Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Mar 9;36(2):e70132. doi: 10.1111/jon.70132

Quantitative Choroid Plexus Gadolinium Enhancement Is Related to Diffuse Brain Tissue Injury in Multiple Sclerosis

Sneha Senthil 1,2, Ian Tagge 1, Dumitru Fetco 1,3, Cheng Hsun Hsieh 1, Haz‐Edine Assemlal 1,3, Zahra Karimaghaloo 1,3, Emily Fetco 3, G R Wayne Moore 2,4, Douglas L Arnold 1, David A Rudko 1,5, Sridar Narayanan 1,2,✉
PMCID: PMC12972605  PMID: 41804056

ABSTRACT

Background

Recent studies suggest that disruptions of the blood–cerebrospinal fluid (CSF) barrier within the choroid plexus (ChP) may contribute to multiple sclerosis (MS) pathogenesis. We investigated the relationship between a quantitative marker of ChP enhancement and markers of focal and diffuse brain tissue injury in MS.

Methods

A group of 34 MS participants and 21 healthy participants underwent 7T MRI including magnetization prepared 2 rapid acquisition gradient echoes (MP2RAGE) and fluid attenuated inversion recovery (FLAIR) acquisitions. The MS group received contrast, and delta T1 (ΔT1) maps were computed to assess enhancement. ChP, white matter lesions (WML), normal‐appearing white matter (NAWM), and gray matter (GM) were segmented. Pre‐contrast quantitative T1 (qT1) values were compared between groups, and linear regression with mean ChP ΔT1 was performed for WML volume and pre‐gadolinium (Gd) mean qT1 of WML, NAWM, and GM.

Results

Mean qT1 of ChP, NAWM, and GM, as well as ChP volume, were higher in MS compared to controls (p < 0.001). ChP ΔT1 was significantly associated with pre‐Gd qT1 of NAWM (β = 0.20, R2 = 0.54, p < 0.001) and GM (β = 0.18, R2 = 0.49, p < 0.001), but not WML volume (p = 0.3) or WML qT1 (p = 0.05).

Conclusions

The association between ChP enhancement and diffuse tissue injury, together with elevated qT1 values and ChP volumes in MS, supports a mechanism of brain injury involving CSF‐mediated toxicity distinct from classic lesion pathology in MS.

Keywords: blood–CSF barrier, choroid plexus, tissue damage

1. Introduction

Multiple sclerosis (MS) is a chronic demyelinating disease that results in a wide range of neurological deficits. While immune cell trafficking into the CNS through disruptions in the blood–brain barrier (BBB) is a well‐established feature of MS [1, 2, 3, 4], recent studies have implicated the choroid plexus (ChP) in MS pathogenesis [5, 6, 7]. The ChP, a highly vascularized structure within the brain's ventricular system, serves as an entry point for lymphocytes and plays a pivotal role in modulating inflammatory processes [5, 8, 9].

In addition to its immune functions, the ChP plays a vital role in the production of cerebrospinal fluid (CSF), which is essential for maintaining the homeostasis of the CNS. It forms the blood–CSF barrier (BCSFB), a selective interface of epithelial tight junctions (6), and harbors immune cells that protect against pathogens and regulate immune surveillance via chemokine signaling (8). Disruption of the BCSFB, however, can compromise CNS defense by increasing ChP permeability, allowing harmful microbes to infiltrate into the CSF [10]. In inflammatory conditions, alterations in ChP permeability allows pathogenic T‐lymphocytes to migrate through both the BBB and the BCSFB, contributing to increased infiltration of immune cells and pro‐inflammatory factors into the CNS [11]. Given the central role of neuroinflammation in MS, investigating the contribution of increased BCSFB permeability to brain tissue injury is crucial, yet remains largely unexplored.

Quantitative T1 (qT1) (longitudinal relaxation time constant) mapping on MRI offers noninvasive assessment of brain tissue microstructure. MRI‐histology studies have shown that prolonged qT1 reflects increased water content, demyelination, and axonal loss in white matter (WM), serving as a marker of tissue injury [12, 13, 14]. Previous studies have also demonstrated that qT1 correlates with tissue damage in MS, with elevated qT1 observed in WM lesions (WMLs) compared to normal‐appearing white matter (NAWM) [15], and in NAWM [16, 17, 18, 19, 20, 21, 22] and normal‐appearing gray matter (NAGM) [18, 23, 24] compared to healthy controls (HCs). Recently, the magnetization prepared 2 rapid acquisition gradient echoes (MP2RAGE) technique has been employed for qT1 measurement, providing both T1‐weighted MPRAGE‐like images and qT1 maps [25]. MP2RAGE can be acquired at high spatial resolution within clinically acceptable scan times, and is self‐bias‐field corrected, making it well‐suited for ultra‐high‐field applications (> = 7T) [25]. It has been demonstrated that a change in qT1 in tissue after administration of a gadolinium (Gd)‐based contrast agent (GBCA) can be measured through voxel‐wise subtraction of qT1 maps pre‐ and post‐Gd [26].

