NeuroImage: Clinical 34 (2022) 102967 Contents lists available at ScienceDirect NeuroImage: Clinical journal homepage: www.elsevier.com/locate/ynicl T1/T2-weighted ratio in multiple sclerosis: A longitudinal study with clinical associations Mateus Boaventura a, Jaume Sastre-Garriga a, Aran Garcia-Vidal b, Angela Vidal-Jordana a, Davide Quartana a, René Carvajal a, Cristina Auger b, Manel Alberich b, Mar Tintoré a, Àlex Rovira b, Xavier Montalban a, Deborah Pareto b, * a Department of Neurology-Neuroimmunology, Multiple Sclerosis Centre of Catalonia (Cemcat), Barcelona, Spain b Section of Neuroradiology, Department of Radiology, Vall d’Hebron University Hospital, Barcelona, Spain A R T I C L E I N F O A B S T R A C T Keywords: Background: T1w/T2-w ratio has been proposed as a clinically feasible MRI biomarker to assess tissue integrity in Clinically isolated syndrome multiple sclerosis. However, no data is available in the earliest stages of the disease and longitudinal studies Magnetic resonance imaging analysing clinical associations are scarce. Multiple sclerosis Objective: To describe longitudinal changes in T1-w/T2-w in patients with clinically isolated syndrome (CIS) and T1/T2-weighted ratio multiple sclerosis, and to investigate their clinical associations. Methods: T1-w/T2-w images were generated and the mean value obtained in the corresponding lesion, normal- appearing grey (NAGM) and white matter (NAWM) masks. By co-registering baseline to follow-up MRI, evolved lesions were assessed; and by placing the mask of new lesions to the baseline study, the pre-lesional tissue integrity was measured. Results: We included 171 CIS patients and 22 established multiple sclerosis patients. In CIS, evolved lesions showed significant T1-w/T2-w increases compared to baseline (+7.6%, P < 0.001). T1-w/T2-w values in new lesions were lower than in pre-lesional tissue (-28.2%, P < 0.001), and pre-lesional tissue was already lower than baseline NAWM (-7.8%, P < 0.001). In CIS at baseline, higher NAGM T1-w/T2-w was associated with multiple sclerosis diagnosis, and longitudinal decreases in NAGM and NAWM T1-w/T2-w were associated with disease activity. In established multiple sclerosis, T1-w/T2-w was inversely correlated with clinical disability and disease duration. Conclusion: A decrease in T1-w/T2-w ratio precedes lesion formation. In CIS, higher T1-w/T2-w was associated with multiple sclerosis diagnosis. In established multiple sclerosis, lower T1-w/T2-w values were associated with clinical disability. The possible differential impact of chronic inflammation, iron deposition and demyelination should be considered to interpret these findings. 1. Introduction treatments (Wattjes et al., 2015). However, at least at individual level, these sequences have poor correlation with clinical findings. Several Magnetic resonance imaging (MRI) is the most important para­ novel MRI techniques have been developed to probe MS related cerebral clinical tool in diagnosing and monitoring multiple sclerosis (MS), and a changes in a more specific and quantitative manner. Magnetization key surrogate endpoint in clinical trials. However, pathophysiological transfer imaging, diffusion tensor imaging, and myelin-water imaging research has shown that conventional MRI has limited sensitivity to the have been described as more specific techniques to unravel the diversity microstructural changes, both at the lesional level and the normal- of microstructural tissue changes not visible with conventional MRI. appearing grey (NAGM) and white matter (NAWM) (Filippi et al., Nevertheless, these techniques have limitations, such as longer acqui­ 2019). In clinical routine, new T2-weighted (w) lesions and contrast- sition times, poor spatial resolution, lack of harmonization or the need enhancing lesions are commonly used for monitoring subclinical dis­ for expertise in image post-processing (Enzinger et al., 2015). ease activity and evaluating the effectiveness of pharmaceutical The ratio of T1-w and T2-w image intensities (T1-w/T2-w) from * Corresponding author at: Radiology Department, Vall d’Hebron University Hospital, Psg, Vall d’Hebron 119-129, Barcelona 08036, Spain. E-mail address: deborah.pareto.idi@gencat.cat (D. Pareto). https://doi.org/10.1016/j.nicl.2022.102967 Received 3 December 2021; Received in revised form 4 February 2022; Accepted 14 February 2022 Available online 16 February 2022 2213-1582/© 2022 Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). M. Boaventura et al. NeuroImage: Clinical 34 (2022) 102967 standard MRI protocols was initially described as a biomarker of cortical mm3), and 3) transverse fast FLAIR (TR = 9,000 ms, TE = 87 ms, TI = myelin content (Glasser and Van Essen, 2011; Glasser et al., 2014; Shams 2,500 ms, flip angle = 120◦ , voxel size = 0.49 × 0.49 × 3.0 mm3). et al., 2019) although further studies have established T1-w/T2-w as a measure of microstructural integrity (Righart et al., 2017; Arshad et al., 2.3. MRI qualitative analysis 2017; Uddin et al., 2019; Pareto et al., 2020). Two works showed the application of the T1-w/T2-w ratio in cortical pathology in MS (Righart To define any MRI abnormalities, to fulfill McDonald 2017 diag­ et al., 2017; Nakamura et al., 2017). Both studies have demonstrated an nostic criteria and to exclude alternative diagnosis, all MRI scans were association between reduced T1-w/T2-w values and tissue damage. analyzed by an expert neuroradiologist. This analysis included the Furthermore, two other studies have found reduced T1-w/T2-w ratios in number and location of hyperintense lesions on T2-FLAIR scans, the the NAWM of MS patients compared to healthy controls (Beer et al., presence of gadolinium-enhancing lesions and the number of new T2-w 2016; Cooper et al., 2019). Taking it all into consideration, both low lesions. A normal brain MRI was defined as displaying no WM abnor­ myelin content and/or loss of microstructural integrity may explain malities ≥ 3 mm suggestive of MS (Filippi et al., 20192019). most of these findings. However, longitudinal