Characterizing Proton-Induced Biological Effects in a Mouse Spinal Cord Model: A Comparison of Bragg Peak and Entrance Beam Response in Single and Fractionated Exposures

Abstract

Proton relative biological effectiveness (RBE) is a dynamic variable influenced by factors like linear energy transfer (LET), dose, tissue type, and biological endpoint. The standard fixed proton RBE of 1.1, currently used in clinical planning, may not accurately represent the true biological effects of proton therapy (PT) in all cases. This uncertainty can contribute to radiation-induced normal tissue toxicity in patients. In late-responding tissues such as the spinal cord, toxicity can cause devastating complications. This study investigated spinal cord tolerance in mice subjected to proton irradiation and characterized the influence of fractionation on proton- induced myelopathy at entrance (ENT) and Bragg peak (BP) positions. Cervical spinal cords of 8-week-old C57BL/6J female mice were irradiated with single- or multi-fractions (18x) using lateral opposed radiation fields at 1 of 2 positions along the Bragg curve: ENT (dose-mean LET = 1.2 keV/μm) and BP (LET = 6.9 keV/μm). Mice were monitored over 1 year for changes in weight, mobility, and general health, with radiation-induced myelopathy as the primary biological endpoint. Calculations of the RBE of the ENT and BP curve (RBEENT/BP) were performed. Single-fraction RBEENT/BP for 50% effect probability (tolerance dose (TD50), grade II paresis, determined using log-logistic model fitting) was 1.10 ± 0.06 (95% CI) and for multifraction treatments it was 1.19 ± 0.05 (95% CI). Higher incidence and faster onset of paralysis were seen in mice treated at the BP compared with ENT. The findings challenge the universally fixed RBE value in PT, indicating up to a 25% mouse spinal cord RBEENT/BP variation for multifraction treatments. These results highlight the importance of considering fractionation in determining RBE for PT. Robust characterization of proton-induced toxicity, aided by in vivo models, is paramount for refining clinical decision-making and mitigating potential patient side effects.