We aimed to use qT1 mapping to assess the relationship of T1 shortening (delta T1 [ΔT1]) in the ChP after intravenous Gd‐contrast administration, with markers of focal and diffuse brain tissue injury in MS, namely, WML volume, qT1 within WML, and qT1 in NAWM and GM. We hypothesize that greater ChP enhancement (higher ΔT1) correlates with greater tissue damage, supporting the idea that systemically circulating molecules and/or inflammatory cells entering the CNS through the ChP and diffusing throughout the brain via the CSF is a possible mechanism of tissue injury in MS. Although ΔT1 measures were available only in the MS cohort, we compared pre‐Gd qT1 values between MS participants and HCs to help assess whether these relationships reflect MS‐specific tissue injury.

2. Methods

2.1. Study Participants

We recruited 34 people with MS (relapsing‐remitting MS [RRMS] = 24, secondary progressive MS [SPMS] = 10, age = 53.2 ± 9.8, M = 7, F = 27, Expanded Disability Status Scale [EDSS] = 2.5 [0–6.5], disease duration = 22.6 ± 8.6 years), followed at the Montreal Neurological Institute‐Hospital. A convenience group of 21 HCs (age = 51.0 ± 12.6 years; M = 9, F = 12) with no history of brain disease and who were imaged with a matched protocol were also included.

Inclusion criteria for MS participants were diagnosis based on the 2017 revised McDonald criteria [27], and being either untreated or on a stable disease modifying therapy (DMT) for ≥ 6 months. Exclusion criteria were concomitant neurological disease in addition to MS, a history of claustrophobia, and contraindications to MRI. All study procedures were approved by the hospital's Research Ethics Board and participants gave written consent.

2.2. MRI Acquisition

Participants were scanned on a 7T whole‐body MR scanner (MAGNETOM Terra, Siemens Healthineers, Erlangen, Germany) with an 8 Tx/32 Rx‐channel head coil (Nova Medical Inc., Massachusetts, USA) in parallel transmission mode. The imaging protocol included an MP2RAGE acquisition (resolution = 0.7 mm [3] isotropic, FOV = 240 × 240 × 172 mm3, TI1/TI2 = 800 ms/2700 ms, TR/TE = 6000 ms/2.74 ms, acquisition time = 10:14 min). MS participants also underwent MP2RAGE 8 min post‐Gd injection. A T1 map was computed using the Siemens MapIt software for each MP2RAGE acquisition. A 3D T2‐weighted fluid attenuated inversion recovery (FLAIR) sequence was acquired 20 min post‐Gd injection in MS participants (resolution = 0.7 mm [3] isotropic, FOV = 225 × 225 × 179.2 mm3, TR/TE = 10560 ms/301 ms, TI = 2500 ms, acquisition time = 9:32).

2.3. Image Processing

All images were processed using the MINC toolkit (v.2, Montreal Neurological Institute, Montreal, Quebec, Canada; https://bic‐mni.github.io/) [28]. ​​ MRI images for each participant were individually co‐registered to their native pre‐Gd MP2RAGE‐T1w space. Intensity inhomogeneities in the FLAIR images were corrected by the N4 bias field algorithm [29]. ΔT1 maps were calculated by voxel‐by‐voxel subtraction of post‐contrast from pre‐contrast T1 maps of each participant.

2.4. Tissue and Lesion Segmentations

Manual segmentation of the ChP (on pre‐Gd‐T1‐weighted images, with pre‐Gd INV2 and post‐Gd FLAIR images used as additional anatomical reference to improve tissue boundary delineation) was performed by a trained rater (S.S., > 3 years of experience) using the Display software (v2.0, Montreal Neurological Institute, Montreal, Quebec, Canada;http://www.bic.mni.mcgill.ca/software/Display). A semi‐manual thresholding approach was applied on the INV2 images to facilitate the identification of ChP tissue based on its characteristic signal intensity relative to surrounding CSF, followed by manual refinement. Intrarater variability of ChP segmentation was low, with a coefficient of variation of 0.93% across repeated segmentations performed 6 months apart. For the MS cohort, T2‐hyperintense WMLs were manually segmented on FLAIR images by an experienced medical image analyst (D.F., > 20 years of experience). WML volume was calculated for each MS participant using MINC tools.