studies with clinical as­ sociations are scarce and few data is available in the earliest stages of the 2.4. Processing of the T1-w/T2-w ratio imaging data disease. In this study, we set out to assess, in patients with clinically isolated T1-w/T2-w images were generated as proposed by Ganzetti et al., syndrome (CIS) and in a cohort of relapsing remitting MS patients (Ganzetti et al., 2014) with the MRTool - Multimodal Mapping extension treated with disease modifying drugs, baseline and longitudinal changes (v. 1.2, Swiss Federal Institute of Technology, Zurich, Switzerland, in T1-w/T2-w ratio in NAGM and NAWM with clinical associations. We http://www.fil.ion.ucl.ac.uk/spm/ext) for Statistical Parametric Map­ also aimed to investigate the dynamics of T1-w/T2-w ratio in focal ping (SPM) 12 (University College London, London, UK, http://www.fil. lesional tissue to gauge the microstructural changes underlying lesion ion.ucl.ac.uk/spm). The processing includes Intensity Non-Uniformity formation and evolution. correction and rigid registration of the T2-w images to the T1-w im­ ages, with linear calibration of image intensity modes using non-brain 2. Subjects/Materials and methods tissue masks (eye and temporalis muscle from both the T1-w and T2-w images). The output includes: 1) a normalized, scaled and unbiased 2.1. Patients and clinical assessments T1-w and T1-w/T2-w ratio image, and 2) a normalized unbiased T1-w image, which can be used for tissue segmentation (to allow the evalu­ We analyzed two cohorts described as CIS cohort and MS cohort. First, ation of regional values in T1-w/T2-w in the same space). For an over­ we assessed previously acquired data from a prospective cohort of pa­ view of the entire processing pipeline see Fig. 1. tients with CIS (CIS cohort) who underwent brain MRI for diagnosis and for monitoring disease evolution (Tintore et al., 2015). CIS was defined 2.5. Brain tissue segmentation and T1-w/T2-w for NAWM and NAGM as an episode suggestive of CNS inflammatory demyelination (McDo­ nald et al., 2001). At baseline, the demographic data, previous history of Tissue segmentation into GM, WM, and CSF was achieved using the neurological abnormalities, the CIS topography, and disability accord­ segment tool implemented in SPM12. This procedure was applied to the ing to the Expanded Disability Status Scale (EDSS) score were recorded. normalized T1-w image. Lesion masks were subtracted from the WM and The patients were evaluated on a regular basis (every 3 to 6 months) to GM masks to create NAWM and NAGM masks. Only voxels with a WM or assess both the EDSS score and the occurrence of relapses. A lumbar GM probability higher than 0.7 in T1-w were used to define the NAGM puncture to assess the presence of IgG oligoclonal bands (OB) was per­ and NAWM masks. Mean intensity values on the T1-w/T2-w images formed within the first 3 months of disease onset. Baseline MRI was were obtained for the NAGM and NAWM masks in each subject. An obtained within the first 3 to 5 months after the onset of symptoms and additional analysis was performed to generate separate masks for deep the second, at 12 months after clinical onset. After excluding alternative and cortical GM. Deep GM was segmented with FIRST, (Patenaude et al., diagnosis, all CIS patients were analyzed, regardless of the presence or 2011) the corresponding mask was created and removed from the absence of brain lesions that met criteria for MS at first relapse (e.g, CIS NAGM, generating, as a result, the cortical GM masks. that are already clinically definite MS patients at baseline, and also isolated myelitis or optic neuritis with a normal brain MRI). Fulfillment 3. Generation of lesion masks and T1-w/T2-w for lesional tissue of McDonald 2017 diagnostic criteria was assessed at baseline (Thompson et al., 2018). From this first cohort, the clinical data used 3.1. Baseline, follow-up and evolved lesion masks (EDSS and relapses) were analyzed up to 3 years of follow-up. In the second cohort (MS cohort), which consisted of established relapsing Baseline and follow-up WM lesion masks were obtained with the remitting MS patients starting on disease-modifying treatment, a base­ Lesion Segmentation Tool (LST) toolbox version 1.2.3. for SPM12, using line brain MRI was acquired at treatment onset, and a follow-up brain T2-FLAIR and T1-w images in all cohorts and the corresponding vol­ MRI 12 months afterwards. In the MS cohort, baseline and 1-year follow- umes, calculated. In the CIS cohort and the MS cohort, baseline MRI and up clinical data was retrospectively collected, including EDSS and MS lesion mask were co-registered to the follow-up MRI scan to create a relapses. mask of the evolved lesional tissue, regardless of their appearance on The study was approved by the local ethical committee and a written follow-up scans (see Fig. 2). informed consent was signed by the participating patients. 3.2. New lesions and pre-lesional tissue masks 2.2. MRI acquisition In the CIS cohort, new T2-w lesions visually detected by a trained All patients from the CIS cohort and the MS cohort underwent brain technician on follow-up MRI and were annotated on T2-proton density MRI at baseline and follow-up on the same 3 T magnet (Tim Trio; (PD) images by using the semiautomatic tool included in Jim 5.0 (http:/ Siemens, Erlangen, Germany) with a 12-channel phased array head coil. www.xinapse.com/home.php). New lesion masks were later revised and The MRI protocol included the following sequences: 1) sagittal T1-w 3D confirmed by an expert neuroradiologist. For technical reasons, new MPRAGE (TR = 2,300 ms, TE = 2.98 ms, TI = 900 ms, voxel size = 1.0 × lesion masks were only available for 52 out of 59 patients with new 1.0 × 1.2 mm3), 2) transverse proton density (PD)- and T2-w fast spin- lesions. T2-PD images and masks for new lesions were then co-registered echo (TR = 3,080 ms/TE = 21–91 ms, voxel size = 0.78 × 0.78 × 3.0 to baseline MRI scans using SPM12 (first to follow-up and then to 2 M. Boaventura et al. NeuroImage: