Pediatric radiotherapy is an anxiety-provoking situation that can, on one hand, impair the child's cooperation and require sedation, and on the other hand,burden the care pathway by generating significant anxiety in both the child and their parents. Companion robots represent an innovative avenue for patient support, but they have never been evaluated in the context of radiotherapy. This study examines the effect of MIROKI, the first social robot capable of communication through a Large Multimodal Model (LMM), on children's state anxiety during radiotherapy. An experimental single-case design (SCED) will compare phases with and without the robot (withdrawal/reversal), both in the waiting room and in the irradiation bunker. The primary outcome variable is anxiety, assessed in several ways: through questionnaires at the beginning and end of treatment; through an objective measure of cardiac activity (heart rate in bpm) during the radiotherapy session, with or without the MIROKI robot; and through an observational grid assessing the child's behavior. The intensive repeated measure (bpm) constitutes the preferred variable in this SCED design, while behavioral observation and anxiety questionnaires will support interpretation. The results could pave the way for reducing sedation and improving the overall care pathway in pediatric radiotherapy.