ESTRO 2021 Abstract Book

S580

ESTRO 2021

selection of the best patient-specific plan, without increasing planning workload.

PD-0750 When your MR Linac is down: can an automised pipeline bail you out of trouble? L. Placidi 1 , D. Cusumano 1 , A. Alparone 2 , L. Boldrini 1 , G. Chiloiro 1 , A. Romano 1 , M. Nardini 1 , V. Valentini 1 , L. Indovina 1 1 Fondazione Policlinico Universitario A. Gemelli IRCCS, Radiation Oncology, Rome, Italy; 2 Tecnologie Avanzate, Tecnologie Avanzate, Turin, Italy Purpose or Objective The unique treatment delivery technique provided by magnetic resonance guided radiotherapy (MRgRT) can represent a significant drawback when hybrid system fail occurs, since no equivalent technology is generally available to transfer in treatment patients. This retrospective study proposes a pipeline to completely automatize the workflow necessary to shift a MRgRT treatment to a traditional radiotherapy Linac, evaluating the dosimetric, planning, and quality of the obtained plans. Materials and Methods Patients undergoing treatment during the last actual MRgRT system fail were retrospectively included in this exploratory study. The pipeline was based on a tool installed on a third-party Treatment Planning System (TPS) able to mimic the original MR Linac dose distribution and ad-hoc Python script have been implemented to completely automatize patient’s reassignment process to standard linacs in order to carry on the interrupted treatment. Voxel-based dose mimicking optimization converts the predicted dose distribution to a complete treatment plan with dose calculation using a collapsed cone convolution (CCC) dose engine. The so obtained dose distribution (AUTO) has been compared with the distribution obtained in the conventional radiotherapy Linac, where patients have been transferred to (MAN). Plan comparison has been performed in terms of time required to obtain the final dose distribution, DVH parameters, dosimetric indexes (CI, HI and gradient measure) and visual analogue scales (VAS) scoring by expert radiation oncologists that scored the plans without any information on how the mimicked dose distribution was achieved. Results Automatic generation of the plans has been obtained within 10 minutes, for all the six considered cases. An example of mimicked dose is depicted in figure 1.