Clinical 34 (2022) 102967 Fig. 1. Overview of the image acquisition and processing pipeline. Flow diagram of steps to generate tissue segmentation and T1- w/T2-w images in clinically isolated syn­ drome and multiple sclerosis cohorts. See text for further details. Abbreviations: wT1- w = normalized T1-w; LST = lesion seg­ mentation tool; SPM = statistical parametric mapping; GM = grey matter; WM = white matter; NAGM = normal-appearing grey matter; NAWM = normal-appearing white matter; basLT = baseline lesion tissue mask; evolLT = mask of the evolved lesions; newLT = mask for new lesions; preLT = pre- lesional tissue mask; longLT = longitudinal lesion tissue mask; PD = proton density; Co- reg = corregistration. baseline MPRAGE images) to create a mask of baseline pre-lesional tis­ normality. Unpaired parametric data using unpaired t-test and unpaired sue (see Fig. 2). non-parametric data using independent- samples Mann-Whitney U test. Paired analyses were compared with paired t-test for parametric data 3.3. T1-w/T2-w Ratio in lesion masks and related samples Wilcoxon signed rank test for non-parametric data. For multigroup comparisons, repeated measures analysis of variance Since the normalized T1-w/T2-w image from MRTool and the lesion (ANOVA) was performed to compare T1-w/T2-w values between masks masks were in the native space and differ topographically across the and times, followed by Bonferroni multiple comparisons to identify brain, we added an extra step to analyze the corresponding lesion between which masks and times the differences occur. topography. On one side, the raw T1-w images and their corresponding In the CIS cohort, to evaluate differences between patients with or lesion masks, and, on the other side, the calibrated and unbiased T1-w without suggestive MS diagnosis at CIS onset (abnormal MRI, positive and T1-w/T2-w images, all were registered to the MNI152 T1-w 1 mm OB and 2017 McDonald criteria fulfillment) and with or without in­ brain template from the Montreal Neurological Institute. Mean intensity flammatory disease (new lesions and new relapses) in T1-w/T2-w values on the T1-w/T2-w images were obtained for all lesion masks. An values, we performed a general linear model with adjustment for age additional exploratory analysis was performed, assessing separately T1- and sex. Bivariate correlations were analyzed with Pearson correlation w and T2-w signal intensity at baseline and follow-up lesion masks to for parametric data and Spearman’s rank correlation coefficient for non- analyze whether the numerator or denominator was the main driver of parametric data. Partial correlations were assessed using covariates of the changes observed. interest (age and sex). A two-tailed P < 0.05 was considered statistically significant. 3.4. Statistical analysis 4. Data availability Statistical analysis was performed by using IBM SPSS Statistics 26 for Requests for access to the data reported in this article will be Mac (IBM, Armonk, USA). Kolmogorov-Smirnov test was used to assess 3 M. Boaventura et al. NeuroImage: Clinical 34 (2022) 102967 Table 1 Demographics, clinical and MRI characteristics. Cohort CIS MS p (n ¼ 171) (n ¼ 22) Female, n (%) 110 (64.3) 14 (63.6) 0.949 Age, mean (SD) 33.2 (7.5) 36.5 (6.8) 0.052* Disease duration, years, mean (SD) 0.3 (0.12) 8.9 (7.7) <0.001* Follow-up time, years, mean (SD) 3.2 (1.7) 2.4 (1.0) 0.119* MRI interval time, months, mean 8.9 (4.5) 13.0 (6.0) <0.001* (SD) CIS topography, n (%) Optic nerve 67 (39.2) NA Brainstem 44 (25.7) NA Spinal cord 45 (26.3) NA Other 15 (8.8) NA EDSS, median (range) Baseline 1.5 (0–4.5) 1.5 (0–8.5) 0.012£ 1 year follow-up 1.0 (0–3.5) 1.5 (0–8.5) <0.001£ 3 years follow-up 1.0 (0–3.5) NA Positive OB, n (%) 82/151 17/19 0.003 (54.3) (89.5) Baseline brain MRI 0.002 Normal, n (%) 54 (31.6) 0 (0) Abnormal, n (%) 117 (68.4) 22 (1 0 0) Gadolinium T1 enhancement, n (%) 40/130 9/22 (40.9) 0.347 (30.8) Lesion volume, mL, mean (SD) Baseline 4.0 (9.3) 15.3 (26.8) 0.094* Follow-up 3.2 (5.9) 14.5 (24.7) 0.063* Fig. 2. Example of lesion evolution. Baseline (A and C) and follow-up (B and Baseline McDonald 2017 D) T1-w/T2-w images from a 33-year-old male patient who presented a spinal Fulfillment, n (%) 69 (40.4) 22 (1 0 0) <0.001 cord clinically isolated syndrome, with new brain lesions at 12 months scan. New relapses The red arrows show an example of a lesion that reduces in size and increases in 1-year follow-up, n/n (%) 26/171 1 (4.5) 0.323# signal in the T1-w/T2-w image. The blue arrows show an example of a new (15.2) 3 years follow-up, n/n (%) 37/145 NA lesion (in B) and its corresponding pre-lesional tissue (in A). Overlaid Masks of (25.5) all lesions: red (in C) represents baseline lesion tissue co-registered to follow-up, New lesions (12 months), n (%) 59 (34.5) 4 (18.2) 0.007 where evolution lesion tissue is also colored in red (in D); blue represents new Patients on DMT, n (%) 52 (30.4) 22 (1 0 0) <0.001 lesion tissue in D and pre-lesional tissue in C. (For interpretation of the refer­ ences to colour in this figure legend, the reader is referred to the web version of Notes: Data is presented as mean ± SD (standard deviation), median (range) or n this article.) (%). Age in years at baseline MRI. Disease duration means time from the first relapse to baseline MRI. Follow-up time means time from first assessment to the last visit in CIS cohort, and in MS cohort the time from baseline assessment at considered by the corresponding author. DMT starting to the last visit. CIS = clinical isolated syndrome; MS = multiple sclerosis; EDSS = Expanded Disability Status Scale; OB = oligoclonal bands; 5. Results MRI = magnetic resonance imaging; DMT = disease-modifying treatment. Chi- square test; # Fisher’s exact test; * Student’s t-test; £ Mann-Whitney’s test. 5.1. Demographic, clinical and radiological characteristics P < 0.001) and follow-up (cortical GM: 0.97 ± 0.08 vs deep GM 0,99 ± A total of 171 CIS patients and 22 treated MS patients were included 0,09, P < 0.001) in the CIS cohort. In the lesional tissue, a significantly in the study. All clinical, demographic and MRI data is summarized in higher percentage reduction in T2-w compared with the increase in T1- Table 1. w signal intensity was observed (percentual change in T2-w: − 2.89% ± 10.96 vs T1-w: +1.70% ± 4.03; P < 0.001) in CIS patients. 