Tissue segmentation was performed using the Brain Tissue Composition‐Net (BTC‐Net) pipeline developed at NeuroRx. This pipeline employs a 3D convolutional neural network, inspired by the U‐Net architecture [30] and trained on co‐registered MRI scans at 1 mm3 isotropic resolution, to generate tissue probability maps as well as binary tissue class masks. All MP2RAGE inputs were resampled from 0.7 mm3 to 1 mm3 for tissue segmentation, to match the model architecture. Deep and cortical GM were treated as distinct tissue classes by BTC‐Net. Rather than use the standard segmentation labels output by BTC‐Net, which would be appropriate for volume estimations, we started with the probabilistic tissue maps output by BTC‐Net and applied tissue‐specific probability thresholds to generate high‐confidence tissue masks. These thresholds were empirically determined through visual assessment for each tissue class, balancing two competing requirements: (i) reduction of partial‐volume contamination at tissue interfaces and (ii) preservation of adequate spatial coverage for reliable regional measurements. We retained high‐confidence tissue voxels thresholded at ≥ 93% probability for cortical GM, ≥ 95% probability for deep GM, and ≥ 90% probability for WM. NAWM masks were generated by explicitly masking out the WML mask from the BTC‐Net generated tissue masks. All resulting tissue masks were manually reviewed and subsequently resampled back to each participant's native space (0.7 mm3 isotropic) prior to analysis. A representative example of tissue and lesion segmentations is illustrated in Figure 1. Mean ΔT1 of ChP was calculated by applying the ChP segmentation mask to corresponding ΔT1 maps in the MS cohort. Pre‐contrast qT1 values for ChP, GM, and NAWM were obtained by applying the corresponding masks to the pre‐Gd qT1 map in all participants. Similarly, pre‐Gd qT1 of WMLs was extracted for the MS cohort. All volumetric measures, including ChP, NAWM, deep GM, and cortical GM volumes, were normalized for interindividual head size differences using a scaling factor from skull‐based linear registration to a standard atlas.

FIGURE 1.

FIGURE 1

Open in a new tab

Representative tissue and lesion segmentations for a relapsing‐remitting multiple sclerosis (RRMS) participant included in this study. The segmentations are co‐registered to the participant's native 7T fluid‐attenuated inversion recovery (FLAIR) image space. The top row displays the pre‐gadolinium (Gd) T1‐weighted (T1‐w) image, pre‐contrast quantitative T1 (qT1) map, ΔT1 map, and FLAIR image (left to right). The bottom row displays the segmentation masks for high‐confidence normal‐appearing white matter (NAWM), high‐confidence gray matter (GM; deep GM in green and cortical GM in dark blue), choroid plexus (ChP), and white matter lesions (WML) (left to right).

2.5. Statistical Analysis

All statistical analyses were performed using R (v.4.4.0; R Foundation for Statistical Computing, Vienna, Austria). Group differences between MS and HCs of ChP volume and qT1 of ChP, NAWM, and GM were computed using an unpaired two‐tailed t‐test. In the MS cohort, differences between qT1 measures of WML and NAWM were assessed using a paired two‐tailed Student's t‐test. Separate linear regression analyses were carried out between ChP ΔT1 as the predictor variable and the following response variables, including age as covariate: 1) WML volume, 2) pre‐Gd mean qT1 of WML, 3) pre‐Gd mean qT1 of NAWM, and 4) pre‐Gd mean qT1 of GM. The ChP volume of each participant was included as a covariate in these linear models, as smaller ChP ROIs may have proportionally greater partial volume effects, potentially biasing the measured T1 change. Pearson correlation coefficients were determined to assess the strength of each of these relationships, with 95% confidence intervals provided for each analysis. The relationship between ChP permeability and additional global measures of damage, including head‐size corrected WM and GM volume, were assessed using the same linear regression framework. To evaluate whether the magnitude of ChP ΔT1 was related to the pre‐Gd qT1 of ChP, Pearson's correlation was computed. Since this relationship was being directly assessed, pre‐Gd ChP qT1 was not included as an additional covariate. To examine the relationship between ΔT1 of ChP and disease severity, as assessed by EDSS scores, a Spearman's rank correlation was conducted. Statistical significance was evaluated using a significance level of α = 0.05. p values were Bonferroni‐corrected for multiple comparisons (α = 0.05/5 = 0.01).

3. Results

3.1. Study Population

Demographic and clinical characteristics of the recruited MS cohort are shown in Table 1. The MS cohort included 24 participants with RRMS (70.6%) and 10 participants with SPMS (30.4%). Most participants (64.7%) were on DMTs at the time of their scan. The HC group did not receive contrast administration, showed no abnormalities on structural MRI, and served as a pre‐Gd comparison group for pre‐contrast qT1 analyses.

TABLE 1.

Detailed demographic information of the recruited MS participant cohort.

