Keeping It Consistent: Assessing Measurement Repeatability of a Novel Anisotropic Phantom for Higher Order Diffusion Tensor MRI Sequences

Abstract

Diffusion MRI (dMRI) provides valuable insight into tissue microstructure for clinical and research applications. Traditional diffusion tensor imaging (DTI) models diffusion as a Gaussian process, which limits its accuracy in regions with complex fibre configurations (such as crossing or bifurcating tracts). Higher-order models, including diffusion kurtosis imaging (DKI) and constrained spherical deconvolution (CSD), address these limitations by capturing non-Gaussian diffusion behaviour or resolving multiple fibre orientations within a voxel. Despite their theoretical advantages, these models are more sensitive to noise and acquisition variability, raising concerns about repeatability. Currently, there is no standardized method to perform quality assurance (QA) on dMRI data, and there are limited studies measuring the repeatability of higher-order tensor metrics.

This thesis evaluates the repeatability of higher-order dMRI metrics using a novel anisotropic phantom developed by PreOperative Performance (Toronto, ON). The phantom contains fibre modules with controlled geometries (linear, branching, and crossing bundles) designed to mimic white matter tract architecture. Six diverse regions of interest (ROIs) were assessed across 11 independent imaging sessions using DTI, high angular resolution diffusion imaging (HARDI), and DKI protocols. Repeatability was quantified using coefficient of variation (CoV) and intraclass correlation coefficient (ICC).

DTI-derived metrics, including fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD), exhibited excellent repeatability (CoV<10%, ICC>0.9). HARDI acquisitions (60- and 90-direction) yielded slightly improved repeatability over DTI. DKI metrics, including mean kurtosis (MK), axial kurtosis (AK), radial kurtosis (RK), and kurtosis fractional anisotropy (KFA), showed greater variability, particularly in ROIs with fibre crossings. Generalized fractional anisotropy (GFA), derived from CSD, demonstrated increasing ICCs with higher angular resolution: 0.66 (DTI), 0.80 (HARDI-60), and 0.85 (HARDI-90). These findings support the phantom’s utility for repeatability testing and contribute toward developing QA standards for advanced dMRI models.