ESTRO 2025 - Abstract Book

S2151

Interdisciplinary – Education in radiation oncology

ESTRO 2025

Conclusion: LLMs have the potential to be effective tools for pretreatment education in pediatric radiation oncology. However, only advanced models such as GPT-4 provide information comparable to that of a radiation oncologist, though occasional low-quality responses still occur. Caution should be exercised when using GPT-3.5 models, as they are more prone to providing irrelevant answers to patient questions.

Keywords: Pediatric radiation oncology, LLMs, Education

References: 1. Ogasawara R, Katsumata N, Toyooka T, Akaishi Y, Yokoyama T, Kadokura G. Reliability of Cancer Treatment Information on the Internet: Observational Study. JMIR Cancer 2018 Dec 17;4(2):e10031 2. Clusmann J, Kolbinger FR, Muti HS, Carrero ZI, Eckardt J-N, Laleh NG, Löffler CML, Schwarzkopf S-C, Unger M, Veldhuizen GP, Wagner SJ, Kather JN. The future landscape of large language models in medicine. Commun Med 2023 Oct 10;3(1):141 3. Ayers JW, Poliak A, Dredze M, Leas EC, Zhu Z, Kelley JB, Faix DJ, Goodman AM, Longhurst CA, Hogarth M, Smith DM. Comparing Physician and Artificial Intelligence Chatbot Responses to Patient Questions Posted to a Public Social Media Forum. JAMA Intern Med 2023 Apr 28

173

Digital Poster Thermal modeling of ultra-high dose rate radiotherapy. Aashini Rajpal, Julie Colnot, Massinissa Abbas, Eric Deutsch, Charlotte Robert UMR 1030 - Molecular Radiotherapy and Therapeutic Innovation, Gustave Roussy, Villejuif, France Purpose/Objective: Ultra-high dose rate radiotherapy (UHDR-RT) has emerged as a promising modality, [1,2] demonstrating a differential response in irradiated healthy versus cancerous tissues compared to conventional dose rate RT (CONV RT). Given the distinct temporal structures of CONV-RT and UHDR-RT, there is potential for these techniques to produce different thermal responses during radiation-matter interactions. In fact, ionizing radiation delivers significant energy to target molecules, which can result in the molecule’s Coulombic explosion, with the atoms having high kinetic energies (which can be indicative of temperature change based on the Maxwell-Boltzmann distribution) , known as "hot-atom chemistry". [3] This work aims to study temperature variations in tissues when irradiated with UHDR, an underexplored area, [4] in an attempt to understand the observed benefits of UHDR-RT, through in-silico studies. Material/Methods: The Bioheat transfer module of COMSOL Multiphysics which utilizes Pennes’ Bioheat equation, was used to simulate heat diffusion within target tissues (skin and prostate) irradiated by a FLASHKNIFE system (THERYQ, Rousset, France) operating at 60 mA. A 3D box representing biological tissue was modeled, to assess heat distribution following irradiation with a UHDR electron beam of radius 5 cm (Pulse width = 2 µs with a repetition rate of 100 Hz, Energies = 6 MeV and 10 MeV, Doses studied = 11 Gy and 21 Gy). The input data for energy deposition in the tissues (at 1 mm depth), to perform the heat transfer modeling, was obtained from the experimental dose data or the GATE simulations. The temperature changes were recorded at each time step (1 µs) during the simulation. Herein, the blood perfusion rate was kept constant with temperature and the thermal damage (irreversible molecular and structural damage) was estimated using the Arrhenius equation. [5]