5.2. T1-w/T2-w In NAGM, NAWM and lesional tissue In the MS cohort, no longitudinal changes in T1-w/T2-w values in NAGM and NAWM were observed after 12 months of therapy: NAGM In the entire CIS cohort with valid data at baseline and follow-up (168 (baseline vs follow-up: 0.97 ± 0.13 vs 0.93 ± 0.1; P = 0.13) and NAWM patients), no significant longitudinal T1-w/T2-w changes in NAGM (1.49 ± 0.21 vs 1.44 ± 0.19; P = 0.24). T1-w/T2-w in baseline lesional (0.96 ± 0.08 vs 0.97 ± 0.08, P = 0.64) and NAWM (1.52 ± 0.16 vs 1.53 tissue was lower compared to baseline NAGM (0.78 ± 0.14 vs 0.96 ± ± 0.16, P = 0.69) were observed. In the subgroup of patients with 0.12, P < 0.001) and NAWM (0.78 ± 0.14 vs 1.47 ± 0.21, P < 0.001). abnormal MRI at baseline (117 patients), T1-w/T2-w in baseline lesional Baseline lesional tissue T1-w/T2-w was not different from evolved tissue was lower compared to T1-w/T2-w ratio in baseline NAWM (0.85 lesional tissue (0.78 ± 0.14 vs 0.80 ± 0.17, P = 0.67). ± 0.1 vs 1.52 ± 0.15, P < 0.001) and NAGM (0.85 ± 0.1 vs 0.97 ± 0.08, P < 0.001). T1-w/T2-w ratio in evolved lesional tissue was significantly higher (+7.6%) than baseline lesional tissue (0.91 ± 0.15 vs 0.85 ± 0.1, 5.3. T1-w/T2-w And clinical outcomes P < 0.001). In the subgroup of patients with new lesions (52 patients), T1-w/T2-w in new lesions tissue was lower than in pre-lesional tissue In the CIS cohort at baseline, a higher NAGM T1-w/T2-w ratio was (-28.2%, 1.02 ± 0.24 vs 1.42 ± 0.24, P < 0.001) (Fig. 3). Interestingly, observed in patients with an abnormal brain MRI, positive OB and T1-w/T2-w in pre-lesional tissue was significantly lower than baseline fulfilment of 2017 McDonald criteria. In NAWM, only a trend for a NAWM (-7.8%, 1.42 ± 0.25 vs 1.54 ± 0.18, P < 0.001) (Fig. 3) but higher T1-w/T2-w was observed in patients fulfilling 2017 McDonald higher than baseline lesional tissue (+65.1%, 1.42 ± 0.25 vs 0.86 ± 0.1, criteria (Table 2). Over follow-up, larger decreases in T1-w/T2-w values P < 0.001) (Fig. 3). in both NAGM and NAWM were observed in patients fulfilling 2017 T1-w/T2-w values were higher in deep GM as compared to cortical McDonald criteria, presence of OB, and an abnormal brain MRI at GM, both at baseline (cortical GM: 0.96 ± 0.08 vs deep GM 0,99 ± 0,09, baseline, although these differences did not reach statistical 4 M. Boaventura et al. NeuroImage: Clinical 34 (2022) 102967 Table 2 T1w/T2-w and clinical outcomes in the CIS cohort. NAGM T1-w/ p value NAWM T1-w/ p T2-w T2-w value Baseline Brain MRI Abormal 0.97 (0.01) <0.001 1.51 (0.01) 0.729 Normal 0.92 (0.01) 1.50 (0.02) Oligoclonal Bands Positive 0.97 (0.01) 0.013 1.52 (0.02) 0.263 Negative 0.94 (0.01) 1.50 (0.02) 2017 McDonald criteria Fulfillment 0.98 (0.01) 0.001 1.53 (0.02) 0.080 Non-fulfillment 0.94 (0.01) 1.49 (0.01) Change over follow- up Brain MRI Abormal − 0.18 (0.58) 0.061 0.08 (0.63) 0.249 Normal 1.78 (0.86) 1.38 (0.93) Oligoclonal Bands Positive − 0.56 (0.71) 0.137 − 0.40 (0.72) 0.261 Negative 0.98 (0.75) 0.79 (0.77) 2017 McDonald criteria Fulfillment − 0.39 (0.80) 0.180 − 0.32 (0.87) 0.248 Non-fulfillment 0.97 (0.61) 0.94 (0.66) Fig. 3. T1-w/T2-w mean values in different tissues and time points in the New lesions at 12 m CIS cohort. Multigroup comparisons with repeated measures analysis of vari­ Yes − 1.04 (0.79) 0.019 − 0.85 (0.86) 0.048 ance (ANOVA), followed by Bonferroni multiple comparisons. All masks in this No 1.35 (0.62) 1.32 (0.67) New relapses at 12 graph showed a statistically significant difference among them (P < 0.001). m NAWM = normal-appearing white matter. Yes − 2.54 (1.17) 0.005 − 2.00 (1.25) 0.026 No 1.08 (0.52) 1.08 (0.56) significance. Patients presenting disease activity during the follow-up New relapses at 36 (new T2-w lesions at 12 months MRI, new relapses one and three m Yes 1.69 (1.02) 0.016 − 1.73 (1.08) 0.032 years after CIS) showed a decrease in T1-w/T2-w values in NAGM and No 1.25 (0.64) 1.05 (0.68) NAWM, while in stable patients an increase in T1-w/T2-w values was observed (Table 2). For cortical and deep GM, these same associations Notes: Data is presented as mean/SD (standard deviation) and changes as per­ centage. The associations were assessed by using general linear model (adjusted are demonstrated in Supplementary Table 1. Longitudinal changes in for age and sex). Differences were considered statistically significant at P < 0.05. T1-w/T2-w values in lesional tissue were not associated with clinical Abbreviations: CIS = clinical isolated syndrome; NAGM = normal-appearing activity during the follow-up (data not shown). After age and sex grey matter; MRI = magnetic resonance imaging; OB = oligoclonal bands; adjustment, no significant correlations were observed between baseline NAWM = normal-appearing white matter. EDSS and baseline T1-w/T2-w values in NAGM (r = 0.015, P = 0.87), NAWM (r = 0.032, P = 0.74) or baseline lesional tissue (r = 0.094, P = In both cohorts, cross-sectional analysis at baseline revealed lower 0.32). T1-w/T2-w values in lesional tissue compared to normal-appearing tis­ In the MS cohort, after adjusting for age and sex, disease duration and sue, what may probably reflect differences in both myelin concentration baseline EDSS were inversely correlated with baseline T1-w/T2-w in and tissue integrity. This finding is in line with histopathological evi­ NAGM (disease duration: r = -0.557, P = 0.031; and; baseline EDSS: r = dence showing that the pathological abnormalities affecting lesions -0.505, P = 0.044) and NAWM (disease duration: r = -0.522, P = 0.046; consist of major demyelination and severe tissue destruction (Bitsch baseline EDSS: r = -0.517, P = 0.049). et al., 2001; Barkhof et al., 2003; Laule et al., 2006). In contrast, ab­ Partial correlations between T1-w/T2-w at