165 - A Novel Biomarker for Treatment Response and Biologically Adaptive Radiotherapy: Real-Time Quantification of Radiation-Produced Free Radical Generation on Clinical MR-Linac

Purpose/Objective(s): Despite sub-millimeter spatial precision in modern radiotherapy, current approaches fail to account for biological heterogeneity in tumor and normal tissue response. Free radical generation (FRG), particularly superoxide radicals, is a key mediator of radiation-induced damage and normal tissue toxicity. Despite its critical role, no real-time, non-invasive method currently exists to quantify FRG in vivo during treatment. To address this gap, we developed Free radical IRradiation Emergent (FIRE) MRI, a novel framework leveraging quantitative MRI on a clinical MR-Linac (MRL) to dynamically measure real-time FRG. This approach establishes FRG as a patient-specific biomarker for treatment response, enabling spatially and biologically adaptive radiotherapy.

Materials/Methods: A high-temporal-resolution quantitative T₁ mapping protocol was developed on a 0.35T MRL for simultaneous imaging and irradiation. Radiation chemistry simulations were performed to model FRG dynamics, identifying superoxide radicals as the dominant paramagnetic species affecting change in T₁ relaxivity, leading to the experimental derivation of a superoxide relaxivity constant to model T₁ changes. Validation was performed in 35 biologically relevant phantoms, including hydroxyl-specific scavengers (coumarin), superoxide scavengers (MitoTEMPO), generalized free radical scavengers (glutathione), radiosensitizing gold nanoparticles, and control solutions, pure water and phosphate buffer solution (PBS). Simultaneous imaging and irradiation experiments were conducted and real-time T₁ dynamics were correlated with FRG kinetics. Real-time T₁ dynamics were analyzed using area under the curve (AUC) for ?T₁ (ms·s) and superoxide concentration (µM·s), along with superoxide production (nM/s) and dose-normalized rates (nM/Gy).

Results: FIRE MRI successfully quantified real-time radiation-produced FRG, demonstrating sensitivity and specificity to superoxide radicals. Mean T₁-derived superoxide radical production rates ranged from 0.27–3.45 nM/s (2.64–38.39 nM/Gy) with AUC values 25.08–113.62 µM·s. Superoxide specificity was validated with MitoTEMPO (0.27±0.12 nM/s), which reduced FRG, whereas gold nanoparticles (1.32±0.47 nM/s), known to enhance hydroxyl radical generation, did not significantly alter superoxide production.

Conclusion: This study establishes FIRE MRI as the first real-time, non-invasive method for quantifying radiation-produced superoxide FRG on a clinical MRL, as a novel biomarker for treatment response. This shows potential to enable FRG-based dosimetry and biologically adaptive radiotherapy, facilitating patient- and spatially-specific dose modifications based on real-time molecular responses. These findings represent a significant step towards precision radiotherapy leveraging radiation chemistry, with potential for clinical translation to improve treatment outcomes.