Impact of Proton Beam Arrangements on LETd and RBE in Organs at Risk for Nasopharyngeal Cancer 📝

Author: Mauricio Acosta, Zachary Fellows, Alonso N. Gutierrez, Noah S. Kalman, Kristen A. McConnell, Rob Rohe, Andrew J. Wroe, Sreenija A Yarlagadda 👨‍🔬

Affiliation: Miami Cancer Institute, Baptist Health South Florida 🌍

Abstract:

Purpose:
Beam selection is crucial in proton therapy to ensure optimal physical and biological dose conformity to the target while maximizing sparing of normal tissue. This work investigates the impact of proton beam arrangement on LETd and biological dose for nasopharyngeal carcinoma.
Methods:
Six retrospective nasopharyngeal carcinoma cases were planned in RayStation2023B with two beam arrangements: Beamset1(B1): LAO/RAO/ASO and Beamset2(B2) LAO/LPO/RAO/RPO. CTV target sizes ranged from 109-734cc with plans normalized to achieve nominal plan CTV coverage at D99=Rx. All cases were robustly optimized and evaluated (setup uncertainty=3mm, range uncertainty=3.5%) using Monte Carlo v5.5 (0.3% uncertainty) and a proton RBE factor of 1.1. Beam contributions were limited to 70% from a single field to allow for OAR sparing. LETd distributions were used to calculate variable RBE dose using the Carabe, McNamara, and Wedenberg models with an α/β=3Gy. Mean differences (B1-B2) in D1% and Dmean were compared while average LETd was reported for OARs.
Results:
B1 delivered more conformal dose distributions posterior to the target with reduced D1% to brainstem (5.6Gy), chiasm (4.5Gy) and cord (1.4Gy) for RBE=1.1 when compared to B2. Average LETd is significantly higher given the anterior beam arrangement of B1 at 6.0, 3.8 and 5.1keV/um for brainstem, chiasm and cord compared to B2 at 2.4, 2.7 and 3.7keV/um, respectively. Despite the elevated LETd of B1, variable RBE models showed a change in mean D1% of less than 1.5Gy when compared to RBE=1.1. Conversely, OARs lateral to the target experienced reduced mean doses (RBE1.1) for the B2 beam arrangement, including lens (48%) and retina (22%) as well as reduced mean D1% for optic nerve (5%).
Conclusion:
B1 yielded favorable LETd-weighted dose distributions for cord, chiasm and brainstem, while B2 limited dose to lateral optic structures. Future work will focus on identifying geometrically favorable cases for each beam arrangement.

Back to List