Investigating the oxygen dependence of FLASH-RT using electron paramagnetic resonance imaging.

PURPOSE: The FLASH effect occurs when radiation given at ultra-high dose rates spares healthy tissue while achieving isoeffective tumor control relative to conventional (CONV) dose rate irradiation. Limited data has implicated oxygen in the FLASH effect. This study investigated the effect of tissue oxygen tension (pO2) on gastrointestinal (GI) FLASH sparing in mice. METHODS: pO2 was quantified in vivo by electron paramagnetic resonance imaging (EPRI)- noninvasive oximetry through spin probe injection and magnetic resonance co-registration- and a phosphorescent quenching microsensor. B6(Cg)-Tyrc-2 J/J female mice were imaged while breathing 10%, 21%, and 95% oxygen. C57BL/6 mice of both sexes underwent microsensor probing and FLASH or CONV total abdominal irradiation under these conditions, and GI toxicity was quantified with regenerating crypt assay. RESULTS: EPRI-measured pO2 within the liver, bowel, and kidney positively correlated with breathing-gas oxygen concentration. In the liver, the microsensor further confirmed this correlation and demonstrated lower pO2's in male mice under 10% and 21% oxygen with a similar trend under 95% oxygen compared to female mice. In both sexes, GI toxicity decreased with decreasing pO2 after FLASH and CONV. Male mice showed significant FLASH GI sparing across breathing-gas conditions, but female mice showed nonsignificant trends. For all gas conditions and both dose-rate regimes, male mice had lower toxicity (more spared crypts) than female mice, except for CONV under 21% oxygen. CONCLUSIONS: Our findings demonstrate that both CONV- and FLASH-induced GI toxicity correlate with in-vivo oxygen tension and that age-matched male and female mice differ in FLASH radiosensitivity across the breathing-gas conditions examined.