ESTRO 2025 - Abstract Book

S3531

Physics - Optimisation, algorithms and applications for ion beam treatment planning

ESTRO 2025

address the interplay effect. However, this approach is statistically and scenario-based limited and may result in conservative plans regarding robust CTV dose. We aimed: (i) to introduce a 4D probabilistic robustness evaluation (4DPRE) to comprehensively assess plan robustness against breathing motion, geometrical and range errors and (ii) to probabilistically evaluate plan robustness for 17 EC patients. Material/Methods: Clinical IMPT plans for 17 EC patients were included. Patients were treated to 41.4 (14/20) and 50.4 GyRBE (3/20), in 1.8 GyRBE/fraction, planned on the average 4DCT-scan (Avg-4DpCT). Patients with >5mm target motion received layered rescanning (x4). Dose was prescribed (D pres ) to the VWmin dose and evaluated at the Avg-4DpCT, In-4DpCT and Ex-4DpCT:VWmin-D 98%,CTV ≥0.95D pres . The 4DPRE simulated the dose delivery over the 4DCT phases using 30 breathing signals and a fast patient-specific dose engine model using Polynomial Chaos Expansion. It considered geometrical random errors (σ=2.5mm:1SD) and breathing signals per fraction and systematic geometrical (Σ=2.5mm:1SD) and range errors (ρ=1.5%:1SD) for all fractions, followed by a 4D dose accumulation (Figure 1). Per patient, we performed a probabilistic evaluation of 10.000 completed fractionated treatments, enabling DVH probability distributions for the CTV and the heart. Plan robustness and toxicity was evaluated using a probabilistic CTV goal: D 99.8%,CTV ≥0.95D pres in 90% of the treatments [1] and a 2-year mortality probability (2YM-NTCP) [2].