Acute Diffusion Tensor and Kurtosis Imaging and Outcome following Mild Traumatic Brain Injury

Abstract

In this prospective cohort study, we investigated associations between acute diffusion tensor imaging (DTI) and diffusion kurtosis imaging (DKI) metrics and persistent post-concussion symptoms (PPCS) 3 months after mild traumatic brain injury (mTBI). Adult patients with mTBI (n = 176) and community controls (n = 78) underwent 3 Tesla magnetic resonance imaging (MRI) within 72 h post-injury, estimation of cognitive reserve at 2 weeks, and PPCS assessment at 3 months. Eight DTI and DKI metrics were examined with Tract-Based Spatial Statistics. Analyses were performed in the total sample in uncomplicated mTBI only (i.e., without lesions on clinical MRI), and with cognitive reserve both controlled for and not. Patients with PPCS (n = 35) had lower fractional anisotropy (in 2.7% of all voxels) and kurtosis fractional anisotropy (in 6.9% of all voxels), and higher radial diffusivity (in 0.3% of all voxels), than patients without PPCS (n = 141). In uncomplicated mTBI, only fractional anisotropy was significantly lower in patients with PPCS. Compared with controls, patients with PPCS had widespread deviations in all diffusion metrics. When including cognitive reserve as a covariate, no significant differences in diffusion metrics between patients with and without PPCS were present, but patients with PPCS still had significantly higher mean, radial, and axial diffusivity than controls. In conclusion, patients who developed PPCS had poorer white matter microstructural integrity acutely after the injury, compared with patients who recovered and healthy controls. Differences became less pronounced when cognitive reserve was controlled for, suggesting that pre-existing individual differences in axonal integrity accounted for some of the observed differences.

Keywords: biomarkers; brain concussion; diffusion kurtosis imaging; diffusion tensor imaging; post-concussion syndrome.

FIG. 1.

Significant contrasts from Tract-Based Spatial Statistics: PPCS+ vs PPCS-. Patients with (PPCS+) vs without (PPCS-) persistent post-concussion symptoms; total sample (A), uncomplicated sample (B). Statistically significant voxels in red and yellow. FA, fractional anisotropy; KFA, kurtosis fractional anisotropy; RD, radial diffusivity.

FIG. 2.

Significant contrasts from Tract-Based Spatial Statistics: PPCS+ vs Controls. Patients with persistent post-concussion symptoms (PPCS+) versus healthy controls; total sample (A), uncomplicated sample (B). Statistically significant voxels in red and yellow. AD, axial diffusivity; FA, fractional anisotropy; Kax, axial kurtosis; KFA, kurtosis fractional anisotropy; Kmean, kurtosis mean; Krad, radial kurtosis; MD, mean diffusivity; RD, radial diffusivity.

FIG. 3.

Significant contrasts from Tract-Based Spatial Statistics: PPCS- vs Controls. Patients without persistent post-concussion symptoms (PPCS-) versus healthy controls, total sample (A), uncomplicated sample (B). Statistically significant voxels in red and yellow. Kax, axial kurtosis; Kmean, kurtosis mean; Krad, radial kurtosis.