S2997
Invited Speaker
ESTRO 2026
5308 Creating and maintaining large multi-dimensional datasets Gareth Price Division of Cancer Sciences, University of Manchester, UK, United Kingdom This presentation will examine the opportunities and challenges involved in creating and maintaining large multi-dimensional real-world patient datasets, and the role of medical physicists in this rapidly evolving space. By drawing on examples across different modalities, the talk will illustrate how linking heterogeneous data sources enables impactful research, supports learning healthcare system approaches, and can drive improvement in patient care. However, achieving this requires more than advanced technology: it demands appropriate governance, clear data provenance, high- quality data pipelines, and research questions that are appropriately matched to the underlying dataset. The presentation will highlight how medical physicists, positioned at the intersection of clinical practice, science, technology, and regulation, are well-placed to support these efforts. Finally, it will outline future challenges and the need for coordinated interdisciplinary communities to ensure these datasets deliver real clinical benefit. 5309 The power of local and nationwide radiotherapy quality registries Eva Onjukka 1,2 , Giovanna Gagliardi 1,2 , Tufve Nyholm 3 1 Nuclear Medicine and Medical Physics, Karolinska University Hospital, Stockholm, Sweden. 2 Oncology Pathology, Karolinska Institutet, Stockholm, Sweden. 3 Diagnostics and Intervention, Umeå University, Umeå, Sweden The Swedish radiotherapy quality registry, SKvaRT, is designed to collect detailed and standardised treatment data from all Swedish radiotherapy departments with a high degree of automation. All 17 Swedish radiotherapy departments are collecting treatment data in MIQA (Medical Information Quality Archive) and, so far, data from 9 departments are regularly uploaded from MIQA to SKvaRT, indicating good progress towards including all patients treated with radiotherapy in Sweden. A critical component is the unified Swedish nomenclature for radiotherapy structures. Currently, SKvaRT contains data for over 62,000 patients, starting from 2010. The registry is intended for national quality assurance, monitoring of guideline adherence, and research. It has an interactive web-based report. The establishment of Swedish quality registries started in the 1970s. Today, care providers report to 38
acknowledging the fact that each hospital may present unique challenges. A particular issue to address is the “black-box” character of AI-based sCT algorithms, especially relevant for hospitals acquiring commercial solutions and not having information on or access to the training data. Image quality and dosimetric performance testing is necessary for sCT adoption into the clinic and current efforts for establishing benchmark tolerance levels will be discussed. Existing experience in sCT QA protocol development is primarily based on implementation on C-arm linacs and its translation to MR guided radiotherapy (MRgRT) remains nontrivial. The capability of MRgRT machines for online plan adaptation introduces unique challenges for sCT image quality controls. Thus, this talk will finalize with a brief overview of the QA tools needed to achieve fast and safe MRgRT workflows. 5307 Proton therapy: Getting the stopping power right Christian Richter High-Precision Proton Therapy, OncoRay - National Center for Radiation Research in Oncology, Dresden, Germany Accurate dose calculations in proton therapy rely critically on precise stopping power ratio (SPR) prediction derived from CT imaging. This talk reviews advanced methodologies to improve SPR accuracy, with a particular focus on spectral CT approaches, including dual-energy CT (DECT) and emerging photon- counting CT (PCCT). Spectral CT enables a more direct characterization of tissue composition, reducing ambiguities inherent to conventional single-energy CT (SECT) and thereby improving SPR estimation. Special emphasis will be placed on first clinical experiences with PCCT in radiotherapy, based on the world’s first installation of a single-source PCCT system in a radiotherapy department at OncoRay in Dresden. This technology offers the potential for improved image quality and material differentiation at reduced imaging dose. Importantly, independent of the chosen CT modality, robust SPR prediction requires careful attention to acquisition and reconstruction protocols. In particular, minimizing beam hardening effects and ensuring CT number consistency remain essential prerequisites for reducing uncertainties. The session will provide both a forward-looking perspective on the role of spectral CT in proton therapy, and practical guidance to improve SPR accuracy in current clinical workflows. Ultimately, closing the loop from image acquisition to robust patient-specific models for CT-to-SPR conversion is key to turning improved imaging into clinically meaningful gains in proton therapy.
Made with FlippingBook - Share PDF online