[Stereotactic irradiation in head and neck cancers]

Stereotactic radiotherapy is increasingly used in head and neck tumours, either as a boost for dose escalation/early salvage, or in the reirradiation setting. We aimed to assess the level of evidence for each clinical setting and to discuss the different dose and frationation regimens. A search of the French and English literature was performed on PubMed until December 2013. Stereotactic reirradiation of locally recurrent squamous cell carcinomas can be performed with overall survival rates of about 12 months with good quality of life, and acceptable toxicity, based on several phase 2 trials and retrospective studies. Nasopharyngeal carcinomas may be irradiated with even better control rates. Late severe toxicities yield up to 20-30%. Patient and tumour selection criteria (limited volume) and dose constraints to the carotids (cumulative dose 110 Gy or less, to avoid the risk of potentially lethal carotid blowout) must be carefully chosen. Fractionated regimens (at least five fractions) should be preferred (30 Gy in five fractions to 36 Gy in six fractions). Methods derived from stereotactic, intensity-modulated radiotherapy (IMRT) may be used with conventional fractionation for larger tumours. Stereotactic irradiation may be associated with cetuximab; data with chemotherapy or other targeted therapies are still lacking. Stereotactic irradiation is also used as a boost after 46 Gy IMRT in several institutions or for early salvage (8 to 10 weeks following full dose irradiation with evidence of residual tumour) in squamous or nasopharyngeal carcinomas. Such indications should be evaluated prospectively in clinical trials. Data in salivary gland and sinonasal neoplasms are still scarce. In conclusion, stereotactic body radiation therapy has the potential as a boost or in the reirradiation setting to improve local control in head and neck tumours. Careful hypofractionation with planning caring for the dose to the main vessels is highly recommended. Prospective studies with prolonged follow-up (at least 2 years) should be encouraged.