ESTRO 2024 - Abstract Book

S4616

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

ESTR0 2024

Figure 2: Comparison of normal tissue complication probabilities (NTCP) calculated from a variable RBE-weighted dose based on the LETd distribution obtained from Monte Carlo (MC) simulation and the deep learning (DL) model. The dashed black line indicates the identity function for perfect correspondence. Results for acute ocular toxicity (left) and memory impairment (right) are shown.

Conclusion:

In this study, a 3D CNN was developed that successfully predicts the proton LETd based on the planned dose distribution. Observed deviations in the LETd had little impact on the clinically more relevant prediction of treatment related side effects. We therefore conclude that DL models can approximate LETd distributions from MC simulations for the clinically relevant task of side effect prediction. The developed 3D CNN can make a significant contribution to the investigation of the variable RBE of protons, particularly for retrospective patient data and for clinics without the capability of MC-based LET calculation.

Keywords: linear energy transfer,CNN

References:

[1] Paganetti, H. "Relative biological effectiveness (RBE) values for proton beam therapy. Variations as a function of biological endpoint, dose, and linear energy transfer." Physics in Medicine & Biology 59.22 (2014): R419.

[2] Eulitz, J., et al. "Increased relative biological effectiveness and periventricular radiosensitivity in proton therapy of glioma patients." Radiotherapy and Oncology 178 (2023): 109422.

[3] Ronneberger, O., et al. “U-net: Convolutional networks for biomedical image segmentation” MICCAI 2015, 2015

[4] Myronenko, A. "3D MRI brain tumor segmentation using autoencoder regularization." BrainLes 2018, 2018.

[5] Hatamizadeh, A., et al. "Unetr: Transformers for 3d medical image segmentation." IEEE/CVF, 2022.

[6] Bahn, E., et al. "Late contrast enhancing brain lesions in proton-treated patients with low-grade glioma: clinical evidence for increased periventricular sensitivity and variable RBE." International Journal of Radiation Oncology* Biology* Physics 107.3 (2020): 571-578.