Whole brain white matter integrity measures on patients with moderate to severe traumatic brain injury (TBI). FA and entropy analysis

Introduction Mean Fractional Anisotropy (FA) along with Longitudinal and Transverse Apparent Diffusion Coefficients (ADC) are the established measures in assessing white matter (WM) integrity in TBI patients. A question is raised whether mean FA performs better than Shannon Entropy (SE) as a measure of discrimination between controls and patients with moderate to severe TBI. Purpose To assess WM integrity in moderate to severe TBI patients utilizing the established diffusion tensor measures, and compare FA with Shannon Entropy as measures of assessment of WM integrity. Materials and methods A group of 13 patients with moderate to severe TBI, 2–8 years post-ictus, has been compared with a group of 14 matched healthy controls. Diffusion Tensor Imaging (DTI) data were obtained at 3T with a pulse sequence employing 32 diffusion-encoding gradient directions. After data preprocessing, Tract Based Spatial Statistics (FSL-FMRIB) for FA, whole-brain histogram analysis for SE and Receiver Operating Curve (ROC) analysis for both FA and SE have been performed. Results Higher axial and transverse diffusivity and lower FA have been observed in TBI patients ( p P  = 0.007), Global mean FA (0.8, P  = 0.0066) and SE (0.8, p  = 0.008). Conclusion Axial and transverse diffusivity are increased in TBI patients compared to controls. Diffuse Axonal Injury (DAI) is evident as global reduction of FA in TBI patients, which is driven by increased transverse diffusion changes. No significant difference is observed between FA and SE in TBI patient discrimination. Disclosure The authors declare no conflicts of interest including financial, consultant, institutional, and other relationships that might lead to bias or a any other conflict of interest. This work was supported by the Cyprus Research Promotion Foundation through a grant co-funded by the Cyprus Government and the European Regional Development Fund (F.C., PI; NEW INFRASTRUCTURE/STRATEGIC/0309/37). The authors would like to thank the study volunteers who participated in the project.