Radiotherapy (RT) is commonly used for the treatment of primary brain tumors and metastases but cause cognitive impairments¹. Several preclinical studies have shown that radiation-induced memory and attention deficits are related to demyelination and necrosis, blood-brain barrier and neurogenesis alterations². Therefore, non-invasive and sensitive biomarkers of radiation-induced injury are needed to monitor brain damage and help in optimizing RT protocols. Previous studies showed that diffusion MRI is sensitive to brain radiation-injury^3-7. Here we aimed at establishing the time-courses of structural and metabolic changes occurring in a mouse model of brain irradiation using ¹H MRS and S-index, a new diffusion biomarker sensitive to changes in tissue microstructure.

The heads of 3-months old male C57BL/6RJ mice (n=15 for both control and irradiated groups) were exposed to a radiation dose of 15Gy (3 times 5Gy every 48h) from a ⁶⁰Co source⁸. MR acquisitions were performed one week before irradiation and at different times post-irradiation on a 11.7 T Bruker BioSpec MRI scanner equipped with a CryoProbe. Animals were anesthetized using isoflurane (1-2%). Body temperature was monitored and maintained at 37°C±0.5°C. Whole-brain anatomical (T_2w-TurboRare, resolution=0.05x0.05x0.450mm; 16 slices) and DWI images (PGSE-EPI, TE/TR=24/2500ms, 16 slices, δ/Δ=4/11.5ms, 34 b-values from 10 to 3500 s/mm², 3 orthogonal directions) were acquired as well as ¹H MR spectra (LASER, TE/TR=25/3500ms, 128 averages) from the hippocampus, striatum and olfactory bulbs (volumes = 5.6, 7.5 and 6µL respectively). Non-Gaussian diffusion parameters, ADC₀ (Apparent Diffusion Coefficient) and K (kurtosis), and IVIM parameters fIVIM (flowing blood volume fraction) and D* (pseudo-diffusion coefficient) were estimated on a voxel-by-voxel basis and in selected ROIs by fitting the signal obtained at all b values according to the IVIM/non Gaussian (kurtosis) diffusion model⁹. Moreover, a newly developed diffusion composite marker, S-index¹⁰, aimed at directly detecting minute changes in the diffusion-weighted MRI signals acquired at key b values was also calculated. This composite index was calibrated using databases of diffusion and IVIM parameters previously established on brain of healthy mice. MR spectra were analyzed using LCModel¹¹and a set of simulated spectra. The signal of macromolecules (MM) was parameterized as described elsewhere¹² and implemented in LCModel. Metabolite concentrations were derived using the total Creatine (Cr+PCr) signal as an internal reference of concentration ([Cr+PCr]=8mmol/L). All data were presented as mean ± SD. Statistical analyses were obtained using Student’s t-test (*p<0.05, **p<0.01 and ***p<0.0001) or two-ways analysis of variance (ANOVA) with multiple comparisons using Bonferroni post-hoc test (^$p<0.05, ^$$p<0.01 and ^$$$p<0.0001).

No obvious anatomical lesions, such as edema or necrosis were observed at any time (Fig. 1A), beside a slight brain atrophy (5% in the irradiated group relative to control animals, p<10^-4) from 2 months after radiation until the end of experiment (Fig. 1B). A significant decrease in S-index values was observed transiently 3 days after radiation in the hippocampus (Fig.2). The S-index remained persistently lower in the subventricular zone (SVZ) (Fig. 2) and in the olfactory bulbs of irradiated mice compared to control (Fig. 3). Those S-index drops were mirrored by increases in ADC₀ and/or decreases in K, but those changes were less or not significant (Fig. 3). No significant changes were found for fIVIM and D* in those neurogenic areas, and no differences in diffusion and perfusion parameters were observed in other regions (especially the cortex, the thalamus and the striatum) (Fig. 2). MRS revealed a longitudinal decrease in taurine in the olfactory bulb of irradiated mice compared to control group (p<0.01) (Fig. 4). The other metabolite concentrations in hippocampus were relatively similar between the groups, apart from a decrease in neuronal (NAA and GABA) and glial metabolites (myo-inositol) only at 1 month after irradiation.The observed decrease in S-index values in the olfactory bulbs and SVZ of irradiated mice are consistent with the neurogenesis decline induced by high-dose irradiation^8,13. SVZ is one of the few regions in the brain in which neurogenesis continues throughout adulthood (cells from this region can proliferate and migrate via the rostral migratory stream to the olfactory bulbs where they differentiate into neurons). Taurine is also a likely factor in neurogenesis^14,15 and its specific decrease in the olfactory bulbs is consistent with a weakened neurogenesis. The decrease in S-index reflects a decrease in diffusion hindrance and suggests a decrease in the cell population of affected areas. Immunohistological assays are underway to investigate changes in cell density post-irradiation.This preclinical study suggests that DWI, especially the composite S-index, could be a relevant biomarker to monitor non-invasively brain radiation injury and probe structural changes underlying the radiation-induced cognitive deficits.