A generalized phenomenological model of oxygen depletion for FLASH particle radiotherapy: evidence fromin vitrodata

Ultra-high dose rate (uHDR: FLASH) radiotherapy reduces normal tissue toxicity while preserving tumour control. The oxygen depletion hypothesis posits that rapid oxygen consumption during FLASH irradiation transiently reduces radiosensitivity. However, experimental measurements suggest oxygen depletion may be limited, particularly under normoxia. This study quantitatively assesses the adequacy of this hypothesis in reproducing experimentally observed sparing effects by comparing mechanistic simulations with in vitro cell survival data across a wide range of irradiation conditions. Approach. We developed a phenomenological framework for radiolytic oxygen depletion incorporating temporal dose delivery, dose rate, and linear energy transfer (LET). LET dependence was included through oxygen consumption efficiency and the maximum oxygen enhancement ratio. For direct comparison, simulation results were linked to clonogenic cell survival using a damage reduction factor (DRF) derived from the linear-quadratic model. Simulated DRFs were compared with experimental DRFs from six independent in vitro studies involving electrons, protons, helium ions, and carbon ions over broad ranges of LET, oxygenation, and dose rates using parity plots, residual analyses, stratified weighted statistics, and correlation analyses. Main results. Under most irradiation conditions, simulated DRFs remained close to unity (minimum values approximately 0.96), indicating limited predicted protection from oxygen depletion. Conversely, experimental DRFs exhibited substantially larger deviations. Stratified analyses revealed systematic discrepancies between simulated and experimental DRFs across oxygenation, LET, and dose rate, with pronounced disagreement under normoxia, high LET, and ultra-high dose rates. Residual analyses demonstrated underestimation of the sparing effect by the model. Correlation analyses showed statistically insignificant associations between simulated and experimental metrics. Significance. The oxygen depletion model considered in this work insufficiently reproduces the observed sparing effect under the in vitro conditions considered. Additional biological and physicochemical mechanisms likely play a significant role. This framework provides a quantitative basis for evaluating alternative mechanisms and integrating future experimental data toward a more complete understanding of FLASH radiotherapy.