidities, organ dysfunction, or incomplete healing after salvage surgery, are unlikely to benefit from reirradiation. Noteworthy volumes to be reirradiated must be established between the head and neck surgeon and the radiation oncologist: the definition of the clinical target volume should be taken into account, the natural history of recurrent tumors, especially with regard to extension modalities, and the absence of strict correlation between imaging and histological real extension. This is even more critical with the advent of new irradiation techniques. Chemotherapy associations and new radiosensitizing agents are also under investigation. Comparison between reirradiation modalities is difficult because most trials are phase 2 mono-institutional trials. As selection of patients is a key issue, only phase 3 multiinstitutional trials can provide definitive results.