# The Impact of Gadolinium on Quantitative Myelin Metrics in Ex Vivo Spinal Cord

## TL;DR

This study examines how gadolinium affects MRI measurements of myelin in spinal cord tissue, finding significant changes in relaxation parameters but minimal impact on myelin content metrics.

## Contribution

The study provides empirical evidence on the specific effects of gadolinium on quantitative MRI metrics in ex vivo spinal cord tissue.

## Key findings

- T1 of the free proton pool and T2 of water other than myelin showed the largest differences with gadolinium.
- Myelin water fraction and related metrics changed by 2%–12% with gadolinium.
- Gadolinium flattened trends between relaxation parameters and axon diameter surrogates.

## Abstract

The goal of this work was to evaluate the impact of gadolinium on quantitative myelin MRI metrics in ex vivo tissue. Ex vivo ferret spinal cords were imaged with and without gadolinium. Mean metrics were calculated across white and gray matter. T1 of the free proton pool and T2 of water other than myelin showed the largest differences between gadolinium and without gadolinium samples. Myelin water fraction, bound pool fraction, and myelin water T2 showed a 2%–12% change between gadolinium and no gadolinium data. The relationships between the myelin water fraction, bound pool fraction, and myelin water T2 and a recently established axon diameter surrogate were not affected by gadolinium; however, the trends observed between T1 of the free proton pool and other water T2 and the axon diameter surrogate without gadolinium were flattened in the samples with gadolinium. Overall, the impacts of gadolinium were primarily observed in the relaxation parameters of the non‐myelin proton pools.

Quantitative MRI myelin metrics were evaluated in ex vivo ferret spinal cord with and without gadolinium (Gd). T1 of the free proton pool (T1f) and T2 of the intracellular and extracellular spaces (OWT2) were both significantly shorter with Gd, as expected. Metrics reporting on myelin content show significant differences between Gd and no‐Gd datasets when averaged across the white matter; however, the magnitude of these differences are likely not of concern in many applications.