Characteristics Full cohort (n = 34) RRMS (n = 24, 70.6%) SPMS (n = 10, 29.4%)
Age in years 53.2 ± 9.8 51.2 ± 10.8 58.2 ± 4.2
Female (%) 27 (79.4%) 19 (79.2%) 8 (80%)
EDSS (median[min–max]) 2.5 (0–6.5) 2 (0–4.5) 5.75 (3–6.5)
Average WML volume in mm3 8174.7 ± 10730.1 5988.458 ± 9416.6 13796.4 ± 12573.6
Total number of Gd‐enhancing lesions 3 (0.08%) 2 (0.08%) 1 (0.01%)
Disease duration (years) 22.6 ± 8.6 19.7 ± 7.4 29.5 ± 7.6
On DMT (%) 22 (64.7%) 18 (75%) 4 (40%)
Fingolimod 1 (2.9%) 1 (4.2%) 0
Glatiramer acetate 4 (11.7%) 4 (16.6%) 0
Ocrelizumab 3 (8.8%) 2 (8.3%) 1 (10%)
Dimethyl fumarate 2 (6%) 2 (8.3%) 0
Cladribine 3 (8.8%) 3 (12.5%) 0
Interferon 4 (11.7%) 3 (12.5%) 1 (10%)
Natalizumab 1 (2.9%) 0 1 (10%)
Teriflunomide 3 (8.8%) 3 (12.5%) 1 (10%)
Open in a new tab

Note: All the data represent mean ± standard deviation unless otherwise indicated.

Abbreviations: DMT = disease‐modifying therapy, EDSS = expanded disability status scale, Gd = gadolinium, n = number of participants, RRMS = relapsing‐remitting MS, SPMS = secondary progressive MS, WML = white matter lesion volume.

3.2. Comparison: MS to HC

MS participants showed significantly higher ChP volume and elevated pre‐Gd qT1 values in ChP, NAWM, and GM compared to HCs (all p < 0.001). Corresponding group distributions are shown in Figure 2.

FIGURE 2.

FIGURE 2

Open in a new tab

Boxplots showing group differences in choroid plexus (ChP) volume and pre‐gadolinium (Gd) quantitative T1 (qT1) values in the ChP, normal‐appearing white matter (NAWM), and gray matter (GM) between multiple sclerosis (MS) participants and healthy controls (HCs). MS participants exhibited significantly higher values across all measures (all p < 0.001).

3.3. Comparisons and Relationships Within the MS Group

Table 2 summarizes the data obtained from ΔT1 and qT1 maps in the MS cohort. Pre‐Gd qT1 in WML was significantly higher than that of NAWM (paired t‐test, p < 0.001). No significant differences were seen between pre‐Gd qT1 of lesions of RRMS and SPMS participants.

TABLE 2.

Summary of qT1 and ΔT1 metrics obtained in the MS cohort.

Segmented mask Measure MS cohort HC cohort
ChP ΔT1 (ms) 1629.9 ± 225.2 NA
qT1 pre‐Gd (ms) 2860.9 ± 326.5 2504.4 ± 171.2
qT1 post‐Gd (ms) 1281.2 ± 177.7 NA
Volume (mm3) 2071.4 ± 887.9 1446.1 ± 407.3
WML ΔT1 (ms) 86.7 ± 28.7 NA
qT1 pre‐Gd (ms) 1974.2 ± 275.9 NA
qT1 post‐Gd (ms) 1861.8 ± 281.1 NA
Volume (mm3) 9222.1 ± 10538.7 NA
NAWM ΔT1 (ms) 38.8 ± 11.4 NA
qT1 pre‐Gd (ms) 1355.8 ± 132.2 1152.7 ± 26.2
qT1 post‐Gd (ms) 1332.0 ± 147.4 NA
GM (consisting of deep and cortical GM) ΔT1 (ms) 324.3 ± 150.8 NA
qT1 pre‐Gd (ms) 1958.1 ± 210.6 1725.6 ± 39.1
qT1 post‐Gd (ms) 1633.8 ± 157.7 NA
Deep GM ΔT1 (ms) 182 ± 110.1 NA
qT1 pre‐Gd (ms) 1547.5 ± 118.0 1733.6 ± 21.4
qT1 post‐Gd (ms) 1365.5 ± 153.5 NA
Cortical GM ΔT1 (ms) 206.5 ± 55.1 NA
qT1 pre‐Gd (ms) 1987.3 ± 218.2 1758.4 ± 44.2
qT1 post‐Gd (ms) 1652.8 ± 157.9 NA
Open in a new tab

Note: All the data represent mean ± standard deviation unless otherwise indicated. To get a positive measure of ChP enhancement, the ΔT1 map was obtained by subtracting the post‐qT1 maps from the pre‐qT1 maps.

Abbreviations: ΔT1 = change in qT1 following administration of Gd‐based contrast agent, ChP = choroid plexus, Gd = gadolinium, GM = gray matter, HC = healthy control, MS = multiple sclerosis, NA = normal‐appearing, NAWM = normal‐appearing white matter, qT1 = quantitative T1 relaxation time, WML = white matter lesion volume.

