Purpose/Objective: The aim of the project was to study the role of modified radiotherapy (RT) on overall survival. Materials/Methods: The Meta-Analysis of Radiotherapy in Carcinomas of Head & Neck (MARCH) group performed a meta-analysis using updated individual patient data comparing conventional RT to hyperfractionated or accelerated RT, or both, as radical treatment, in locally advanced head and neck squamous cell carcinoma Trials including only nasopharyngeal carcinoma were not eligible. The logrank-test, stratified by trial, was used for comparison and the hazard ratio (HR) of death or loco-regional failure (LRF) was calculated. Results: Fifteen randomized trials (1970-1998) with 6,515 patients were included. One trial had four-arms and contributed to three comparisons. Its control group was counted three time in the analysis that was then performed on 7,073 patients. The median follow-up period was 5.8 years. The pooled HR of death was 0.91 (95% confidence interval: 0.86-0.97; p=0.003) corresponding to an absolute survival benefit of 3% for modified RT, from 36% to 39%, at 5 years. The pooled HR of LRF was 0.82 (0.77-0.88; p<0.0001) corresponding to an absolute benefit on loco-regional control of 7% for modified RT, from 46% to 53%, at 5 years. Three groups of trials were defined according to the total dose in the modified RT arm as compared to the conventional arm: increased (hyperfractionated RT), similar (accelerated RT) and decreased (very accelerated RT). There was a significant interaction (p=0.03) for survival, but not for loco-regional control (p=0.14), between the type of RT and treatment effect (table). The largest effect was observed with increased total dose. The heterogeneity of treatment effect between trials disappeared after exclusion of the 3 trials having included less than 110 patients for both endpoints. There was no evidence that any subgroup of patients defined according to tumor site benefited more or less of modified RT than any other group. Conclusions: Modified radiotherapy led to a small but significant improvement in survival and loco-regional control. The benefit on survival is dependent on the type of radiotherapy. Grants from Association Recherche Cancer & Programme Hospitalier de Recherche Clinique. Tabled 1RT type according to total doseTrial numberPatient numberHazard ratio of deathHazard ratio of LRFIncreased41,3500.780.76Similar83,8180.960.80Decreased51,9050.930.91Total177,0730.910.82 Open table in a new tab