Combining in vivo proton exchange rate (kex) MRI with quantitative susceptibility mapping to further stratify the gadolinium-negative multiple sclerosis lesions

Abstract

BackgroundConventional gadolinium (Gd)-enhanced MRI is currently used for stratifying the lesion activity of multiple sclerosis (MS) despite limited correlation with disability and disease activity. The stratification of MS lesion activity needs further improvement to better support clinics.PurposeTo investigate if the novel proton exchange rate (kex) MRI combined with quantitative susceptibility mapping (QSM) may help to further stratify non-enhanced (Gd-negative) MS lesions.Materials and methodsFrom December 2017 to December 2020, clinically diagnosed relapsing-remitting MS patients who underwent MRI were consecutively enrolled in this IRB-approved retrospective study. The customized MRI protocol covered conventional T2-weighted, T2-fluid-attenuated-inversion-recovery, pre- and post-contrast T1-weighted imaging, and quantitative sequences, including kex MRI based on direct-saturation removed omega plots and QSM. Each MS lesion was evaluated based on its Gd-enhancement as well as its susceptibility and kex elevation compared to the normal appearing white matter. The difference and correlation concerning lesion characteristics and imaging contrasts were analyzed using the Mann–Whitney U test or Kruskal–Wallis test, and Spearman rank analysis with p < 0.05 considered significant.ResultsA total of 322 MS lesions from 30 patients were identified with 153 Gd-enhanced and 169 non-enhanced lesions. We found that the kex elevation of all lesions significantly correlated with their susceptibility elevation (r = 0.30, p < 0.001). Within the 153 MS lesions with Gd-enhancement, ring-enhanced lesions showed higher kex elevation than the nodular-enhanced ones’ (p < 0.001). Similarly, lesions with ring-hyperintensity in QSM also had higher kex elevation than the lesions with nodular-QSM-hyperintensity (p < 0.001). Of the 169 Gd-negative lesions, three radiological patterns were recognized according to lesion manifestations on the kex map and QSM images: Pattern I (kex+ and QSM+, n = 114, 67.5%), Pattern II (only kex+ or QSM+, n = 47, 27.8%) and Pattern III (kex– and QSM–, n = 8, 4.7%). Compared to Pattern II and III, Pattern I had higher kex (p < 0.001) and susceptibility (p < 0.05) elevation. The percentage of Pattern I of each subject was negatively correlated with the disease duration (r = –0.45, p = 0.015).ConclusionAs a potential imaging biomarker for inflammation due to oxidative stress, in vivo kex MRI combined with QSM is promising in extending the clinical classification of MS lesions beyond conventional Gd-enhanced MRI.

Quantitative susceptibility mapping in multiple sclerosis: A systematic review and meta-analysis

“…The number of hyperintense QSM lesions in reference to total lesion count on conventional scans (e.g., T2, FLAIR) was most frequently documented. Seventeen studies provided data on the percentages of hyperintense QSM lesions, showing varying proportions that ranged from 13 % to 88 % ( Guo et al, 2021 , Liao et al, 2023 ; Haacke et al, 2021; Castellaro et al, 2017 , Chawla et al, 2016 , Chawla et al, 2018 , Chen et al, 2014 , Cronin et al, 2016 , Gillen et al, 2021 , Harrison et al, 2016 , Kakeda et al, 2015 , Li et al, 2016 , Pelizzari et al, 2020 , Rahmanzadeh et al, 2022 , Tolaymat et al, 2020 , Zhang et al, 2016a , Zhang et al, 2016c ). Additionally, rim lesion counts on QSM were recorded in 12 studies, reporting proportions that ranged from 4 % to 44 % ( Chawla et al, 2018 , Cronin et al, 2016 , Guo et al, 2021 , Gillen et al, 2021 , Li et al, 2016 , Harrison et al, 2016 , Kaunzner et al, 2019 , Huang et al, 2022 , Rahmanzadeh et al, 2022 , Yao et al, 2018 , Zhang et al, 2022 ).…”

B1 inhomogeneity corrected CEST MRI based on direct saturation removed omega plot model at 5T

“…The currently adopted DS removed omega plot model is a more general form by integrating multiple saturation transfer effects, which has been implemented in examining ischemic stroke and multiple sclerosis. [29][30][31] This study utilized the linear relationship between the inverse of the residual Z spectral signal with the inverse of the square of B 1 levels, making B 1 inhomogeneity correction from two CEST acquisitions possible. It is worth mentioning that the linearity of the omega plot is B 1 dependent.…”

“…In contrast, endogeneous diamagnetic CEST metabolites are vulnerable to the DS effect due to small chemical shifts, making removal of DS in the omega plot model necessary. The currently adopted DS removed omega plot model is a more general form by integrating multiple saturation transfer effects, which has been implemented in examining ischemic stroke and multiple sclerosis 29–31 . This study utilized the linear relationship between the inverse of the residual Z spectral signal with the inverse of the square of B 1 levels, making B 1 inhomogeneity correction from two CEST acquisitions possible.…”