The results of separate linear regression analyses evaluating the relationships between ChP ΔT1 and markers of brain tissue injury are shown in Figures 3 and 4. We observed significant predictive relationships between ChP ΔT1 and mean pre‐Gd qT1 of both NAWM (β = 0.20, R2 = 0.54, F(2, 31) = 18.21, p < 0.001) and GM (β = 0.18, R2 = 0.49, F(2, 31) = 14.92, p < 0.001). No significant association of ChP ΔT1 with WML volume (β = −6.12, p = 0.3) or WML pre‐Gd qT1 (β = 0.31, p > 0.05) was seen. Pearson correlation analyses showed a strong positive association between ΔT1 of ChP and pre‐Gd qT1 of NAWM (r = 0.76, p < 0.001) and pre‐Gd qT1 of GM (r = 0.70, p < 0.001). In contrast, neither the correlation between ΔT1 of ChP and pre‐Gd qT1 of WML (r = 0.48, p > 0.05) nor the correlation between ΔT1 of ChP and WML volume (r = −0.24, p > 0.05) were significant. In exploratory analyses, ChP ΔT1 was not associated with total normailzed WM volume (r = 0.0, p = 0.99) but showed a moderate negative correlation with normalized GM volume (r = −0.43, p = 0.01). When GM was partitioned into deep and cortical compartments, ChP ΔT1 was not significantly correlated with normalized deep GM volume (r = −0.21, p = 0.23), but demonstrated a moderate negative correlation with normalized cortical GM volume (r = −0.43, p = 0.01). ChP ΔT1 was strongly associated with pre‐Gd qT1 (r = 0.83, p < 0.001). Finally, no significant correlation was observed between ΔT1 of ChP and EDSS (Spearman's ρ = 0.04, p = 0.822).

FIGURE 3.

FIGURE 3

Open in a new tab

Scatterplots showing relationships within the MS cohort between ΔT1 of the choroid plexus (ChP) and (A) pre‐contrast mean qT1 of normal‐appearing white matter (NAWM) (r = 0.76, p < 0.001), (B) pre‐contrast mean qT1 of gray matter (GM) (r = 0.70, p < 0.001), (C) pre‐contrast mean qT1 of white matter lesions (WML) (r = 0.48, p > 0.05), and (D) WML volume (r = −0.24, p > 0.05). The r and p values from the linear regression models are indicated in the top‐left corner of each panel. Shaded areas represent 95% confidence intervals.

FIGURE 4.

FIGURE 4

Open in a new tab

Scatterplots showing relationships within the MS cohort between ΔT1 of the choroid plexus (ChP) and (A) age‐adjusted normal‐appearing white matter (NAWM) volume (r = 0, p = 0.9987) and (B) age‐adjusted gray matter (GM) volume (r = −0.43, p = 0.01). Linear regression r and p values are indicated in the top‐left corner of each panel. Shaded areas represent 95% confidence intervals.

4. Discussion

In this study, we applied qT1 mapping to assess the relationship between the degree of ChP enhancement after GBCA injection, a measure influenced by its stromal vascular permeability, and markers of brain tissue injury in MS. Before assessing these relationships, we first confirmed that MS participants had higher ChP volume and elevated pre‐Gd qT1 in ChP, NAWM, and GM compared to controls. These findings align with prior reports of ChP enlargement [31, 32] and diffuse microstructural abnormalities in MS, which supports the presence of underlying tissue injury even outside overt lesions [31].

Pre‐Gd qT1 values were then used to assess brain tissue injury as they reflect tissue characteristics unconfounded by the presence of GBCA in parenchyma. We found that contrast‐related T1‐shortening (ΔT1) of the ChP in MS was a significant predictor of mean pre‐Gd qT1 of NAWM and GM. Individuals with greater ChP contrast uptake also exhibited more pronounced diffuse tissue abnormalities, supporting a relationship between ChP vascular–stromal changes and widespread microstructural injury in MS. In contrast, we did not observe significant correlations of ΔT1 of the ChP with WML volume nor with pre‐Gd qT1 values within WML, where the overt pathology associated with BBB breakdown and focal lesion formation is expected to predominate over more diffuse pathology [33]. Previous studies using histopathological analyses have revealed that focal WMLs primarily arise due to localized immune cell infiltration through a compromised BBB [34]. Unlike the BCSFB, which regulates CSF composition and, when disrupted, may contribute to global tissue injury, the BBB is a highly selective barrier that, when disrupted, allows the local entry of circulating immune cells [35]. This infiltration triggers a cascade of inflammatory processes, including oxidative stress, fibrinogen leakage, complement activation, and sustained microglial activation, all of which drive lesion formation. However, it is also important to consider that WMLs represent a more advanced stage of pathology than the abnormalities seen in NAWM. Prior to lesion formation, the tissue in these regions may have undergone injury similar to NAWM, suggesting that BCSFB dysfunction could have played a contributing role in the early stages of lesion development. This would be consistent with our prior findings that regions of NAWM destined to become lesions exhibit abnormal estimates of myelin volume fraction and axon volume fraction on quantitative MRI [36]. Given that there is bulk flow of CSF through the brain parenchyma, toxic factors released into the CSF may contribute not only to diffuse pathology of the NAWM, but also plant seeds of injury that later result in the formation of focal lesions [37].

