ESTRO 2025 - Abstract Book

S3406

Physics - Machine learning models and clinical applications

ESTRO 2025

Conclusion: Results demonstrate that LLMs can potentially eliminate or reduce the time-consuming process of manually renaming OAR structures, enabling more efficient data sharing across institutions and countries. The use of a local and pre-trained open-source/open-weights LLM shows promise for rapid clinical application even with confidential patient information.

Keywords: large language models (LLMs), structure renaming

References: [1] Mayo, CS. et al. (2018). American Association of Physicists in Medicine Task Group 263: Standardizing Nomenclatures in Radiation Oncology. International Journal of Radiation Oncology, Biology, Physics, 100 (4), 1057 – 1066, DOI: 10.1016/j.ijrobp.2017.12.013 [2] Dubey, A. et al. (2024). The llama 3 herd of models. arXiv preprint arXiv:2407.21783.

2809

Digital Poster Artificial intelligence for in-vivo dosimetry using EPID in external beam photon radiotherapy Lorenzo Marini 1,2 , Carlotta Mozzi 3 , Aafke Kraan 2 , Francesca Lizzi 2 , Michele Avanzo 4 , Alessandra Retico 2 , Cinzia Talamonti 3,5 1 Department of Computer Science, University of Pisa, Pisa, Italy. 2 Istituto Nazionale di Fisica Nucleare, Istituto Nazionale di Fisica Nucleare, Pisa, Italy. 3 Department of Experimental and Clinical Biomedical Sciences “Mario Serio”, University of Florence, Florence, Italy. 4 Centro di Riferimento Oncologico (CRO) di Aviano, Centro di Riferimento