S2027
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
automated VMAT planning at our institution (Figure 2), demonstrating feasibility for translating research algorithms into clinical treatment planning workflows. The platform has seen increasing adoption, averaging over 1,000 downloads per month in the past six months.
Informatics, The University of Edinburgh, Edinburgh, United Kingdom. 3 Department of Clinical Medicine, aarhus university, Aarhus Centrum, Denmark. 4 Varian Medical Systems,, Siemens Healthineers, New York, USA Purpose/Objective: We have developed PortPy (Planning and Optimization for Radiation Therapy in Python), an open-source package designed to accelerate research and development in radiotherapy treatment planning optimization. PortPy provides curated benchmark datasets and a suite of optimization methods for IMRT, VMAT, and IMPT, including both classical model-based and AI-based dose prediction approaches. It enables transparent, reproducible research across topics such as automated and adaptive planning and supports emerging photon and proton modalities. In addition, PortPy interfaces with commercial TPS, allowing rapid translation of research algorithms into clinical workflows. Material/Methods: We have implemented a variety of classical planning optimization techniques for IMRT, VMAT, and IMPT, including fluence map optimization, leaf sequencing, direct-aperture optimization, and column-generation. We have also incorporated an AI-based 3D dose prediction approach. For research on non-convex planning tasks (e.g., DVH-based planning, VMAT optimization, beam angle selection), PortPy provides mixed-integer programming formulations capable of computing globally optimal plans. Although these methods are computationally intensive and not practical in routine clinical use, they provide invaluable ground-truth results for benchmarking novel and computationally efficient techniques. PortPy includes publicly released datasets of 100 lung and 129 prostate cancer patients, each containing CT imaging, contours, clinical plans, and dose-influence matrices extracted from Eclipse via the scripting API. Using PortPy seamless TPS integration, we have clinically deployed PortPy to support automated planning, in which PortPy performs optimization and Eclipse performs final dose calculation. Results: Various use cases in IMRT, VMAT, and IMPT planning— leveraging both classic and AI-based optimization techniques—are demonstrated in more than ten easy- to-follow Jupyter notebook examples on the PortPy GitHub organization (https://github.com/PortPy- Project/PortPy). As illustrated in Figure 1, the platform comprises three core modules: (1) Data Management for accessing curated patient datasets, (2) Plan Generation using classical optimization or AI-based planning methods, and (3) Plan Evaluation for visualization and dose metric assessment. PortPy has been integrated into Eclipse and is used in routine
Conclusion: PortPy enables transparent and reproducible research in radiotherapy treatment planning optimization. Its native TPS integration enables direct clinical evaluation of new planning methods, helping bridge the gap between algorithm development and clinical implementation. References: Zarepisheh et al. (Medical Physics, 2019) DOI: 10.1002/mp.13572Hong et al. (Advances in Radiation Oncology, 2019) DOI: 10.1016/j.adro.2019.11.0052 Keywords: treatment planning optimization, IMRT, VMAT Digital Poster 5081 Optimizing MRI Sequences for Liver SBRT Planning: A Retrospective Analysis Michael I Lock 1 , Michael S Kim 2 , Robyn Murphy 1 , Daniele Wiseman 3 , Jacob Cottreau 1 , Gary Brahm 1 1 Oncology, London Health Sciences Centre, London, Canada. 2 Oncology, University of Maryland, College Park, USA. 3 Radiology, London Health Sciences Centre, London, Canada
Made with FlippingBook - Share PDF online