Abstract
BACKGROUND AND PURPOSE: Proton therapy has become widely used for non-small cell lung cancer (NSCLC) due to its ability to spare normal tissue. However, the commonly assumed relative biological effectiveness (RBE) of 1.1 may underestimate the therapy's true biological effects on lung, potentially leading to a higher than expected incidence of radiation-induced pneumonitis (RP). In this retrospective study, we evaluated the variability of organ-specific RBE of proton therapy using RP grade 2 or higher (RP2+) as an endpoint. MATERIALS AND METHODS: Data from 270 NSCLC patients across two clinical trials were analyzed, with 134 given photon therapy and 136 given proton therapy. The physical dose was obtained for all patients, and the dose-averaged linear energy transfer (LET(d)) was computed for proton therapy recipients. The Lyman-Kutcher-Burman (LKB) model was used to predict RP2+ incidence. RBE was determined using 1) the ratio of the photon dose to the proton dose achieved the same RP2+ incidence and 2) an LET-based RBE estimation derived from the LET-enhanced LKB model. RESULTS: The estimated RBE for RP2+ in proton therapy recipients was variable and consistently exceeded 1.1. The RBE estimated from the photon-to-proton dose ratio model between 1.10 and 1.35, with an average of 1.28. While the LET-enhanced LKB model yielded an RBE range of 1.13 to 1.33, with an average of 1.23. CONCLUSION: This study demonstrated that the RBE of proton therapy for RP2+ is variable and higher than the conventional 1.1 assumption, emphasizing the need to account for variable RBE in treatment planning.