Prediction of radiation necrosis in a rodent model using magnetic resonance imaging apparent transverse relaxation (R₂^*).

BACKGROUND AND PURPOSE: Radiation necrosis remains an irreversible long-term side-effect following radiotherapy to the brain. The ability to predict areas that could ultimately develop into necrosis could lead to prevention and management of radiation necrosis.
 Materials and Methods: Fischer 344 rats were irradiated using two platforms (micro-CT irradiator and x-RAD 225 IGRT) with radiation up to 30 Gy for the micro-CT and 40 Gy for the xRAD-224 to half the brain. Animals were subsequently imaged using a 9.4T MRI scanner every 2-4 weeks for up to 28 weeks using a 7-echo gradient echo sequence. The apparent transverse relaxation constant (R2*) was calculated and retrospectively analyzed.
 Results: Animals irradiated with the low-dose rate micro-CT did not exhibit and symptoms or imaging changes associated with RN. Animals irradiated with the xRAD-225 exhibited imaging changes consistent with RN at week 24. Analysis of the R2* coefficient within the lesion and hippocampus shows the potential for detection of RN up to 10 weeks prior to morphological changes.
 Conclusions: The ability to predict areas of RN and increases of R2* within the hippocampus provides a method for long-term monitoring and prediction of RN.