In order to contextualize our findings, it is essential to explore the potential underlying mechanisms by which the BCSFB integrity can be compromised. The BCSFB is characterized by the presence of fenestrated capillaries within the ChP, where the true “barrier” function is performed by the ChP epithelium. This barrier is formed by tight junctions between adjacent epithelial cells, composed of key proteins such as occludin, claudins (including claudin‐1, ‐2, ‐3, and ‐11), and zona occludens‐1, which regulate selective permeability and maintain the integrity of the blood–CSF interface [38]. Among these, claudin‐2 in tight junctions plays a crucial role in modulating permeability at the BCSFB, which is inherently more permeable than the BBB [39]. During periods of inflammation, claudin‐2 expression is upregulated [40], increasing paracellular permeability and allowing the passage of cytokines, immune cells, and proteins into the CSF. Although GBCAs do not directly reveal the permeability of the epithelial tight junctions at the BCSFB, they do reflect the degree of stromal vessel permeability and the amount of space available for contrast to accumulate in the stroma. One could hypothesize that increased endothelial permeability within the stroma permits the extravasation of circulating molecules, which may exert toxic or modulatory effects on adjacent epithelial cells, potentially disrupting tight junction integrity and further compromising barrier function. This process is likely mediated by pro‐inflammatory cytokines, which upregulate claudin‐2 expression [41]. Concurrently, the downregulation or mislocalization of other tight junction proteins, such as claudin‐11, may further disrupt the BCSFB, weakening its ability to regulate CSF composition and maintain CNS homeostasis [42].

In the MS cohort, the mean pre‐Gd qT1 values in NAWM and GM were 1356.17 ± 132.69 ms and 2553.59 ± 365.90 ms, respectively, in concordance with previously published ranges at 7T [43]. Values of pre‐Gd qT1 in WMLs were significantly prolonged compared to those in NAWM, consistent with previous studies [23, 44], indicating microstructural damage within WMLs [7]. Interestingly, we also found that pre‐Gd ChP qT1 and ChP ΔT1 were highly correlated (r = 0.83, p < 0.001). Since pre‐Gd qT1 is strongly related to fluid content, this may indicate microstructural changes within the ChP, likely related to edema, inflammation, or expansion of the stromal space, that are strongly related to increased retention of GBCA within the ChP (larger ΔT1).

ChP ΔT1 showed a moderate negative association with head‐size corrected cortical GM volume, suggesting that increased ChP permeability may contribute to cortical loss. As ChP‐driven CSF inflammation reaches cortical and perivascular pathways, these regions may be especially vulnerable.

Several limitations warrant consideration. First, the inversion times of our MP2RAGE acquisition were optimized for maximizing GM–WM contrast and the calculation of qT1 in brain parenchyma, and were not ideal for assessing qT1 or qT1 change in CSF, which has a very long T1. Second, the absence of contrast‐enhanced MRI of healthy participants due to ethical concerns limited direct comparison of ChP ΔT1 between MS participants and HCs. Third, sex was not included as a covariate due to the small number of male participants and the lack of sex effects in exploratory analyses, although this imbalance remains a limitation. The sample size was modest and weighted toward RRMS, restricting generalizability, though the consistency of associations increases confidence in the findings. Finally, the cross‐sectional design precludes evaluating temporal changes in ChP enhancement or links to disease progression, underscoring the need for larger longitudinal studies with more balanced representation of progressive MS.

Funding

This study was supported in part by the Canadian Institutes of Health Research, grant #153005 (to S.N.); the United States Department of Defense, Multiple Sclerosis Research Program, Investigator‐Initiated Research Award (award no. W81XWH19‐1‐0486) (to D.A.R.); and an Investigator‐Initiated award from Novartis Canada (to D.L.A.). S.S. thanks the Fonds de Recherche Québec–Santé for personal support.

Conflicts of Interest

D.L.A. received partial funding for data acquisition from Novartis Canada. The remaining authors declare no conflicts of interest.

Acknowledgments

The authors are grateful to all the participants who made this study possible. We also thank Rozie Arnaoutelis for help in project coordination and Risavarshini Thevakumaran for helpful discussions. This work was previously presented at the European Committee for Treatment and Research in MS (2025).

