Abstract Background Direct exposure of implantable cardioverter-defibrillators (ICDs) during radiotherapy is still considered potentially harmful, or even unsafe, by manufacturers and current recommendations. The effects of photon beams on ICDs are unpredictable, depending on multiple factors, and malfunctions may present during exposure. Purpose To evaluate transient ICD malfunctions by direct exposure to doses up to 10 Gy during low-energy RT, forty-three contemporary wireless-enabled ICDs, with at least 4 months to elective replacement indicator (ERI) were evaluated in a real-time in-vitro session in three different centres. Methods All ICDs had baseline interrogation. Single chamber devices were programmed to the VVI/40 mode and dual or triple chamber devices were programmed to the DDD/40 mode. Rate response function and antitachycardia therapies were disabled, with the ventricular tachycardia (VT)/ventricular fibrillation (VF) detection windows still active. A centring computed tomography was performed to build the corresponding treatment plan and the ICDs were blinded randomized to receive either 2-, 5- or 10-Gy exposure by a low photon-energy linear accelerator (6MV) in a homemade water phantom (600 MU/min). The effective dose received by the ICDs was randomly assessed by an in-vivo dosimetry. During radiotherapy, the ICDs were observed in a real-time session using manufacturer specific programmer, and device function (pacing, sensing, programmed parameters, arrhythmia detections) was recorder by the video camera in the bunker throughout the entire photon exposure. All ICDs had an interrogation session immediately after exposure. Results During radiotherapy course, almost all ICDs (93%) recorded major or minor transient electromagnetic interferences. On detail, sixteen ICDs (37.2%) reported atrial and/or ventricular oversensing, with base-rate-pacing inhibition and VT/VF detection. Twenty-four ICDs (55.8%) recorded non clinically relevant noise, and no detections were observed. Only three ICDs (7%) reported neither transient malfunction nor minor noise, withstanding direct radiation exposure. At immediate post-exposure interrogation, the ICDs that recorded major real-time malfunctions had VT/VF detections stored in the device memory. In none of the ICDs spontaneous changes in parameter settings were reported. Malfunctions occurred regardless of either 2-, 5- or 10-Gy photon beam exposure. Conclusions Transient electromagnetic interferences were observed in most of the contemporary ICDs during radiotherapy course, regardless of photon dose. To avoid potentially life-threatening ICD malfunctions such as pacing inhibition or inappropriate shock delivery, magnet application on the pocket site or ICD reprogramming to the asynchronous mode are still suggested in ICD patients ongoing even low energy radiotherapy exposure. Funding Acknowledgement Type of funding source: None