S3025
Invited Speaker
ESTRO 2026
5392 Modelling RBE for clinical applications Chiara La Tessa Radiation Oncology, University of Miami, Miami, USA Relative biological effectiveness (RBE) links the physical dose delivered in radiotherapy to its biological impact on both tumor and normal tissues. Uncertainties in RBE estimation directly affect treatment efficacy and the risk of complications, making accurate modelling a central challenge in modern particle therapy. While carbon therapy inherently requires variable RBE models for treatment planning, proton therapy has traditionally relied on a constant RBE of 1.1, under the assumption that biological variability is limited. However, increasing preclinical and clinical evidence has demonstrated that RBE varies significantly with dose, tissue type, linear energy transfer (LET), and biological context. This has raised concerns, particularly for normal tissue toxicity, and highlighted the need for more advanced and clinically robust RBE modelling approaches in proton therapy. In this talk, I will focus on two complementary directions to address current challenges in RBE modelling. I will present two approaches that extend beyond conventional RBE models. The first is the Generalized Stochastic Microdosimetry Model (GSM 2 ) 1,2 , which relies on an advanced microdosimetry-based description of radiation to capture fine-scale variations in energy deposition that are directly relevant to biological damage. GSM 2 structure is designed to preserve this level of detail in outcome prediction, enabling the characterization of high-resolution RBE variations in both proton and carbon ion therapy. I will also present a machine and deep learning–based approach, the Artificial iNtelligence bAsed model for (radiation- induced) cell KIlliNg prediction (ANAKIN) 3 , trained on more than 500 cell survival experiments across multiple radiation types. Second, I will present a machine learning-based alternative methodology for modelling normal tissue effects that does not rely on RBE 4 . This approach directly relates physical descriptors of the radiation field to clinical outcomes, thereby bypassing some of the intrinsic uncertainties and assumptions embedded in RBE-based formulations. RBE inherently relies on referencing biological effects to photons through dose scaling, implicitly assuming a common biological baseline. However, I will present clinical evidence across multiple normal tissue toxicities indicating the limitations of this assumption. By avoiding this normalization, the proposed framework provides a more direct and potentially more robust pathway toward clinically translatable models for predicting toxicity in particle therapy.
Together, the presented approaches suggest a shift beyond conventional RBE paradigms toward more physically and clinically grounded modelling
frameworks. References:
1. Cordoni, F. G. et al. Cell Survival Computation via the Generalized Stochastic Microdosimetric Model (GSM2); Part I: The Theoretical Framework. Radiat. Res. 197, 218–232 (2022). 2. Missiaggia, M. et al. Cell Survival Computation via the Generalized Stochastic Microdosimetric Model (GSM2); Part II: Numerical Results. Radiat. Res. 201, (2024). 3. Cordoni, F. G. et al. An artificial intelligence-based model for cell killing prediction: development, validation and explainability analysis of the ANAKIN model. Phys. Med. Biol. 68, 085017 (2023). 4. Cartechini, G. et al. Proton-specific dose and radiation quality constraints to reduce acute oral mucositis in head and neck cancer patients. Radiother. Oncol. 219, 111522 (2026). 5397 Can we cure and preserve the organ without cutting? Current state-of-the-art Peter S.N. van Rossum Radiation Oncology, Amsterdam UMC, Amsterdam, Netherlands This presentation will address the conditions under which organ-preserving treatment can be considered a responsible alternative to routine surgery in esophageal and rectal cancer. Organ preservation is defined as a curative strategy in which surgery is avoided, postponed, or selectively reserved, with the aim of maintaining organ function and quality of life without unacceptable loss of oncological safety. The presentation will argue that organ preservation will require more than a complete clinical response alone. Key prerequisites will include appropriate patient selection, a realistic likelihood of major or complete response, accurate response assessment, a feasible and intensive surveillance protocol, the availability of safe salvage surgery, and a meaningful expected benefit in function and quality of life. Differences between rectal and esophageal cancer will be highlighted. In rectal cancer, watch-and-wait strategies will be presented as relatively well supported by response evaluation tools and by favorable functional and patient-reported outcomes, particularly after total neoadjuvant treatment. In esophageal cancer, the presentation will show that response assessment after chemoradiotherapy will remain more challenging, which will complicate safe implementation of surveillance approaches. Available evidence on active surveillance after neoadjuvant
Made with FlippingBook - Share PDF online