ESTRO 2024 - Abstract Book

S344

Beachytherapy - Physics

ESTRO 2024

Prostate and BRIGHT plans. We also find that none of the clinically used plans from Oncentra Prostate achieve all sparing aims, whereas only one of the selected plans generated in BRIGHT marginally missed the sparing aims.

Applying the F-test we find for the bladder and urethra, a significant (p-values <0.002) reduction in dose when using BRIGHT, see Figure 2. It is also found that the bladder receives a maximum dose below 200% in all cases. Regarding the rectum for the BRIGHT plans, we see an increase in the mean D2cm3 from 5.3 Gy to 5.6 Gy with a p-value of 0.0016.

Conclusion:

All plans selected from the manifold found by BRIGHT were found to be clinically acceptable. Compared to the clinical plans, a significant reduction in dose to the bladder and urethra is obtained without reducing coverage of the prostate. This comes with a moderate increase in dose to the rectum that does not exceed the clinical aims. We can therefore conclude that BRIGHT is well-capable of finding high-quality plans in different clinical settings, opening the door to a more widespread clinical introduction.

Keywords: HDR, Treatment Optimization.

References:

1. Bouter, A. et al. GPU-accelerated bi-objective treatment planning for prostate high-dose-rate brachytherapy. Med Phys 46, (2019).

2. Barten, D. L. J. et al. Towards artificial intelligence-based automated treatment planning in clinical practice: A prospective study of the first clinical experiences in high-dose-rate prostate brachytherapy. Brachytherapy 22, (2023). 3. Bouter, A., Luong, N. H., Witteveen, C., Alderliesten, T. & Bosman, P. A. N. The multi-objective real-valued gene pool optimal mixing evolutionary algorithm. in GECCO 2017 - Proceedings of the 2017 Genetic and Evolutionary Computation Conference (2017). doi:10.1145/3071178.3071274.

1589

Proffered Paper

Finding the best options to spare organs at risk with AI-based prostate HDR brachytherapy planning

Renzo J. Scholman 1,2 , Danique L.J. Barten 3 , Bradley R. Pieters 3 , Peter A.N. Bosman 1,2 , Tanja Alderliesten 4

1 Centrum Wiskunde & Informatica, Evolutionary Intelligence, Amsterdam, Netherlands. 2 Delft University of Technology, Software Technology, Delft, Netherlands. 3 Amsterdam University Medical Centers / University of Amsterdam, Radiation Oncology, Amsterdam, Netherlands. 4 Leiden University Medical Center, Radiation Oncology, Leiden, Netherlands

Purpose/Objective: