RBE-Guided Treatment Planning, LET Optimization, and Implications of Proton Arc Therapy for the Sparing of Nervous Tissue in Head and Neck Proton Therapy

Background/Objectives: Traditionally, proton therapy assumes a fixed relative biological effectiveness (RBE) value of 1.1, which can lead to inaccurate dose calculations in treatment planning. This study examines the effect of dynamic arc delivery and pencil beam scanning (PBS) linear energy transfer (LET) optimization on LET-dependent RBE dose escalation to nervous tissue structures in head and neck (H&N) cancer patients undergoing proton beam therapy, utilizing two RBE dose models. Methods: Fifteen head and neck cancer patients previously treated with PBS proton therapy at high risk of nervous tissue toxicity were retrospectively analyzed. Three plans were developed for each patient: PBS, dynamic arc, and PBS with LET optimization. RBE-weighted dose distributions were calculated and compared for all patient plans using a linear LET-weighted model and a tissue-specific α/β-dependent model. LET-dose (LETd) constraints were systematically tested to determine optimal values for plan quality and efficacy in altering LET spatial distribution in regions of concern. LET-dependent RBE enhancement was calculated for the three different planning methods. Results: Dynamic arc plans increased RBE enhancement when compared to PBS, while LET optimization for PBS consistently reduced RBE-enhanced dose to nervous tissue compared to non-optimized plans for both RBE models. Patient-specific variability in optimization benefit was observed, with the most significant improvements in cases with greater initial RBE enhancement. A maximum LETd constraint of 2.5 µm/keV above 80% of the maximum structure dose threshold was found to balance plan quality and RBE mitigation. Conclusions: Dynamic arc delivery increased RBE enhancement relative to static PBS. LET optimization successfully modified LETd spatial distributions to minimize RBE enhancement to nervous tissue structures when compared to non-LET optimized PBS and dynamic arc plans. Patient-specific risk stratification should be used when clinically deploying LET optimization.