References

  • 1. Cramer S. P., Simonsen H., Frederiksen J. L., Rostrup E., and Larsson H. B. W., “Abnormal Blood–Brain Barrier Permeability in Normal Appearing White Matter in Multiple Sclerosis Investigated by MRI,” NeuroImage: Clinical 4 (2013): 182–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Ortiz G. G., Pacheco‐Moisés F. P., Macías‐Islas M. Á., et al., “Role of the Blood–Brain Barrier in Multiple Sclerosis,” Archives of Medical Research 45, no. 8 (2014): 687–697. [DOI] [PubMed] [Google Scholar]
  • 3. Zierfuss B., Larochelle C., and Prat A., “Blood–Brain Barrier Dysfunction in Multiple Sclerosis: Causes, Consequences, and Potential Effects of Therapies,” Lancet Neurology 23, no. 1 (2024): 95–109. [DOI] [PubMed] [Google Scholar]
  • 4. Donatelli G., Cecchi P., Migaleddu G., et al., “Quantitative T1 Mapping Detects Blood–Brain Barrier Breakdown in Apparently Non‐Enhancing Multiple Sclerosis Lesions,” NeuroImage: Clinical 40 (2023): 103509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Vercellino M., Votta B., Condello C., et al., “Involvement of the Choroid Plexus in Multiple Sclerosis Autoimmune Inflammation: A Neuropathological Study,” Journal of Neuroimmunology 199, no. 1–2 (2008): 133–141. [DOI] [PubMed] [Google Scholar]
  • 6. Bergsland N., Dwyer M. G., Jakimovski D., et al., “Association of Choroid Plexus Inflammation on MRI With Clinical Disability Progression Over 5 Years in Patients With Multiple Sclerosis,” Neurology 100, no. 9 (2023): e911–e920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Ricigliano V. A. G., Morena E., and Colombi A., “Choroid Plexus Enlargement in Inflammatory Multiple Sclerosis: 3.0‐T MRI and Translocator Protein PET Evaluation,” Radiology 301, no. 1 (2021): 166–177. [DOI] [PubMed] [Google Scholar]
  • 8. Solár P., Zamani A., Kubíčková L., Dubový P., and Joukal M., “Choroid Plexus and the Blood–Cerebrospinal Fluid Barrier in Disease,” Fluids and Barriers of the CNS 17, no. 1 (2020): 35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Reboldi A., Coisne C., Baumjohann D., et al., “C‐C Chemokine Receptor 6‐Regulated Entry of TH‐17 Cells Into the CNS Through the Choroid Plexus Is Required for the Initiation of EAE,” Nature Immunology 10, no. 5 (2009): 514–523. [DOI] [PubMed] [Google Scholar]
  • 10. Dando S. J., Mackay‐Sim A., Norton R., et al., “Pathogens Penetrating the Central Nervous System: Infection Pathways and the Cellular and Molecular Mechanisms of Invasion,” Clinical Microbiology Reviews 27, no. 4 (2014): 691–726. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Kaur C., Rathnasamy G., and Ling E. A., “The Choroid Plexus in Healthy and Diseased Brain,” Journal of Neuropathology and Experimental Neurology 75, no. 3 (2016): 198–213. [DOI] [PubMed] [Google Scholar]
  • 12. Harkins K. D., Xu J., Dula A. N., et al., “The Microstructural Correlates of T1 in White Matter,” Magnetic Resonance in Medicine 75, no. 3 (2016): 1341–1345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Seewann A., Vrenken H., van der Valk P., et al., “Diffusely Abnormal White Matter in Chronic Multiple Sclerosis: Imaging and Histopathologic Analysis,” Archives of Neurology 66, no. 5 (2009): 601–609. [DOI] [PubMed] [Google Scholar]
  • 14. Shen T., Sheriff S., Qu Y., et al., “Correlations Between Postmortem Quantitative MRI Parameters and Demyelination, Axonal Loss and Gliosis in Multiple Sclerosis: A Systematic Review and Meta‐Analysis,” Brain Imaging Behavior 19, no. 2 (2025): 323–335, 10.1007/s11682-025-00971-5. [DOI] [PubMed] [Google Scholar]
  • 15. Tur C., Battiston M., Yiannakas M. C., et al., “What Contributes to Disability in Progressive MS? A Brain and Cervical Cord‐Matched Quantitative MRI Study,” Multiple Sclerosis 30, no. 4–5 (2024): 516–534. [DOI] [PubMed] [Google Scholar]
  • 16. Srinivasan R., Henry R., Pelletier D., and Nelson S., “Standardized, Reproducible, High Resolution Global Measurements of T1 Relaxation Metrics in Cases of Multiple Sclerosis,” AJNR American Journal of Neuroradiology 24, no. 1 (2003): 58–67. [PMC free article] [PubMed] [Google Scholar]