the different compart­ normalities affecting normal-appearing tissue mainly consist of axonal ments and clinical variables (EDSS, disease duration and lesion volume) damage and loss and microglial activation, (Kutzelnigg et al., 2005; are presented as supplementary data (Supplementary Table 2), for both Howell et al., 2010) but do not include major demyelination. In other cohorts and time-points. words, focal damage in lesional tissue may have a higher myelin-specific nature, which could contribute to the main pathological substrate of the 6. Discussion lowest T1-w/T2-w values found in our study, while the normal- appearing tissue show a non-myelin-specific nature of damage. This In this study, we investigated changes in T1-w/T2-w within NAGM, gradient is in line with a study that demonstrated a significant reduction NAWM, and focal lesional tissue in CIS and MS patients. To the best of of T1-w/T2-w ratio in cortical and WM lesions, compared with homol­ our knowledge, this is the first study to investigate longitudinal associ­ ogous non-lesioned areas (Granberg et al., 20172017). ations with clinical outcomes concerning T1-w/T2-w at all tissue levels. In the CIS cohort, a significant increase in T1-w/T2-w comparing In CIS patients, we found that T1-w/T2-w changes occurring in lesions evolved lesional tissue with baseline lesional tissue was observed. This seem to indicate tissue damage and recovery and that alterations in T1- same finding was not observed in the MS cohort, where T1-w/T2-w w/T2-w precede lesion formation. In NAGM and NAWM, discordant values in lesional tissue did not change over follow-up. Besides edema associations have been found in the CIS cohort: higher baseline T1-w/T2- resolution (probably driving more T2-w signal decreasing), we assume w values were associated with a diagnosis of MS but, in contrast, lon­ that increasing T1-w/T2-w values in CIS lesions depend, to some extent, gitudinal decreases were associated with worse outcomes. In established on the level of remyelination in the lesion itself and might somehow MS, lower T1-w/T2-w ratio in NAGM and NAWM was associated with indicate tissue recovery. The mechanisms and targets of demyelination longer disease duration and higher EDSS and remained unchanged after in MS may be different in the distinct subgroups or stages of the disease, one year of therapy. 5 M. Boaventura et al. NeuroImage: Clinical 34 (2022) 102967 (Lucchinetti et al., 2000) or in the different types of lesions (Faizy et al., Concerning T1-w/T2-w values in cortical and deep GM, ferritin iron 2016). An important difference between our cohorts is disease duration. has a strong effect on decreasing T2-w signal, but only a weak effect on A study showed that remyelination is a frequent event in early MS increasing T1-w signal, (Vymazal et al., 1995) leading to an slight in­ stages, but in chronic lesions, the extent of remyelination was limited in crease in T1-w/T2-w. This could explain the higher T1-w/T2-w in deep most patients (Goldschmidt et al., 2009). Thus, myelin dynamics in MS GM compared to cortical GM, as the basal ganglia have been shown to depends on individual remyelination potential, justifying differences accumulate iron. between the cohorts studied. Therapy effects are other potential vari­ Compared to previous studies, we could highlight some positive as­ ables for changes in ratio, which are not evaluated in this study. pects of ours, namely the first complete T1-w/T2-w brain tissue analysis Interestingly, we have showed that T1-w/T2-w ratio in pre-lesional of a large CIS cohort with prospective clinical and radiological evalua­ tissue was already decreased when compared to baseline NAWM. tion. Concerning our methodology, some limitations should be noted. These results suggest that changes in NAWM of patients with MS occur We did not include healthy controls for comparison. Although T1-w/T2- before lesions become evident on conventional MRI scans. Comparable w has usually been applied to high resolution 3D T1-w and 3D T2-w findings have been identified in previous studies with other methods, sequences, we used 2D T2-w sequences with 3 mm slice thickness, such as MRI frequency shifts using susceptibility-weighted imaging which may have introduced a partial volume effect in our T1-w/T2-w. (Wiggermann et al., 2013); magnetization transfer ratio (MTR) (Filippi Cerebral atrophy may also influence tissue segmentation, inducing et al., 1998; Goodkin et al., 1998; Pike et al., 2000) and proton MR partial volume effects. However, the short duration of disease in the CIS spectroscopy (Tartaglia et al., 2002). cohort could mitigate this effect, due to minor atrophy observed in this We have found unexpected associations in NAGM and NAWM in CIS phase. Another limitation is the lack of B1 + residual correction of the patients: higher baseline T1-w/T2-w values were associated with a T1-w/T2-w maps, something that is not yet widely available, but means diagnosis of MS but, conversely, longitudinal decreases were associated that it might be wise to regress measures of head size (i.e. all head tis­ with concurrent disease activity. Hypercellular content state could sues, not just brain size in the subject’s own physical space) and body justify the unexpected higher T1/T2 ratio (Blystad et al., 2016; Thaler mass index (BMI) out of any statistical analyses as covariates of no in­ et al., 2017) in pathological CIS patients, decreasing with inflammation terest. We have considered age and sex. But it would be very interesting resolution. These findings also show the low specificity of this marker. to investigate the role of BMI in future studies. An additional assessment Two studies (Righart et al., 2017; Nakamura et al., 2017) demonstrated that might be included is the cortical lesional tissue, using MRI se­ correlations between reduced values and tissue damage, although in quences as double inversion recovery and phase sensitive inversion re­ different topographies and pathological substrates: in demyelinated covery, to evaluate T1-w/T2-w