Patients who suffer a traumatic brain injury (TBI) often exhibit symptoms which are unexplained by conventional neuroimaging, which lacks sensitivity (to microscopic injury), and specificity (to focal abnormalities)¹. Sodium MRI can assess loss of Na⁺ homeostasis², but has not yet been applied to TBI, despite strong basic science evidence that ionic imbalances involving the Na⁺ ions are the driving force behind the cascade of cell damage after TBI³. We propose therefore the use of ²³Na MRI to characterize long-term aspects of TBI injury that have not been studied before, specifically, perturbations in the intracellular Na⁺ concentration (C₁) and extracellular volume fraction (α₂)^4,5. Our hypothesis is shown in Fig. 1A. In this preliminary study, we compared global and local C₁ and α₂ measurements in patients in chronic stage of TBI with healthy controls.

MRI scans: Four TBI patients (4 men; median age 36 yr, range 22-72 yr) and 6 age- and gender-matched controls were scanned at 3 T (Prisma, Siemens) with an 8-channel transmit-receive ¹H/²³Na head coil built in our RF Laboratory. The Glascow Coma Scale (GCS) was known for three TBI patients (GSC = 6, 8, 14), yielding classifications of severe and mild TBI. No GCS was obtained in 1 patient. All were scanned at the chronic TBI stage (median time from injury 1.4 yr, range 1-12 yr). Injury modes were fall, assault and motor vehicle accident. Two ²³Na MRI were performed: (1) FLORET^4,6: 3 hubs, cone angle 45°, 120 interleaves/hub, FA 80°/1 ms, TE 0.2 ms, TR 100 ms, FOV 320 mm, resolution 5 mm isotropic, 20 averages, TA 12:00 min; (2) FLORET with fluid suppression by inversion recovery (IR): same parameters as (1) except: inversion pulse 180°/6 ms, TI 25 ms, FA 90°/1 ms, 30 averages, TA 18:00 min.

Data processing: Images were reconstructed in Matlab with 3D regridding and nominal isotropic resolution of 2.5 mm. Both ²³Na acquisitions were used to generate C₁ and α₂ maps of the grey and white matters (GM, WM) and whole brain using linear regression of reference Agar gel phantoms and a 3-compartment model (Fig. 1B)^4,5. MPRAGE was used to generate masks of GM and WM.

Statistics: The rank sum test was applied to the mean, skewness and kurtosis of the distributions of C₁ and α₂ values over whole GM, WM, and brain in all subjects to assess the significance of their difference between control and TBI. Statistical difference was defined as p<0.05.

Fig. 2 exhibits examples of ²³Na images with and without IR, MPRAGE and C₁ and α₂ maps. These images reveal regional abnormalities in the TBI patient (severe TBI, 1 year after injury) compared to a similar slice from a control subject: some regions have very low C₁ but high α₂ (no more cells, edema), while some present both high C₁ and α₂ (cell death in progress with concomitant reduction of cell packing and edema).

Fig. 3 presents typical distributions of all C₁ and α₂ values over the whole brain of the same subjects as Fig. 2. We can visually detect that the shapes of these distributions are different, with statistical measures such as skewness and kurtosis significantly different, while mean values not significantly different.

Fig. 4 shows boxplots of the mean, skewness and kurtosis over whole GM and WM, from all subjects. Only C₁ skewness and kurtosis were significantly different between TBI patients and controls. Mean C₁ and α₂ measurements were not significantly different, but with trends for lower C₁ and higher α₂ in TBI, likely driven by focal (MRI-apparent) cell loss, as shown in Fig. 2.

There were no statistically significant findings for global measurements of mean C₁ and α₂, but more conclusive findings from the abnormal C₁ distributions (skewness and kurtosis) in all TBI patients, suggesting large amount of injury heterogeneity. Regional analyses show large variations of mean C₁ and α₂ in lesions detected on ¹H images, compared to normal values (see Fig. 1A). However, these lesions all look similar on the MPRAGE image, but present differences in C₁ and α₂ values between lesions, allowing to distinguish areas of the brain with edema and areas with cell death in progress. These preliminary results on a small number of patients nevertheless suggest that quantitative sodium MRI may be a useful imaging biomarker for loss of homeostasis and cell death in TBI. Both global and regional measurements of C₁ and α₂ could therefore be used to assess the degree of neurodegeneration in chronic TBI patients, and help assess the efficiency of potential treatments in longitudinal studies.