S1963
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
PMC11701999. Keywords: Re-irradiation, NTCP, BED_EUD
Gy and 78.3 Gy, respectively. Figure 1 and 2 show that for Pmax=40%, NTCP-guided plans achieved target coverage comparable to EQD2-based planning. They also yielded reduced high-dose rectal volumes (e.g., V122Gy) and only modest increase in bladder dose spill, with statistically significance observed solely for V140Gy. In contrast at Pmax=20%, SBRT plans could not meet sufficient target coverage criteria.
Digital Poster 3931 Formulating inverse HDR brachytherapy treatment planning as a Binary Quadratic Model: a foundation for quantum optimization Hannes A Loebner 1 , Lars Meuser 2 , Silvan Mueller 1 , Julien Ott 1 , Michael K Fix 1 , Peter Manser 1 1 Division of Medical Radiation Physics and Department of Radiation Oncology, Inselspital, Bern University Hospital, and University of Bern, Bern, Switzerland. 2 Department of Physics, University of Milano-Bicocca, Milan, Italy Purpose/Objective: Optimization in high-dose-rate (HDR) brachytherapy is inherently combinatorial, since discrete dwell-times at discrete dwell-positions must simultaneously satisfy dose guidance for target and organs-of-interest (OOIs). Conventional treatment planning systems approximate this optimization problem as continuous and apply gradient descent (GD) or heuristic methods to solve it, thereby ignoring its discrete nature. Therefore, the aim of this work is to re-formulate this optimization problem as a Binary Quadratic Model (BQM) to explicitly capture the discrete nature of dwell positions and times while encoding minimum and maximum dose guidance [1]. As a first proof-of- concept study, this work provides a direct pathway to solution on classical and quantum-inspired optimizers, and ultimately on real quantum annealers [2], potentially enabling a substantial reduction of A simplified 2-dimensional space HDR-brachytherapy optimization problem was implemented (1 target, 2 OOIs, 36 dwell-positions). Inverse-square-law and point-source description were used to approximate dose deposition (fig.1, top). A Quadratic Unconstrained Binary Optimization (QUBO) model was constructed, encoding target minimum, and target and OOI maximum dose guidance into quadratic penalty terms [1]. The resulting BQM was solved with four optimization algorithms: GD (in-house), simulated annealing (SA, D-wave [3]), Tabu (D-wave [4]), and simulated quantum annealing (SQA, OpenJij [5]). Bit- depth for dwell-time encoding was varied from 1 to 8 bits, enabling dwell-time steps from s to ms, respectively. Optimization cost, dose statistics, and wall-clock time were recorded. optimization time. Material/Methods:
Conclusion: The NTCP-based framework enables conservative, biologically consistent re-irradiation planning without anatomical co-registration. The method may rapidly predict dose distributions serving as a practical decision-support tool in multidisciplinary discussions on side-effect risks. References: Brand DH, Brüningk SC, Wilkins A et al. CHHiP Trial Management Group. The Fraction Size Sensitivity of Late Genitourinary Toxicity: Analysis of Alpha/Beta ( α / β ) Ratios in the CHHiP Trial. Int J Radiat Oncol Biol Phys. 2023 Feb 1;115(2):327-336. doi: 10.1016/j.ijrobp.2022.08.030. Epub 2022 Aug 17. PMID: 35985457.Jongen CAM, Heijmen BJM, Schillemans et al. Normal tissue complication probability modeling for late rectal bleeding after conventional or hypofractionated radiotherapy for prostate cancer. Clin Transl Radiat Oncol. 2024 Nov 10;50:100886. doi: 10.1016/j.ctro.2024.100886. PMID: 39763489; PMCID:
Made with FlippingBook - Share PDF online