ESTRO 2025 - Abstract Book

S3435

Physics - Machine learning models and clinical applications

ESTRO 2025

Conclusion: The combined ERI-dose model for pCR prediction was successfully validated on an external cohort treated with distinct RT regimens. Specifically, a dose escalation of 8.5 Gy in low-to-intermediate responders corresponded to a 21% increase in pCR, while, as predicted by the model, no benefit was observed in the non-responder group. These findings highlight the model's potential to guide personalised RT protocols, enabling dose escalation based on ERI to optimise individual pCR outcomes.

Keywords: personalised RT, Adaptive RT, pCR

References: [1] Ciccheti et al, Quantifying individual dose relationships for pCR in rectal cancer: potentials for a customized dose, Rad. Onc. DOI:10.1016/S0167-8140(22)02446-X [2] Hall et al, Effect of increasing radiation dose on pathologic complete response in rectal cancer patients treated with neoadjuvant chemoradiation therapy. Acta Onc. https://doi.org/10.1080/0284186X.2016.1235797 [3] Broggi et al, Predicting pathological response after radio-chemotherapy for rectal cancer: Impact of late oxaliplatin administration, Rad.Onc. DOI:10.1016/j.radonc.2020.05.019 [4] Vergouwe et al, A closed testing procedure to select an appropriate method for updating prediction models. Stat in Med. https://doi.org/10.1002/sim.7179

3811

Digital Poster Cell survival predictions using machine learning and Monte Carlo simulations Maria Pia Valenzuela, Sophia Galvez, Jorge Jara, Sebastian Martinez, Andrea Russomando Physics, Pontificia Universidad Catolica de Chile, Santiago, Chile

Purpose/Objective: Particle therapy is a powerful cancer treatment, but understanding its impact at the biological level is essential for optimizing its effectiveness. Predicting the effects of radiation, which depend on factors such as LET and the