ESTRO 2025 - Abstract Book

S3475

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

ESTRO 2025

1839

Digital Poster Is there added value to discrete proton arc therapy? Wens Kong 1 , Merle Huiskes 2 , Steven J.M. Habraken 2 , Eleftheria Astreinidou 2 , Coen R.N. Rasch 2 , Ben J.M. Heijmen 1 , Sebastiaan Breedveld 1 1 Department of Radiotherapy, Erasmus MC Cancer institute, Erasmus University Medical Center, Rotterdam, Netherlands. 2 Department of Radiation Oncology, Leiden University Medical Center, Leiden, Netherlands Purpose/Objective: Dynamic proton arc therapy (PAT) still has significant implementation and QA issues [1]. Discrete (”step-and-shoot”) proton arc therapy (PAT) has been proposed as an alternative [2,3]. We developed a novel system for fully automated multi-criterial optimisation of discrete PAT plans with 36 equi-angular fields (PAT-36). Next, discrete PAT 36 plans were compared to automatically generated IMPT plans with 4-10 discrete beam angles in terms of dosimetric quality and treatment time to assess possible added value of discrete PAT. Material/Methods: Starting point for a PAT-36 plan generation was a large candidate spot set divided over a large number of energy layers (both patient-dependent) in 36 fields. The final plan was then generated in a two-step process: (1) preselection of important energy layers through group-sparsity regularisation in energy layers, and (2) generation of the final PAT-36 plan using sparsity-induced spot selection and wish-list driven automated multi-criteria plan generation[4-6]. To assess the added value of PAT-36, for 10 oropharyngeal patients, comparisons were made with plans with 4- and 6-beam clinical beam class solutions (CLIN-4 and CLIN-6) and with 6-, 8-, and 10-beam IMPT plans with individualised beam set-ups (BAO-6, BAO-8, BAO-10), generated with a recently proposed algorithm for beam angle optimisation. All final PAT-36, CLIN-x and BAO-y plans were automatically generated with the same wish-list, avoiding planner bias. Before analyses, all plans were normalised to the same target coverage. To investigate treatment time differences, the total numbers of energy layers and monitor units (MU) were assessed. Results: All PAT-36, CLIN-x and BAO-y plans met all institutional hard constraints. Figure 1a shows that PAT-36 clearly outperformed CLIN-4, CLIN-6 and BAO-6/8 plans, while NTCPs for BAO-10 were similar. Equivalence between PAT-36 and BAO-10 is confirmed in Figure 1b, although BAO-10 had significantly higher D2% in brainstem and spinal cord, but always within constraints. On average, 394 and 375 energy layers were used in PAT-36 and BAO-10, but BAO-10 used slightly more MUs (Figure 2a) and 12% more integral dose. Figure 2b shows the large spread in PAT-36 of energy layers and MUs over the 36 gantry angles, compared to BAO-10.