Background: The prediction of the relapse pattern is an important issue in glioblastoma for personalized approach. Molecular factors, such as MGMT promoter methylation, influence relapse in- or out-field of the initial radiotherapy volume. Recently, a recurrent mutation located at position 289 of the extracellular domain of the epidermal growth factor receptor (EGFR^A289mut) has been associated with a more infiltrative phenotype in glioblastoma. The primary objective of the present study was to explore the impact of EGFR^A289mut on the pattern of relapse after a chemo-radiotherapy based treatment of patients suffering from glioblastoma. Methods: An ancillary study from a monocentric prospective cohort of patients suffering from glioblastoma was conducted. All patients received radiotherapy and concomitant temozolomide. The population was divided into two groups according to EGFR^A289 status (mutated versus wild-type). Primary endpoint was the overlap score (varying from 0 to 1) between initial irradiated tumor volume (Vinit) and relapse volume (Vr). Vinit was the initial 95% isodose of the radiotherapy and Vr was delineated using the enhanced MRI T1-weighted part of the relapse tumor. Secondary endpoints explored the impact of other EGFR extracellular mutations, EGFR amplification and EGFRvIII on the relapse pattern, as well as the impact of EGFR^A289mut on survival. EGFR alterations were identified using next-generation sequencing or droplet-digital PCR based methods on formalin-fixed paraffin-embedded samples collected at initial diagnosis. Results: One hundred patients were included: 11% of the population had EGFR^A289mut glioblastoma (n = 11/100). EGFR^A289mut glioblastomas had a relapse pattern more marginal compared to EGFR^A289wt glioblastomas: a mean overlap score Vinit/Vr of 0.78 was observed in the EGFR^A289mut group versus 0.94 in the EGFR^A289wtgroup (p = 0.021). The proportion of EGFR^A289mutglioblastomas in the outfield relapse (overlap score < 0.8) has a trend to be higher than in the infield relapse (overlap score > 0.8): 25% (n = 2/8) versus 9.8% (n = 9/92) for the EGFR^A289wt population, p = 0.21. Neither EGFR amplification nor EGFRvIII did influence overlap score Vinit/Vr. In our population, EGFR^A289mut did not influence survival. Conclusions:EGFR^A289mutinfluences the relapse pattern in a population of patients suffering from glioblastoma. The role of EGFR^A289mut as a decision-making biomarker for personalized radiotherapy should now be investigated in dedicated clinical trial. Clinical trial information: NCT02617745.