Comparison of water exchange measurements between filter-exchange imaging and diffusion time-dependent kurtosis imaging in the human brain.

Abstract

Purpose: Filter-exchange imaging (FEXI) and diffusion time (t)-dependent kurtosis imaging (DKI(t)) are two diffusion-based methods that have been proposed for in vivo measurements of water exchange rates. Few studies have directly compared these methods. We aimed to investigate whether FEXI and DKI(t) yield comparable water exchange measurements in the human brain in vivo.

Methods: Eight healthy volunteers underwent multiple-direction FEXI and DKI(t) acquisitions on a 3T scanner. We performed region of interest (ROI) analysis to determine correlations between FEXI-derived apparent exchange rate (AXR) and DKI(t)-derived reciprocal of exchange time ( $1/{\tau }_{\mathrm{ex}}$ ).

Results: In both white matter (WM) and gray matter (GM), DKI(t) revealed substantial diffusion-time dependence of diffusivity and kurtosis. However, at t ≥ 100 ms, the diffusivity showed weak time dependence. In WM, this time dependence may be due to water exchange between myelin water and "free" water with different T₁ values, although other factors, such as remaining restrictive effects from microstructural barriers, cannot be excluded. We found a significant correlation between DKI(t)-derived $1/{\tau }_{\mathrm{ex}}$ and FEXI-derived AXR in the axial direction within WM. No such correlation was present in GM, although both values showed similar ranges.

Conclusion: These results suggest that FEXI and DKI(t) could be sensitive to the same water exchange process only when the diffusion time in DKI(t) is sufficiently long, and only in WM. In both GM and WM, the restrictive effect of microstructure is non-negligible, especially at short diffusion times (<100 ms).