in cortical lesion GM. Nevertheless, the cortex correlating with myelin density, and in cortical NAGM correlating total cortical lesion volume is much lower than the total volume of with dendrite density. Still on histological validation at the cortical cortical NAGM, thus we may not expect a change in the mean value. level, despite a good sensitivity to detect cortical demyelination, a In conclusion, no longitudinal changes in T1-w/T2-w values in modest T1-w/T2-w ratio specificity was demonstrated in a recent work NAGM and NAWM were observed over 1 year in both cohorts. In CIS (Zheng et al., 2022). In addition to the cerebral cortex, in two other patients, T1-w/T2-w changes occurring in lesional tissue seem to indi­ studies focusing on different subcortical regions, (Arshad et al., 2017; cate tissue damage and recovery with alterations in T1-w/T2-w ratio Uddin et al., 2019) T1-w/T2-w ratio has been suggested as a general predating lesion formation. This finding can be adressed in future measure of microstructure, which may be more affected by axonal studies, with potential clinical applications in predicting the appearance diameter/density than myelin density. These studies used myelin water of new lesions. In NAWM and NAGM, higher baseline T1-w/T2-w values fraction (MWF) as a validated measure of myelin density, (Laule et al., in CIS patients were associated with baseline MS diagnosis, whereas 2006) showing low concordance between T1-w/T2-w and MWF mea­ longitudinal decreases in T1w/T2w ratio were associated with concur­ surements in cerebral WM and subcortical GM structures. On this basis, rent disease activity. In established MS, lower T1-w/T2-w was associ­ these works have suggested that T1-w/T2-w and MWFs appear to be ated with worse clinical outcomes and disease duration. The differential sensitized to different subcortical microstructural properties. Recently, impact of chronic inflammation, iron deposition and demyelination (Pareto et al., 2020) our group demonstrated moderate correlation be­ should be considered to interpret all these findings. Future research tween MTR and T1-w/T2-w in NAGM, and strong correlation in NAWM should take confounding effects such as iron and inflammation into and lesions. This study suggests again that, besides myelin integrity, account, through multimodal MRI acquisitions and post-mortem studies, other factors may be playing a role in T1-w/T2-w measures. to quantify T1-w/T2-w in healthy and pathological tissue. More recently, three different articles also showed higher T1-w/T2- w values in different CNS regions in other disorders: Parkinson disease Disclosures (Du et al., 2019), Huntington disease (Rowley et al., 2018) and in Alz­ heimer’s patients (Pelkmans et al., 2019). Such differences clearly sug­ M. Boaventura received a grant as ECTRIMS- Clinical Training gest that T1-w/T2-w is not just measuring myelin, and that other factors fellow and has received a speaker honorarium and/or travel expenses for such as iron content, microglia activation and amyloid deposition may scientific meetings from Biogene, Merck and Roche. be playing a role. Certainly, in MS and very early CIS patients, other J. Sastre-Garriga: over the last 12 months, J. Sastre-Garriga has processes such as edema resolution, inflammation, iron uptake and engaged in consulting and/or participating as speaker/chair in events removal, and axonal/dendrite diameter and density may influence these organized by Merck, Bayer, Celgene, Sanofi and Biogen, is director of values. Revista de Neurología and editor for controversies of the Multiple In the MS cohort, T1-w/T2-w values were inversely correlated with Sclerosis Journal. disease duration and EDSS. Previous non-longitudinal studies showed A. Garcia-Vidal has nothing to disclose. different results concerning clinical associations between T1-w/T2-w A. Vidal-Jordana receives support for contracts Juan Rodes (JR16/ and EDSS. In line with our results, T1-w/T2-w values in NAWM were 00024) from Fondo de Investigaciones Sanitarias, Instituto de Salud inversely correlated with EDSS in a study of 244 MS patients (Beer et al., Carlos III, Spain; and has received speaking honoraria and travel ex­ 2016). In contrast, another work did not show any significant associa­ penses from Novartis, Roche, Mylan, Teva, Biogen and Sanofi. tions between T1-w/T2-w and EDSS, although the sample size was much D. Quartana has received travel expenses for scientific meetings smaller (26 patients) (Granberg et al., 20172017). A recent longitudinal from Biogene and Sanofi. work (Cooper et al., 2021) in CIS and early MS patients showed that T1- R. Carvajal has received speaking honoraria and/or travel expenses w/T2-w in NAWM is associated with increasing lesion volume and dis­ for scientific meetings from Roche, Merck, Novartis and Sanofi. ease activity in the first 2 years of disease after first clinical presentation. C. Auger has received speaking honoraria from Novartis and Biogen. 6 M. Boaventura et al. NeuroImage: Clinical 34 (2022) 102967 M. Alberich has been sponsored by Novartis Farmacéutica S.A., 2019. Association between pathological and MRI findings in multiple sclerosis. Lancet Neurol. 18 (2), 198–210. Barcelona, Spain. Wattjes, M.P., Rovira, À., Miller, D., et al., 2015. Evidence-based guidelines: MAGNIMS M. Tintoré has received compensation for consulting services and consensus guidelines on the use of MRI in multiple sclerosis–establishing disease speaking honoraria from Almirall, Bayer Schering Pharma, Biogen-Idec, prognosis and monitoring patients. Nat. Rev. Neurol. 11, 597–606. Genzyme, Merck-Serono, Novartis, Roche, Sanofi-Aventis, and Teva Enzinger, C., Barkhof, F., Ciccarelli, O., Filippi, M., Kappos, L., Rocca, M.A., Ropele, S., Rovira, À., Schneider, T., de Stefano, N., Vrenken, H., Wheeler-Kingshott, C., Pharmaceuticals. MT is co-editor of Multiple Sclerosis Journal-ETC. Wuerfel, J., Fazekas, F., 2015. Nonconventional MRI and microstructural cerebral A. Rovira serves on scientific advisory boards for Novartis, Sanofi- changes in multiple sclerosis. Nat. Rev. Neurol. 