ESTRO 2025 - Abstract Book

S3753

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2025

Conclusion: This study demonstrated the feasibility to computationally quantify dose to circulating blood in 19 AI-segmented organs. Heart, lungs, aorta and superior-vena-cava were found as the major contributing organs to the total blood dose. Furthermore, this work has shown that the blood dose is weakly, but significantly correlated with the during treatment neutrophil-to-lymphocyte ratio. As neutrophil-to-lymphocyte ratio has frequently been associated with poorer survival, understanding its correlation with dose to blood could offer valuable insights for developing immune sparing radiotherapy.

Keywords: radiation dose to blood, NSCLC, modelling

References: [1] Beekman C, et al. A stochastic model of blood flow to calculate blood dose during radiotherapy. Phys Med Biol. 2023;68(22):10.1088/1361-6560/ad02d6 [2] Shin J, Xing S, McCullum L, et al. HEDOS-a computational tool to assess radiation dose to circulating blood cells during external beam radiotherapy based on whole-body blood flow simulations. Phys Med Biol. 2021;66(16):10.1088/1361-6560/ac16ea.

1663

Digital Poster Predicting Tumor Voxel Dose-Response of Head and Neck Cancer Using Deep Learning: Impact of FDG-PET Imaging Feedback Timing and HPV Status Shupeng Chen 1 , Rohan Deraniyagala 2 1 Radiation Oncology, Emory University, Atlanta, USA. 2 Radiation Oncology, Corewell Health, Royal Oak, USA Purpose/Objective: Early and accurate prediction of tumor voxel dose-response matrix (DRM) during radiotherapy is crucial for effective adaptive treatment management in head and neck squamous cell carcinoma (HNSCC). However, this is challenging due to the dynamic fluctuations of FDG uptake in tumor cells and the considerable PET image noise. This study investigated the feasibility of using a deep learning (DL) model to predict DRM during the early weeks of treatment