S2333
Physics - Quality assurance and auditing
ESTRO 2026
Conclusion: A multi-lesion SRS anthropomorphic head phantom was designed and created with end-to-end testing capabilities. This phantom helps ensure accuracy for treatment using multi-lesion SRS techniques, particularly with potential use in credentialing for multi-center clinical trials for SRS. Keywords: Multi-SRS, IROC phantom Proffered Paper 829 National intercomparison of re-irradiation techniques Lucy Griffiths, Colin Kelly, Ciaran Malone, Aodh MacGairbhith Medical Physics, Saint Luke’s Radiation Oncology Network, St. James’ Hospital, Dublin, Ireland Purpose/Objective: Re-irradiation (reRT) presents unique challenges in radiotherapy due to the complexity of accurately assessing cumulative doses to organs at risk (OARs). As patients increasingly undergo multiple treatment courses, understanding the variability in dose accumulation methodologies across institutions has become essential. This national project aimed to evaluate current re-irradiation dose accumulation practices in Ireland and to quantify inter-centre variations in reported OAR doses. Material/Methods: Two anonymised lung cancer patient datasets, each comprising two treatment courses, were distributed to ten radiotherapy centres nationwide. Participating centres were asked to perform image registration, biological dose conversion, and dose accumulation according to their local clinical protocols. Each centre reported the cumulative dose to selected OARs, specifying the registration technique, α / β ratios, and dose summation method used. The compiled results were analysed to assess inter-centre variability and identify methodological differences. Results: Variation was observed in the reported cumulative doses (figure 1), particularly for structures affected by anatomical distortion such as the airway (55.3-104.1 Gy), where EQD2 values ranged widely between centres. Smaller and more geometrically stable organs, such as the oesophagus and brachial plexus, showed improved agreement (49.4±4.2 Gy). The majority of centres (90%, n=10) employed rigid image registration (RIR), while a smaller subset used deformable registration (DIR) or hybrid approaches. Differences were also noted in the use of EQD2 (60%) versus BED (40%), α / β ratio selection, and the application of planning risk volumes (PRVs). Two centres incorporated tissue recovery factor (TRF) when
center of the brain. The two non-dosimetric targets are placed farther than 5 cm from the center. This was chosen to allow for single- or multi-isocenter approaches to be used. X-ray based image-guidance capability was created using two PBT inserts embedded in the forehead and occipital regions of the head, mimicking skull bone.
Made with FlippingBook - Share PDF online