To determine the utility of diffusion kurtosis imaging (DKI) as an advanced neuroimaging biomarker for Mild Traumatic Brain Injury (mTBI).

In the absence of structural abnormalities in mTBI, diffusion tensor imaging (DTI) metrics are often inadequate to determine longitudinal changes in microstructural white matter integrity. DKI extends conventional DTI for greater sensitivity and specificity of microstructural axonal injury.

111 mTBI patients and 32 controls (15-50 years old) were enrolled acutely after mTBI and followed with up to 4 standardized serial assessments over 3 months. Patients were enrolled at either Encounter 1 (E1), within 72 hours, or Encounter 2 (E2), 5-10 days post-injury, and returned for Encounter 3 (E3) at 15-29 days and Encounter 4 (E4) at 83-97 days. Each encounter included a clinical exam, neuropsychological assessment, as well as diffusion kurtosis imaging. Data were analyzed as mean percentage of diffusion kurtosis orthogonal outlier within each region of interest. Voxels were considered an outlier if orthogonal kurtosis was >99.5% percentile of the healthy control distribution.

Within mTBI subjects, mean percentage of orthogonal kurtosis imaging (K_orth) outliers was significantly greater at E1-E3 relative to E4 (p<.05). Between-subjects analysis revealed a significant difference in mean percentage of K_orth outliers at E2 between mTBI and controls (p<.05). Furthermore, rank-sum test of regions of interest (ROI) sorted by ROIs with the greatest number of outliers at E2, identified a subset of 20 regions that differed acutely between mTBI and controls.  

Diffusion kurtosis imaging was sensitive to longitudinal changes in white matter integrity. Additionally, mean percentage of Korth outliers in a subset of 20 ROIs may serve as a potential mTBI diagnostic marker.