Yngve Forslin
Brain gadolinium retention has received notable attention since the first report in 2014.1,2 Several studies have confirmed measurable signal changes attributable to gadolinium retention after multiple administrations of linear gadolinium-based contrast agents (GBCAs). However, results regarding macrocyclic GBCAs are conflicting,3 and the key question remains whether gadolinium retention has any negative effects.
In our previous study, we retrospectively studied gadolinium retention after exposure to linear GBCAs in patients with MS whom we had followed for 18 years and compared with nonexposed matched healthy controls. We could then illustrate long-lasting longitudinal changes in T1 signal intensity ratios and their statistical associations with cognitive impairment.4 As of yet, few other studies have focused on gadolinium retention and cognition.5 Furthermore, only a few research groups have used relaxometry to capture gadolinium retention, which could mitigate the need for a reference region (that may in itself be affected by gadolinium retention).6–8
Based on our previous study, we sought to try and replicate our results in a separate cohort with a quantitative MRI approach. Synthetic MRI is a technique for simultaneous T1 and T2 relaxometry,9 which we use routinely in our research group because we have been involved in implementing it on Siemens MRI systems.10 Therefore, during my PhD studies, I studied different applications of synthetic MRI for patients with MS, and this is naturally a patient group that is typically exposed to GBCAs multiple times.11 In our hospitals, we treat many patients with MS and therefore had a unique opportunity to study the effects of multiple doses of linear and/or macrocyclic types of GBCAs with regards to T1 and T2 relaxometry values.
In the current study, we recruited 85 patients with MS and 21 healthy controls (without exposure to GBCAs) who were stratified depending on the type of GBCAs to which they had been exposed. Another advantage with our methodology is that it is possible to generate synthetic T1 and T2 images that are inherently perfectly aligned with the R1 and R2 maps, which made it practical to fully segment the volumes of the structures on the synthesized T1 and T2 images.
Our study showed that a higher number of administrations of linear, but not macrocyclic, GBCAs was associated with a dose-dependent increase in R1, and to some degree R2, in the studied brain regions (dentate nucleus, globus pallidus, caudate nucleus, and thalamus). Moreover, higher relaxation rates were associated with lower performance in some domains of cognitive performance, but not with increased physical disability or fatigue.