Daniel Krainak PhD, Presenter: Nothing to Disclose

David Thomasson, Abstract Co-Author: Nothing to Disclose

Jennifer Hufton, Abstract Co-Author: Nothing to Disclose

David Soltysik PHD, Abstract Co-Author: Nothing to Disclose

Nadia Madelaine Biassou MD, PhD, Abstract Co-Author: Nothing to Disclose

Sunder S. Rajan PhD, Abstract Co-Author: Nothing to Disclose

Clinicians and researchers use markers derived from diffusion tensor imaging (DTI) to assess disease and interventions. We examined the influence of DTI imaging parameters (signal averages, gradient directions and b-values) on the DTI metrics using an anisotropic phantom.

DTI data of an anisotropic phantom (Brain Innovations, Inc, Maastricht, NL) were acquired at 3T using a 2D single shot EPI sequence (80 2mm slices, 160x160 mm2 FOV, isotropic 2mm voxels). To investigate the sensitivity of DTI to the choice of b-value, images with various b-values (600-1500 s/mm2) were acquired (TR/TE=9000/56 ms). To evaluate the balance between directions and averages we limited imaging time to 13 minutes and acquired combinations of number of gradient directions (low (6), medium (15), and high (32)) and signal averages (1-8) with b-value = 900 s/mm2. Measures of fractional anisotropy (FA) and mean diffusivity (MD) were extracted from the images using NiPyPE (Neuroimaging in Python Pipelines and Interfaces) in combination with FSL (Oxford, UK). The effect of b-value was examined using linear regression. The combinations of signal averages and diffusion gradient directions were assessed by the coefficients of variation (CV) of the DTI metrics.

Anisotropic fiber regions within the phantom were clearly depicted using routine DTI protocols. We observed a positive correlation between b-value and FA (R2 = 0.93) and a negative correlation with b-value and MD (R2 = 0.92) in an anisotropic region-of-interest (ROI). Of the combinations of directions and averages acquired, the CV for all DTI metrics within an ROI was the lowest for the acquisition which used 15 gradient directions and three signal averages.

A balanced approach to signal averaging and increasing directions may reduce variability compared to only increasing averages or directions. Changing b-values cause systematic variations in observed DTI metrics within an ROI, and b-value needs to be considered when comparing results. Clinicians and researchers may improve within region reliability by wisely combining increased gradient orientations with signal averaging.

Optimal selection of signal averaging and diffusion gradients may reduce variability in time-sensitive exams, and users should be aware that choice of b-value affects DTI-derived metrics.

Krainak, D, Thomasson, D, Hufton, J, Soltysik, D, Biassou, N, Rajan, S, Exploring the DTI Acquisition Parameter Space with an Anisotropic Phantom.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12027888.html