ESTRO 2025 - Abstract Book

S3613

Physics - Quality assurance and auditing

ESTRO 2025

to release medical software compliant with MDR requirements, including visual inspection and documentation was eight hours. Ten new releases have been deployed over the last 16 months. In contrast, older applications would only be released once every few years, due to the high workload involved. Conclusion: The introduction of microservice-based architecture and automatic build pipelines has reduced the complexity of testing and releasing in-house built medical software according to the MDR, enabling the deployment of new software releases within eight hours. Automatically generated test results and documentation reports simplify workflows and minimizes workload. Placing the responsibility in a multidisciplinary DevOps team ensures sustainable development, operation and maintenance of the software by spreading knowledge, reduce individual workload, and having a strong collaboration with clinical wishes.

Keywords: Medical software, In-house built, MDR

References: [1] What is Azure Devops. https://learn.microsoft.com/en-us/azure/devops/user-guide/what-is-azure devops. Accessed 2024-10-28. [2] RabbitMQ. https://www.rabbitmq.com/. Accessed 2024-11-04. [3] Rossberg, Joachim. Agile project management with azure DevOps: Concepts, templates, and metrics . Apress, 2019.

1402

Digital Poster Novel customized dynamic cardiac phantom for complete end-to-end-testing of STereotactic Arrhythmia Radioablation (STAR) Melanie Grehn 1 , Adrian Zaman 2 , Jingyang Xie 3 , Marc Delaperriere 1 , Achim Schweikard 4 , Frank-Andre Siebert 1 , Juergen Dunst 1 , Roland R. Tilz 5 , Evgeny Lian 2 , Oliver Blanck 1 1 Department of Radiation Oncology, University Medical Center Schleswig-Holstein, Kiel, Germany. 2 Department of Internal Medicine III, Cardiology, University Medical Center Schleswig-Holstein, Kiel, Germany. 3 Institute of Robotics and Cognitive Systems, University of Luebeck, Luebeck, Germany. 4 Institute of Robotics and Cognitive Systems, University of Luebeck, Kiel, Germany. 5 Department of Rhythmology, University Medical Center Schleswig-Holstein, Luebeck, Germany Purpose/Objective: STereotactic Arrhythmia Radioablation (STAR) is a new treatment option for ventricular tachycardia [1], however, the full interdisciplinary workflow is not yet standardized, and a realistic and complete quality assurance test of the whole treatment-chain (end-to-end-test) does not exist. We now evaluated a novel customized dynamic-cardiac phantom for full end-to-end-testing, starting from (electro)anatomical mapping (EAM) and target definition to radiation oncology with treatment planning and dose application. Material/Methods: For the CIRS dynamic-cardiac-phantom (Sun Nuclear, Melbourne, USA) we customized a new insert dedicated to STAR (Fig1). For each part of the treatment chain, we simulated different cardiac motion patterns (SI:3.1/4.6 mm, AP:1.3/3.7 mm, LR:1.1/4.3 mm, 70 bpm) [2] and a sinus-curve (15 mm amplitude, breathing cycle: 4 s) for breathing motion. EAM was realized with a single-tip mapping-catheter with CARTO-3 (Biosense Webster, New Brunswick, USA) with and without ECG-gating. As planning imaging, we performed ECG-gated 4D-cardiac-CT and 3D/4D planning-CT. Target transfer from marked EAM-screenshots to the planning-CT was realized with CARDIO-RT (University of Luebeck, Germany) [3]. Geometric and dosimetric accuracy of the treatments were measured with ion-chambers, gafchromic-film and EPID-dosimetry on a c-arm linac (TrueBeam STx, Varian, Palo Alto, USA) with and without gating and on a robotic linac (CyberKnife VSI, Accuray, Sunnyvale, USA) with and without tracking.