Free water elimination improves test–retest reproducibility of diffusion tensor imaging indices in the brain: A longitudinal multisite study of healthy elderly subjects

Abstract

Free water elimination (FWE) in brain diffusion MRI has been shown to improve tissue specificity in human white matter characterization both in health and in disease. Relative to the classical diffusion tensor imaging (DTI) model, FWE is also expected to increase sensitivity to microstructural changes in longitudinal studies. However, it is not clear if these two models differ in their test-retest reproducibility. This study compares a bi-tensor model for FWE with DTI by extending a previous longitudinal-reproducibility 3T multisite study (10 sites, 7 different scanner models) of 50 healthy elderly participants (55–80 years old) scanned in two sessions at least one week apart. We computed the reproducibility of commonly used DTI metrics (FA: fractional anisotropy, MD: mean diffusivity, RD: radial diffusivity and AXD: axial diffusivity), derived either using a DTI model or a FWE model. The DTI metrics were evaluated over 48 white matter regions of the JHU-ICBM-DTI-81 white-matter labels atlas, and reproducibility errors were assessed. We found that relative to the DTI model, FWE significantly reduced reproducibility errors in most areas tested. In particular, for the FA and MD metrics there was an average reduction of approximately 1% in the reproducibility error. The reproducibility scores did not significantly differ across sites. This study shows that FWE improves sensitivity and is thus promising for clinical applications, with the potential to identify more subtle changes. The increased reproducibility allows for smaller sample size or shorter trials in studies evaluating biomarkers of disease progression or treatment effects.

Design and Validation of Diffusion MRI Models of White Matter

“…To circumvent this issue, Pasternak and colleagues initially imposed additional constraints: spatial regularization and tighter upper and lower bounds on the tissue volume fraction, dependent on tissue diffusivities. This constrained implementation has been applied to a variety of conditions (Albi et al, 2017; Bergamino et al, 2016; Oestreich et al, 2016; Ofori et al, 2015). …”

Novel Diffusion-Kurtosis-Informed Template Reduces Distortions due to Partial Volume Effects and Improves Statistical between-Group Comparisons

“…The most eminent methods are represented by the DTI signal acquisition method combined with fluid attenuated inversion recovery (FLAIR) [48,49] and the free water elimination (FWE) method [50,51] accounting for mixed diffusion signal within a voxel. It was demonstrated that, in general, suppression of CSF contamination enhances the accuracy and reproducibility of DTI metrics and improves tractography results [48,49,[52][53][54], especially in periventricular regions. The FWE correction was also applied to improve accuracy of DTI-derived metrics of limbic tracts in the study of their developmental trajectories [55].…”

“…Correcting for CSF contamination was shown to be of paramount importance for disentangling atrophy-based artefacts in ageing [54,56] and neurodegenerative pathologies, such as mild cognitive impairment [57], Huntington's disease [52], or Alzheimer's disease [58]. Accurate estimation of diffusion metrics can be especially important in elucidation of very early stages of Alzheimer's disease.…”

Dedicated diffusion phantoms for the investigation of free water elimination and mapping: insights into the influence of T2 relaxation properties

“…14 So far, free water elimination (FWE) DTI models have been shown to reduce bias in DTI metrics and fibre tractography 13,14,19,26 and have also improved test-retest reproducibility. 27 Moreover, the free water volume fraction map has been shown to be a sensitive biomarker in vasogenic oedema 16 and neurodegenerative diseases such as dementia, 28 Alzheimer's disease 29 and Parkinson's disease. [30][31][32] Although the aforementioned methods stabilize the parameter estimation problem, imposing these model assumptions leads to the risk of producing biased results.…”