Click here to

Session: Radiobiology of Particle Therapy [Return to Session]

Comparison of Variable Proton RBE Model Predictions in the Treatment Plans of Patients Who Developed Radiation-Induced Necrosis

D Flint1*, T Zhang2, P Yepes1,3, A Adair1, Q Wang1,3, C Ruff1, G Engeseth1,4,5, B Gunn1, S Shaitelman1, G Sawakuchi1, (1) The University of MD Anderson Cancer Center, Houston, TX, (2) Hofstra University/Northwell Health, Hempstead, NY, (3) Rice University, Houston, TX, (4) Haukeland University Hospital, Bergen, Norway, (5) University of Bergen, Bergen, Norway.


TH-D-202-2 (Thursday, 7/14/2022) 11:00 AM - 12:00 PM [Eastern Time (GMT-4)]

Room 202

Purpose: To compare the predictions of different variable proton relative biological effectiveness (RBE) models in patient treatment plans that were associated with clinical findings of radiation-induced necrosis.

Methods: A cohort of 13 head and neck patients treated with intensity-modulated proton therapy (IMPT) to a dose of 70 Gy(RBE) in 33 fractions were selected. All patients developed radiation-induced necrosis near the treatment site, identified in T1-post-contrast MR scans. Contours from the MR scans were transferred to the treatment planning CTs to delineate the necrotic regions within the CT datasets. We then performed Monte Carlo calculations on the CT datasets using a validated GPU-implemented track-repeating algorithm to score the dose and linear energy transfer (LET) distributions associated with each patient’s treatment. From the voxelized dose and LET distributions, we calculated the RBE-weighted dose distributions according to 3 variable-RBE models: the models by McNamara (2015), Wedenberg (2013), and our model (under review), which we compared to the clinically used RBE=1.1.

Results: For all variable-RBE models, the RBE-weighted mean (Dmean) and maximum (Dmax) doses to the necrotic regions were significantly higher than the treatment planning assumption of a fixed RBE=1.1. Among them, our model predicted the highest RBE-weighted doses, which were considerably higher than the planned dose to the tumor (Dmax = 87.9±1.3 Gy(RBE); Dmean = 78.7±1.9 Gy(RBE)). We also found that the variable RBE-weighted doses to critical structures in the treatment field often exceeded the treatment planning constraints: e.g. for our model, 100% of patients received Dmax>54 Gy(RBE) to the brainstem, 91% of patients received Dmax>54 Gy(RBE) to the optic chiasm, and 88% of patients received Dmax>70Gy(RBE) to the mandible.

Conclusion: Assuming a fixed RBE=1.1 may fail to account for RBE-weighted doses to normal tissues that are considerably higher than the planned treatment dose, and that may be associated with clinically meaningful toxicities.

Funding Support, Disclosures, and Conflict of Interest: This work was supported by the following P01 grant: 1P01CA261669-01. GOS and SFS have research funds from Alpha Tau Medical and Artios Pharma.


RBE, Protons, Modeling


TH- External Beam- Particle/high LET therapy: Proton therapy – dose optimization

Contact Email