Test-retest reliability of high spatial resolution diffusion tensor and diffusion kurtosis imaging

Pashtun Shahim,Joong H Kim,Laurena Holleran,David L Brody

doi:10.1038/s41598-017-11747-3

Abstract

We assessed the test-retest reliability of high spatial resolution diffusion tensor imaging (DTI) and diffusion kurtosis imaging (DKI). Diffusion MRI was acquired using a Siemens 3 Tesla Prisma scanner with 80 mT/m gradients and a 32-channel head coil from each of 3 concussive traumatic brain injury (cTBI) patients and 4 controls twice 0 to 24 days apart. Coefficients of variation (CoV) for DTI parameters were calculated in each DTI Studio parcellated white matter tract at 1.25 mm and 1.75 mm isotropic voxel resolution, as well as DKI parameters at 1.75 mm isotropic. Overall, fractional anisotropy had the best reliability, with mean CoV at 5% for 1.25 mm and 3.5% for 1.75 mm isotropic voxels. Mean CoV for the other DTI metrics were <7.0% for both 1.25 and 1.75 mm isotropic voxels. The mean CoV was ≤4.5% across the DKI metrics. In the commonly injured orbitofrontal and temporal pole regions CoV was <3.5% for all parameters. Thus, with appropriate processing, high spatial resolution advanced diffusion MRI has good to excellent test-retest reproducibility in both human cTBI patients and controls. However, further technical improvements will be needed to reliably discern the most subtle diffusion abnormalities, especially at high spatial resolution.

Highlights

Diffusion tensor imaging (DTI) has proven to be an elegant noninvasive technique for assessing the microstructural characteristics of brain white matter tracts[1]
There was no association between the Coefficients of variation (CoV) and the number of days apart the pair of the scans was acquired (Fig. S3)
The main findings of this study were: (i) the high spatial resolution diffusion imaging protocol yielded mean CoVs below 7.0% for fractional anisotropy (FA), axial diffusivity (AD), radial diffusivity (RD), and mean diffusivity (MD), averaged across white matter regions, (ii) for the 1.25 mm isotropic voxel size diffusion tensor imaging (DTI), the controls had overall lower mean CoVs than concussive traumatic brain injury (TBI) (cTBI) patients, (iii) Inferior ROIs, including the commonly injured temporal pole and orbitofrontal regions, had overall high reliability, (iv) Diffusion kurtosis imaging (DKI) parameters appeared even more reliable than DTI parameters

Summary

Introduction

Diffusion tensor imaging (DTI) has proven to be an elegant noninvasive technique for assessing the microstructural characteristics of brain white matter tracts[1]. DTI allows the characterization of water diffusion in white mater through metrics such as fractional anisotropy (FA), axial diffusivity (AD), radial diffusivity (RD), and mean diffusivity (MD). These metrics have been shown in numerous studies to delineate subtle white matter pathologies in conditions including demyelinating diseases, movement disorders, neurodegenerative diseases, and traumatic brain injury (TBI)[3,4,5,6,7,8,9]. Acquired from several different 3 Tesla scanners including 8–20 channel head coils and showed good reliability (CoV of

Objectives

Discussion

Conclusion