KEY POINTS
- The study analyzed 23 prostate cancer patients treated on a 1.5-T MR-Linac with 36.25 Gy in 5 fractions to the prostate and a 40 Gy focal GTV boost.
- Three non-contrast MRI sequences were acquired at planning and during treatment to generate 3D maps of relative Oxygen Extraction Fraction, or rOEF, allowing repeated physiological imaging without additional ionizing radiation.
- Median average rOEF in healthy prostate tissue decreased from 24% before treatment to 13% by fraction 5. Tumor rOEF also declined, from 19% to 15%.
- Regions with particularly elevated tumor rOEF contracted during treatment. The median fraction of GTV with rOEF >70% fell from 2% at baseline to 0% by fraction 5, with significant negative temporal trends.
- The decline was specific to irradiated tissue: rOEF fell significantly in healthy prostate, GTV and an in-field gluteal muscle region, while out-of-field muscle showed no systematic change.
- Physiological change appeared early. In patients imaged after the first irradiation, the steepest healthy-prostate decline occurred between planning and fraction 1, with a median slope of −3.2, followed by −3.1 between fractions 1 and 2 before progressively flattening.
- The observed tumor trends remained significant when GTV contours were artificially expanded in different directions, suggesting that the findings were relatively robust to modest contouring uncertainty.
- Importantly, rOEF is not a validated hypoxia measurement. Elevated values can reflect altered oxygen delivery, oxygen consumption or both, and the investigators plan comparison with established hypoxia imaging before any biological dose-painting interpretation.
CLINICAL TAKEAWAY
MR-Linac imaging may eventually allow treatment adaptation based on physiology rather than anatomy alone. This study shows that radiation-associated changes can be detected after the first SBRT fraction, but rOEF should currently be regarded as an exploratory oxygen-extraction biomarker, not a map of radioresistant hypoxia.