S1588
Physics - Autosegmentation
ESTRO 2026
Conclusion: We found TotalSegmentator performed well on geometric assessment for several cardiac substructures including the right atrium. Dosimetric assessment indicated AI-contours yield comparable right atrial dose estimates to those obtained from GT- contours. We plan to evaluate AI-contours prospectively within the radiotherapy workflow. Additional cardiac substructures, including the rest of the CAA structures – not currently available within TotalSegmentator – will be incorporated into our institutional autocontouring model development pipeline. References: 1. McWilliam A, et al. Radiation dose to heart base linked with poorer survival in lung cancer patients. Eur J Cancer. 2017 Nov;85:106-113.2. Marchant T, et al. Dosimetric impact of sparing base of heart on organ at risk doses during lung radiotherapy. Radiother Oncol. 2025 Jan;202:110654.3. Ntentas G, et al. Proton Therapy in Supradiaphragmatic Lymphoma: Predicting Treatment-Related Mortality to Help Optimize Patient Selection. Int J Radiat Oncol Biol Phys. 2022 Mar 15;112(4):913-925.4. Duane F, et al. A cardiac contouring atlas for radiotherapy. Radiother Oncol. 2017 Mar;122(3):416-422. Keywords: Cardiac substructures, autocontouring, open-source
Digital Poster Highlight 3802
Developing an in-house target AI contouring model for craniospinal patients – from idea to routine practice Samuel Ingram 1,2 , Peter Sitch 3 , Gillian Whitfield 4,2 , Shermaine Pan 4,2 , Gabrielle Testa 3 , Nicky Thorp 4 , Love Goyal 2,5 , Abiola Fatimilehin 4 , Khalid Abutaleb 4 , Marianne Aznar 2,5 , Matthew Clarke 1 , Matthew Lowe 3,2 1 Radiotherapy Physics, The Christie NHS Foundation Trust, Manchester, United Kingdom. 2 Division of Cancer Sciences, University of Manchester, Manchester, United Kingdom. 3 Proton Therapy Physics, The Christie NHS Foundation Trust, Manchester, United Kingdom. 4 Clinical Oncology, The Christie NHS Foundation Trust, Manchester, United Kingdom. 5 The Christie, The Christie NHS Foundation Trust, Manchester, United Kingdom Purpose/Objective: To develop and implement an in-house CTV AI contouring model for craniospinal patients. Material/Methods: We showcase how this model was used to create and implement AI AutoContouring for craniospinal target contouring in our Proton Beam Therapy service. The model was trained from 129 patient cases treated at our institution between 2019-2023 (Figure 1) using the nnU-Net v2 framework [1] and the 3D full resolution architecture. Blinded randomised qualitative review of manually drawn versus AI contours was carried out by 2 independent consultant oncologists on 18
Digital Poster 3697
Detecting Automation Bias in Radiotherapy Contouring: The Role of Similarity Measures Mark J Gooding 1,2 , Djamal Boukerroui 1 1 -, Inpictura Ltd, Abingdon, United Kingdom. 2 Division of Cancer Sciences, Faculty of Biology, Medicine and Health, The University of Manchester, Manchester, United Kingdom
Made with FlippingBook - Share PDF online