226 - Toward Adaptive Dose Painting with Hypoxia Biomarkers: Feasibility of Oxygen-Enhanced (OE)-MRI T₁ Mapping on Low-Field MR-Linac

C. Park¹, N. Warner^1,2, E. Kaza¹, S. Tanguturi¹, and A. Sudhyadhom¹; ¹Department of Radiation Oncology, Brigham and Women’s Hospital and Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, ²Harvard-MIT Health Sciences and Technology, Massachusetts Institute of Technology, Cambridge, MA

Purpose/Objective(s): Stereotactic MR-guided Adaptive Radiation Therapy (SMART) dose painting in hypoxia shows promise in improving treatment outcomes but clinically feasible imaging and localization remains a challenge. Oxygen-enhanced (OE)-MRI is compelling for hypoxia imaging via preferential O₂ uptake and T₁ relaxation shortening in normoxic tissues, as changes in 1/T₁ (?R) quantifiability reflect tissue oxygenation level dependent (TOLD) contrast. While OE-MRI has been investigated at high-fields, no studies have optimized low-field TOLD T₁ detection and quantification. This work aims to establish the feasibility of clinically-efficient hypoxia imaging with TOLD T₁ contrast on a 0.35T MR-Linac by optimizing a low-field MP2RAGE sequence for time-efficient dynamic T₁ mapping and quantifying T₁ sensitivity and O₂ change detectability limits. Materials/

Methods: Clinically-efficient low-field OE-MRI with dynamic volumetric acquisitions to quantify and map O₂ distributions was achieved by developing an optimized 3D MP2RAGE T₁ mapping method that maximizes signal-to-noise ratio (SNR) and temporal resolution on a 0.35T MR-Linac. High-resolution whole-brain scans were acquired in 5 healthy volunteers to evaluate image fidelity and repeatability. T₁ mapping validation was performed with a phantom containing 8 chemically-formulated inserts of known T₁ values. Mean T₁ values within a volumetric region-of-interest were compared with ground-truth to evaluate T₁ linearity with the optimized sequences. Sensitivity to T₁ changes were evaluated with a dynamic 3D protocol with 20 consecutive measurements. Two-sided 95% prediction intervals were computed to determine the T₁ and ?R detectability margins to define the O₂ detectability limits.
Results: Whole-brain scans showed excellent T₁ concordance with literature values at 0.35T and high repeatability. 3D MP2RAGE optimization for clinically-efficient low-field OE-MRI with a GRAPPA parallel imaging approach achieved a balance of spatial resolution, SNR, and temporal resolution with a 26.6 s whole-brain 3D scan time, capable of 20 measurements within 8:52 min. Phantom scans resulted in T₁ maps with robust linear correlations (R²=0.983) within a clinically-relevant T₁ range. Statistical analysis reveals a median (range) T₁ detectability margin of 8.26 (3.91–16.55) ms for T₁ values 344 (180–703) ms, corresponding to a ?R₁ of 0.028 (0.009–0.097) ms^-1, which establishes high sensitivity to T₁ changes and O₂ change detectability, equivalent or superior to high-field studies.
Conclusion: We optimized a low-field MP2RAGE sequence with the capability for dynamic 3D T₁ mapping with high sensitivity and O₂ change detectability on a 0.35T MR-Linac. This work establishes the feasibility of clinically-efficient hypoxia imaging with OE-MRI TOLD T₁ contrast, enabling hypoxic tissues characterization, and contributing to novel technical advancements to biologically-guided adaption and dose painting to hypoxic spatial distributions.