S2604
Radiobiology - Radiobiological modelling
ESTRO 2026
Purpose/Objective: Radiotherapy is a standard treatment for solid tumors, but its potency is limited by tumor radioresistent hypoxic subvolumes. The classical linear-quadratic (LQ) model provides a static dose-response relationship, estimating cell survival based on delivered dose but without accounting for biological changes occurring over the course of treatment. To overcome these limitations, Jeong et al.[1] developed a mechanistic, time-resolved LQ-based model that incorporates resource competition and dynamic recompartmentalization, improving Tumor Control Probability (TCP) prediction across fractionation regimens. However, preclinical data reveal that once reoxygenated, hypoxic cells remain quiescent and radioresistant for several weeks [2]. We extended Jeong’s model with a quiescent compartment to test whether the biologically grounded model can accurately predict TCP, potentially guiding strategies to target resistant subpopulations. Material/Methods: The original model divides the tumor into three compartments based on oxygen and glucose availability. We extended it by adding a quiescent sub-compartment within the proliferative one (Figure 1), representing reoxygenated, metabolically reduced, and radioresistant cells, assigned a specific oxygen enhancement ratio (OER_Q). Parameters most influencing the model fit were identified by sensitivity analysis and optimized via maximum log-likelihood. Both models were benchmarked on 38 cohorts (2,701 patients) with non-small cell lung cancer (NSCLC) treated with conventional, SBRT, and single-fraction SBRT regimens. The goodness of fit was evaluated using a chi-squared ( χ² ) test and model performance was further evaluated using the Pearson's correlation coefficient (r) and root mean square error (RMSE) between observed and predicted TCP values. Results: Compared to the original model, both normalized to a 2 Gy per fraction response (EQD2) and using identical parameters, the extended model reproduced overall tumor control trends but underestimated cell kill, requiring on average +8 Gy for equivalent control. Sensitivity analysis identified the radiosensitivity parameters α , β , and OER_Q as the most influential on TCP dynamics, and optimization yielded α = 0.45 Gy ⁻ ¹ , β = 0.45 Gy ⁻ ² , and OER_Q = 1.33. Using these optimized parameters, the global fit resulted in a TD ₅₀ of 60.6 Gy ( χ² = 28.25, p = 0.85), accurately reproducing observed TCP across regimens but with slightly reduced predictive accuracy (r = 0.84 vs 0.91; RMSE = 0.116 vs 0.084), (Figure 2). Conclusion: The extended model captures tumor control trends across all regimens while incorporating quiescent cell behavior, offering a more biologically realistic framework. However, further parameter refinement and optimization are needed to improve predictive accuracy and fully exploit its potential for modeling treatment resistance and guiding adaptive radiotherapy strategies.
solely from hydroxyl radicals (9 SSBs). Under cumulative accumulation, SSBs arise not only from hydroxyl radicals alone but also from other radical species, including hydrated electrons and hydrogen atoms, resulting in 21 SSBs.
Conclusion: This study demonstrates that under ultra high dose rate conditions, the instantaneous accumulation of radicals during the chemical stage directly influences DNA damage. When radicals from multiple irradiation events coexist, radical-radical interactions increase and enhance self-quenching, but the higher instantaneous radical density still promotes radical-DNA encounters. Under cumulative accumulation, hydroxyl radicals, hydrated electrons and hydrogen atoms all contribute to DNA single-strand breaks. Overall, DNA chemical damage in high radical-density environments is driven by radical clustering and a more complex reaction network. References: [1] Chatzipapas, K.P., Tran, N.H., Dordevic, M., Zivkovic, S., Zein, S., Shin, W.G., Sakata, D., Lampe, N., Brown, J.M., Ristic - Fira, A. and Petrovic, I., 2023. Simulation of DNA damage using Geant4 - DNA: an overview of the “molecularDNA” example application. Precision Radiation Oncology, 7(1), pp.4-14. Keywords: Instantaneous radical density, UHDR, DNA breaks Digital Poster 3939 Extending mechanistic models of tumor response : Integrating quiescence into Radiotherapy dynamics Lara Wahbi 1 , Apostolos Menegakis 1 , Jeho Jeong 2 , Joseph O. Deasy 2 , Astrid van der Horst 1 , Zeno A. R. Gouw 1 , Jan-Jakob Sonke 1 1 Radiation Oncology, The Netherlands cancer Institute, Amsterdam, Netherlands. 2 Medical Physics, Memorial Sloan Kettering Cancer Center, New York, USA
Made with FlippingBook - Share PDF online