11 (12), 676–686. Genzyme, Icometrix, SyntheticMR, Bayer, Biogen and OLEA Medical, Glasser, M.F., Van Essen, D.C., 2011. Mapping human cortical areas in vivo based on myelin content as revealed by T1- and T2-weighted MRI. J. Neurosci. 31 (32), and has received speaker honoraria from Bayer, Sanofi-Genzyme, 11597–11616. Bracco, Merck-Serono, Teva Pharmaceutical Industries Ltd, Novartis, Glasser, M.F., Goyal, M.S., Preuss, T.M., Raichle, M.E., Van Essen, D.C., 2014. Trends and Roche and Biogen. properties of human cerebral cortex: Correlations with cortical myelin content. Neuroimage 93, 165–175. X. Montalban has received speaking honoraria and travel expenses Shams, Z., Norris, D.G., Marques, J.P., Lundberg, P., 2019. A comparison of in vivo MRI for participation in scientific meetings, has been a steering committee based cortical myelin mapping using T1w/T2w and R1 mapping at 3T. PLoS One 14 member of clinical trials or participated in advisory boards of clinical (7), e0218089. https://doi.org/10.1371/journal.pone.0218089. Righart, R., Biberacher, V., Jonkman, L.E., Klaver, R., Schmidt, P., Buck, D., Berthele, A., trials in the past years with Actelion, Bayer, Biogen, Celgene, Genzyme, Kirschke, J.S., Zimmer, C., Hemmer, B., Geurts, J.J.G., Mühlau, M., 2017. Cortical Merck, Novartis, Roche, Sanofi-Genzyme, Teva Pharmaceutical, pathology in multiple sclerosis detected by the T1/T2-weighted ratio from routine Excemed, MSIF and NMSS. magnetic resonance imaging. Ann. Neurol. 82 (4), 519–529. Arshad, M., Stanley, J.A., Raz, N., 2017. Test-retest reliability and concurrent validity of D. Pareto has received speaking honoraria from Novartis and Sanofi- in vivo myelin content indices: Myelin water fraction and calibrated T 1 w/T 2 w Genzyme and a research grant from Biogen Idec. image ratio. Hum. Brain Mapp. 38 (4), 1780–1790. Uddin, M.N., Figley, T.D., Solar, K.G., Shatil, A.S., Figley, C.R., 2019. Comparisons between multi-component myelin water fraction, T1w/T2w ratio, and diffusion Funding tensor imaging measures in healthy human brain structures. Sci. Rep. 9 (1) https:// doi.org/10.1038/s41598-019-39199-x. This project was developed as a part of Mateus Boaventura ECTRIMS Pareto, D., Garcia-Vidal, A., Alberich, M., Auger, C., Montalban, X., Tintoré, M., Sastre- Clinical Training Fellowship Programme 2018–2019. This study was Garriga, J., Rovira, À., 2020. Ratio of T1-weighted to T2-weighted signal intensity as a measure of tissue integrity: comparison with magnetization transfer ratio in partially supported by the Project PI18/00823, from the Fondo de patients with multiple sclerosis. Am. J. Neuroradiol. 41 (3), 461–463. Investigación Sanitaria (FIS), Instituto de Salud Carlos III. Nakamura, K., Chen, J.T., Ontaneda, D., Fox, R.J., Trapp, B.D., 2017. T1-/T2-weighted ratio differs in demyelinated cortex in multiple sclerosis. Ann. Neurol. 82 (4), 635–639. CRediT authorship contribution statement Beer, A., Biberacher, V., Schmidt, P., Righart, R., Buck, D., Berthele, A., Kirschke, J., Zimmer, C., Hemmer, B., Mühlau, M., 2016. Tissue damage within normal appearing white matter in early multiple sclerosis: assessment by the ratio of T1- and T2- Mateus Boaventura: Conceptualization, Methodology, Data cura­ weighted MR image intensity. J. Neurol. 263 (8), 1495–1502. tion, Writing – original draft, Writing – review & editing. Jaume Sastre- Cooper G, Finke C, Chien C, et al. Standardization of T1w/T2w Ratio Improves Detection Garriga: Supervision, Project administration, Conceptualization, of Tissue Damage in Multiple Sclerosis. Front Neurol; 10. Epub ahead of print 9 April 2019. DOI: 10.3389/fneur.2019.00334. Methodology, Writing – original draft, Writing – review & editing. Aran Tintore, M., Rovira, À., Río, J., Otero-Romero, S., Arrambide, G., Tur, C., Comabella, M., Garcia-Vidal: Methodology, Writing – original draft, Writing – review Nos, C., Arévalo, M.J., Negrotto, L., Galán, I., Vidal-Jordana, A., Castilló, J., & editing. Angela Vidal-Jordana: Writing – original draft, Writing – Palavra, F., Simon, E., Mitjana, R., Auger, C., Sastre-Garriga, J., Montalban, X., 2015. review & editing. Davide Quartana: Methodology, Data curation. René Defining high, medium and low impact prognostic factors for developing multiple sclerosis. Brain 138 (7), 1863–1874. Carvajal: Methodology, Data curation. Cristina Auger: Methodology, McDonald, W.I., Compston, A., Edan, G., Goodkin, D., Hartung, H.-P., Lublin, F.D., Data curation, Writing – review & editing. Manel Alberich: Method­ McFarland, H.F., Paty, D.W., Polman, C.H., Reingold, S.C., Sandberg-Wollheim, M., ology, Data curation. Mar Tintoré: Methodology, Writing – original Sibley, W., Thompson, A., Van Den Noort, S., Weinshenker, B.Y., Wolinsky, J.S., 2001. Recommended diagnostic criteria for multiple sclerosis: Guidelines from the draft, Writing – review & editing. Àlex Rovira: Supervision, Project international panel on the diagnosis of multiple sclerosis. Ann. Neurol. 50 (1), administration, Conceptualization, Methodology, Writing – review & 121–127. editing. Xavier Montalban: Supervision, Project administration, Thompson, A.J., Banwell, B.L., Barkhof, F., Carroll, W.M., Coetzee, T., Comi, G., Correale, J., Fazekas, F., Filippi, M., Freedman, M.S., Fujihara, K., Galetta, S.L., Writing – review & editing. Deborah Pareto: Supervision, Project Hartung, H.P., Kappos, L., Lublin, F.D., Marrie, R.A., Miller, A.E., Miller, D.H., administration, Conceptualization, Methodology, Writing – original Montalban, X., Mowry, E.M., Sorensen, P.S., Tintoré, M., Traboulsee, A.L., draft, Writing – review & editing. Trojano, M., Uitdehaag, B.M.J., Vukusic, S., Waubant, E., Weinshenker, B.G., Reingold, S.C., Cohen, J.A., 2018. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria. Lancet Neurol. 