ESTRO 2024 - Abstract Book

S3696

Physics - Dose prediction, optimisation and applications of photon and electron planning

ESTRO 2024

2. Hoem I.S. Automatic VMAT planning in RayStation for locally advanced cervical cancer. Master thesis, Norwegian University of Science and Technology, Trondheim, June 2022.

3. Pötter R. et al. The EMBRACE II study: The outcome and prospect of two decades of evolution within the GEC ESTRO GYN working group and the EMBRACE studies. Clin Transl Radiat Oncol. 2018 Jan 11;9:48-60.

2658

Poster Discussion

Inverse biological radiotherapy treatment planning optimization using simulated annealing

Diana Aires 1 , Brígida Ferreira 2 , Joana Dias 3 , Humberto Rocha 4

1 Faculty of Sciences of the University of Lisbon, Physics Department, Lisbon, Portugal. 2 Faculty of Sciences of the University of Lisbon, IBEB, Lisbon, Portugal. 3 Faculty of Economics of the University of Coimbra, INESC, Coimbra, Portugal. 4 Faculty of Economics of the University of Coimbra, CeBER, Coimbra, Portugal

Purpose/Objective:

Radiobiological models can be used for radiotherapy treatment plan optimization taking advantage of the direct relation they provide between delivered dose and clinical outcome. This approach can also provide a better personalization of treatments by accounting for the radiosensitivity of the different tissues. In contrast, physical metrics focus only on achieving a desired dose distribution without providing information about the expected treatment outcome. Still, inverse physical optimization treatment planning is currently the most common optimization approach adopted in clinical practice. Biological optimization has only been recently added to commercially available treatment planning systems (TPS) and it is still primarily used as a complement to physical optimization or simply for plan evaluation.

In this work, an optimization algorithm that employs radiobiological models for inverse treatment planning optimization using simulated annealing (SA) algorithms was implemented considering IMRT treatment planning.

Material/Methods:

The optimization approach was developed in MATLAB R2020a and was integrated into the research open-source TPS matRad (Wieser et al., 2017). Treatment plan optimization is achieved by applying simulated annealing, a metaheuristic inspired by the annealing process in metallurgy aiming to maximize the biological objective function uncomplicated tumour control probability, P+. The probability of tumour control (TCP) and the probability of normal tissue complication (NTCP) were estimated using the Linear-Quadratic-Poisson model and the Relative Seriality to account for the heterogeneous dose distribution and the parameters were collected from the available literature. In this optimization strategy the optimization is done one beam at a time, meaning that in each iteration only the fluence weights of a single beam are changed. This approach makes the optimization process significantly faster and does not allow for drastic changes in fluences between neighbouring beamlets making the treatment more realistic in the sense that a deliverable plan is obtained. Another important feature of the algorithm is the use of a virtual normal tissue structure that overlaps with the tumour volumes. Its purpose is to control the maximum dose delivered to the target volume since TCP alone does not penalize these maximum values.