ESTRO 2025 - Abstract Book

S4402

RTT - Treatment planning, OAR and target definitions

ESTRO 2025

segmentation was compared to the expert contours by calculating the volume ratio, Dice similarity coefficient (DSC), and 95% Hausdorff distance (95%HD). A higher DSC (closer to 1) and lower 95%HD indicate better segmentation accuracy. Two experienced radiographers were asked to rate the AI segmentation using a Likert scale ranging from 1 (unusable) to 5 (clinically acceptable) and to provide feedback on the segmentations [2]. Results: The mean DSC for all structures was 0.56±0.26, and was the highest for the brainstem (0.81±0.08) and lowest for the right optic nerve (0.18±0.14). The mean 95%HD was 12.07±16.68 mm and was the highest for the mandible (57.24±3.15 mm) and lowest for the optic chiasm (4.54±3.39 mm). Compared to the gold standard segmentations, the mean AI segmented volume was smaller for all OARs except for the left parotid gland and brainstem. However, the difference was only statistically significant for the left parotid gland and optic nerves (p<0.05). The mean Likert rating provided by the radiographers was 2.47±0.68, with the highest for the right parotid gland (3.10±0.21) and the lowest for the optic chiasm (1.36±0.48). Under-contouring of the mandible and unusable optic chiasm contours were key issues highlighted by the radiographers. Conclusion: The accuracy of the AI segmentations varied, thus highlighting the need for expert editing particularly for complex OARs. References: [1] Guzene L, Beddok A, Nioche C, Modzelewski R, Loiseau C, Salleron J, et al. Assessing Interobserver Variability in the Delineation of Structures in Radiation Oncology: A Systematic Review. International Journal of Radiation Oncology*Biology*Physics 2023;115:1047–60. https://doi.org/10.1016/J.IJROBP.2022.11.021. [2] Baroudi H, Brock KK, Cao W, Chen X, Chung C, Court LE, et al. Automated Contouring and Planning in Radiation Therapy: What Is ‘Clinically Acceptable’? Diagnostics 2023, Vol 13, Page 667 2023;13:667. https://doi.org/10.3390/DIAGNOSTICS13040667. Keywords: Artificial Intelligence, OARs, contouring Digital Poster PARTIAL BREAST IRRADIATION: FREE BREATHING VERSUS DEEP INSPIRATORY BREATH HOLD Patrícia Varzim 1 , Pedro Meireles 1 , João Casalta Lopes 1,2,3 , Daniela Saraiva 1 , Patrícia Peixoto 1 , Filipa Fonseca 1 , Daniela Branco 1 , Diana Alves 1 , Raquel Rocha 1 , Fernando Costa 1 , Armanda Monteiro 1 , Lígia Osório 1 1 Radiotherapy, Unidade Local de Saúde de São João, Porto, Portugal. 2 Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal. 3 H&TRC—Health & Technology Research Center, Coimbra Health School, Polytechnic University of Coimbra, Coimbra, Portugal Purpose/Objective: Radiotherapy is key in breast cancer treatment, reducing recurrence and improving survival. The ultra hypofractionation regimen (26Gy in 5 fractions) shortens treatment time while maintaining effectiveness. Partial Breast Irradiation (PBI) targets the tumor bed with higher doses, sparing healthy tissue. The Deep Inspiratory Breath Hold (DIBH) technique minimizes radiation to critical organs by increasing their distance from the irradiated area. This study compares Free Breathing (FB) and DIBH for PBI using 3D conformal radiotherapy (3DCRT). Material/Methods: Breast cancer patients treated with PBI (26Gy/5 fractions/1 week) between January 2022 and September 2024 were included. CT scans in FB and DIBH were used to create 3DCRT plans. Dosimetric parameters analyzed included V95 and V105 (target), V5, V25, and mean dose (heart), and V30 and mean dose (lung). Data were summarized as mean 4217