Application of K-Space Energy Spectrum Analysis for Inherent and Dynamic B0 Mapping and Distortion Correction in DTI

Introduction Diffusion tensor imaging (DTI) is a powerful technique for the noninvasive characterization of the microstructure of normal and pathological tissue. However, it is typically performed with echo-planar imaging (EPI) and is thus vulnerable to spatial and temporal variations of the static magnetic field (B0) caused by susceptibility effects near air/tissue interfaces, eddy currents induced by the diffusion-weighting gradients, subject motion, physiological noise, and system instabilities. These B0 variations result in geometric distortions and misregistration among different diffusion-weighted images, leading to errors in the derivation of the diffusion tensor and consequently in mean diffusivity or fractional anisotropy (FA) maps as well as in fiber tracking procedures. Many correction methods have been proposed, but result in a reduced signal-to-noise ratio (SNR) (1), require a substantially longer scan time (2–6), involve the coregistration of images acquired with highly variable contrasts (7–10), and cannot effectively correct for all types of artifacts. To address these limitations, we propose a novel method integrating (i) a k-space energy spectrum analysis (KESA) algorithm (11) and (ii) a spin-echo (SE)/asymmetric SE (ASE) acquisition scheme, which can inherently and dynamically generate a B0 map from the k-space data for each baseline (b = 0) and diffusion-weighted image, without requiring any additional data acquisition. These B0 maps are then used to effectively and efficiently correct for the distortions due to both spatial and temporal B0 variations before derivation of the diffusion tensor, resulting in a high spatial fidelity and accuracy.