17 (2), 162–173. Declaration of Competing Interest Filippi M, Preziosa P, Banwell BL, et al. Assessment of lesions on magnetic resonance imaging in multiple sclerosis: practical guidelines. Brain 2019; 142: 1858–1875. Ganzetti M, Wenderoth N, Mantini D. Whole brain myelin mapping using T1- and T2- The authors declare that they have no known competing financial weighted MR imaging data. Front Hum Neurosci; 8. Epub ahead of print 2 September interests or personal relationships that could have appeared to influence 2014. DOI: 10.3389/fnhum.2014.00671. the work reported in this paper. Patenaude, B., Smith, S.M., Kennedy, D.N., Jenkinson, M., 2011. A Bayesian model of shape and appearance for subcortical brain segmentation. Neuroimage 56 (3), 907–922. Acknowledgements Bitsch, A., Kuhlmann, T., Stadelmann, C., Lassmann, H., Lucchinetti, C., Brück, W., 2001. A longitudinal MRI study of histopathologically defined hypointense multiple sclerosis lesions. Ann. Neurol. 49 (6), 793–796. The authors would like to thank Ewerton da Silva Lima for English Barkhof, F., Brück, W., De Groot, C.J.A., Bergers, E., Hulshof, S., Geurts, J., Polman, C.H., writing support. van der Valk, P., 2003. Remyelinated lesions in multiple sclerosis. Arch. Neurol. 60 (8), 1073. https://doi.org/10.1001/archneur.60.8.1073. Laule, C., Leung, E., Li, D.KB., Traboulsee, A.L., Paty, D.W., MacKay, A.L., Moore, G.RW., Appendix A. Supplementary data 2006. Myelin water imaging in multiple sclerosis: quantitative correlations with histopathology. Mult. Scler. J. 12 (6), 747–753. Supplementary data to this article can be found online at https://doi. Kutzelnigg A, Lucchinetti CF, Stadelmann C, et al. Cortical demyelination and diffuse white matter injury in multiple sclerosis. Brain 2005; 128: 2705–2712. org/10.1016/j.nicl.2022.102967. Howell, O.W., Rundle, J.L., Garg, A., Komada, M., Brophy, P.J., Reynolds, R., 2010. Activated microglia mediate axoglial disruption that contributes to axonal injury in References multiple sclerosis. J. Neuropathol. Exp. Neurol. 69 (10), 1017–1033. Granberg T, Fan Q, Treaba CA, et al. In vivo characterization of cortical and white matter neuroaxonal pathology in early multiple sclerosis. Brain 2017; 140: 2912–2926. Filippi, M., Brück, W., Chard, D., Fazekas, F., Geurts, J.J.G., Enzinger, C., Hametner, S., Kuhlmann, T., Preziosa, P., Rovira, À., Schmierer, K., Stadelmann, C., Rocca, M.A., 7 M. Boaventura et al. NeuroImage: Clinical 34 (2022) 102967 Lucchinetti C, Brck W, Parisi J, et al. Heterogeneity of multiple sclerosis lesions: Blystad, I., Håkansson, I., Tisell, A., Ernerudh, J., Smedby, Ö., Lundberg, P., Larsson, E.- Implications for the pathogenesis of demyelination. Ann Neurol 2000; 47: 707–717. M., 2016. Quantitative MRI for analysis of active multiple sclerosis lesions without Faizy, T.D., Thaler, C., Kumar, D., Sedlacik, J., Broocks, G., Grosser, M., Stellmann, J.-P., gadolinium-based contrast agent. Am. J. Neuroradiol. 37 (1), 94–100. Heesen, C., Fiehler, J., Siemonsen, S., Cercignani, M., 2016. Heterogeneity of Thaler, C., Faizy, T.D., Sedlacik, J., Holst, B., Stürner, K., Heesen, C., Stellmann, J.-P., multiple sclerosis lesions in multislice myelin water imaging. PLoS One 11 (3), Fiehler, J., Siemonsen, S., 2017. T1 recovery is predominantly found in black holes e0151496. https://doi.org/10.1371/journal.pone.0151496. and is associated with clinical improvement in patients with multiple sclerosis. Am. Goldschmidt, T., Antel, J., Konig, F.B., Bruck, W., Kuhlmann, T., 2009. Remyelination J. Neuroradiol. 38 (2), 264–269. capacity of the MS brain decreases with disease chronicity. Neurology 72 (22), Zheng, Y., Dudman, J., Chen, J.T., Mahajan, K.R., Herman, D., Fox, R.J., Ontaneda, D., 1914–1921. Trapp, B.D., Nakamura, K., 2022. Sensitivity of T1/T2-weighted ratio in detection of Wiggermann, V., Hernandez Torres, E., Vavasour, I.M., Moore, G.R.W., Laule, C., cortical demyelination is similar to magnetization transfer ratio using post-mortem MacKay, A.L., Li, D.K.B., Traboulsee, A., Rauscher, A., 2013. Magnetic resonance MRI. Mult. Scler. J. 28 (2), 198–205. frequency shifts during acute MS lesion formation. Neurology 81 (3), 211–218. Du, G., Lewis, M.M., Sica, C., Kong, L., Huang, X., 2019. Magnetic resonance T1w/T2w Filippi, M., Rocca, M.A., Martino, G., Horsfield, M.A., Comi, G., 1998. Magnetization ratio: a parsimonious marker for Parkinson disease. Ann. Neurol. 85 (1), 96–104. transfer changes in the normal appering white matter precede the appearance of Rowley CD, Tabrizi SJ, Scahill RI, et al. Altered Intracortical T1-Weighted/T2-Weighted enhancing lesions in patients with multiple sclerosis. Ann. Neurol. 43 (6), 809–814. Ratio Signal in Huntington’s Disease. Front Neurosci; 12. Epub ahead of print 5 Goodkin, D.E., Rooney, W.D., Sloan, R., Bacchetti, P., Gee, L., Vermathen, M., November 2018. DOI: 10.3389/fnins.2018.00805. Waubant, E., Abundo, M., Majumdar, S., Nelson, S., Weiner, M.W., 1998. A serial Pelkmans, W., Dicks, E., Barkhof, F., Vrenken, H., Scheltens, P., Flier, W.M., Tijms, B.M., study of new MS lesions and the white matter from which they arise. Neurology 51 2019. Gray matter T1-w/T2-w ratios are higher in Alzheimer’s disease. Hum. Brain (6), 1689–1697. Mapp. 40 (13), 3900–3909. Pike, G.B., De Stefano, N., Narayanan, S., Worsley, K.J., Pelletier, D., Francis, G.S., Cooper, G., Chien, C., Zimmermann, H., Bellmann-Strobl, J., Ruprecht, K., Kuchling, J., Antel, J.P., Arnold, D.L., 2000. Multiple sclerosis: magnetization transfer MR Asseyer, S., Brandt, A.U., Scheel, M., Finke, C., Paul, F., 2021. Longitudinal analysis imaging of white matter before lesion appearance on T2-weighted images. Radiology of T1w/T2w ratio in patients with multiple sclerosis from first clinical presentation. 215 (3), 824–830. Mult. Scler. J. 27 (14), 2180–2190. Tartaglia, M.C., Narayanan, S., De Stefano, N., Arnaoutelis, R., Antel, S.B., Francis, S.J., Vymazal, J., Brooks, R.A., Patronas, N., Hajek, M., Bulte, J.W.M., Di Chiro, G., 1995. Santos, A.C., Lapierre, Y., Arnold, D.L., 2002. Choline is increased in pre-lesional Magnetic resonance imaging of brain iron in health and disease. J. Neurol. Sci. 134, normal appearing white matter in multiple sclerosis. J. Neurol. 249 (10), 19–26. 1382–1390. 8