  • 17. Vrenken H., Geurts J. J. G., Knol D. L., et al., “Whole‐Brain T1 Mapping in Multiple Sclerosis: Global Changes of Normal‐Appearing Gray and White Matter,” Radiology 240, no. 3 (2006): 811–820. [DOI] [PubMed] [Google Scholar]
  • 18. Neema M., Stankiewicz J., Arora A., et al., “T1‐ and T2‐Based MRI Measures of Diffuse Gray Matter and White Matter Damage in Patients With Multiple Sclerosis,” Journal of Neuroimaging 17, no. s1 (2007): 16S–21S. [DOI] [PubMed] [Google Scholar]
  • 19. Vaithianathar L., Tench C. R., Morgan P. S., Lin X., and Blumhardt L. D., “White Matter T(1) Relaxation Time Histograms and Cerebral Atrophy in Multiple Sclerosis,” Journal of the Neurological Sciences 197, no. 1–2 (2002): 45–50. [DOI] [PubMed] [Google Scholar]
  • 20. Vrenken H., Rombouts S., Pouwels P. J. W., and Barkhof F., “Voxel‐Based Analysis of Quantitative T1 Maps Demonstrates That Multiple Sclerosis Acts Throughout the Normal‐Appearing White Matter,” AJNR American Journal of Neuroradiology 27, no. 4 (2006): 868–874. [PMC free article] [PubMed] [Google Scholar]
  • 21. Castriota‐Scanderbeg A., Fasano F., Filippi M., and Caltagirone C., “T1 Relaxation Maps Allow Differentiation Between Pathologic Tissue Subsets in Relapsing‐Remitting and Secondary Progressive Multiple Sclerosis,” Multiple Sclerosis 10, no. 5 (2004): 556–561. [DOI] [PubMed] [Google Scholar]
  • 22. Parry A., Clare S., Jenkinson M., Smith S., Palace J., and Matthews P. M., “White Matter and Lesion T1 Relaxation Times Increase in Parallel and Correlate With Disability in Multiple Sclerosis,” Journal of Neurology 249, no. 9 (2002): 1279–1286. [DOI] [PubMed] [Google Scholar]
  • 23. Davies G. R., Hadjiprocopis A., Altmann D. R., et al., “Normal‐Appearing Grey and White Matter T1 Abnormality in Early Relapsing‐Remitting Multiple Sclerosis: A Longitudinal Study,” Multiple Sclerosis 13, no. 2 (2007): 169–177. [DOI] [PubMed] [Google Scholar]
  • 24. Niepel G., Tench C. R., Morgan P. S., Evangelou N., Auer D. P., and Constantinescu C. S., “Deep Gray Matter and Fatigue in MS: A T1 Relaxation Time Study,” Journal of Neurology 253, no. 7 (2006): 896–902. [DOI] [PubMed] [Google Scholar]
  • 25. Marques J. P., Kober T., Krueger G., van der Zwaag W., Van de Moortele P. F., and Gruetter R., “MP2RAGE, a Self Bias‐Field Corrected Sequence for Improved Segmentation and T1‐Mapping at High Field,” Neuroimage 49, no. 2 (2010): 1271–1281. [DOI] [PubMed] [Google Scholar]
  • 26. Choi S., Spini M., Hua J., and Harrison D. M., “Blood–Brain Barrier Breakdown in Non‐Enhancing Multiple Sclerosis Lesions Detected by 7‐Tesla MP2RAGE ΔT1 Mapping,” PLoS ONE 16, no. 4 (2021): e0249973. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Thompson A. J., Banwell B. L., Barkhof F., et al., “Diagnosis of Multiple Sclerosis: 2017 Revisions of the McDonald Criteria,” Lancet Neurology 17, no. 2 (2018): 162–173. [DOI] [PubMed] [Google Scholar]
  • 28. Vincent R. D., Neelin P., Khalili‐Mahani N., et al., “MINC 2.0: A Flexible Format for Multi‐Modal Images,” Frontiers in Neuroinformatics 10 (2016): 35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Tustison N. J., Avants B. B., Cook P. A., et al., “N4ITK: Improved N3 Bias Correction,” IEEE Transactions on Medical Imaging 29, no. 6 (2010): 1310–1320. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Isensee F., Jaeger P. F., Kohl S. A. A., Petersen J., and Maier‐Hein K. H., “nnU‐Net: A Self‐Configuring Method for Deep Learning‐Based Biomedical Image Segmentation,” Nature Methods 18, no. 2 (2021): 203–211. [DOI] [PubMed] [Google Scholar]
  • 31. Senthil S., Thevakumaran R., Blinder S., et al. “Combined [11C]PBR28 PET and Quantitative T1 Mapping Reveal Choroid Plexus Alterations in Multiple Sclerosis,” preprint, medRxiv, November 13, 2025, 10.1101/2025.11.10.25339910. [DOI]
  • 32. Krieger B., Bellenberg B., Roenneke A. K., et al., “Relevance of Choroid Plexus Volumes in Multiple Sclerosis,” Fluids and Barriers of the CNS 22, no. 1 (2025): 47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Vos C. M. P., Geurts J. J. G., Montagne L., et al., “Blood–Brain Barrier Alterations in Both Focal and Diffuse Abnormalities on Postmortem MRI in Multiple Sclerosis,” Neurobiology of Disease 20, no. 3 (2005): 953–960. [DOI] [PubMed] [Google Scholar]
  • 34. Kuhlmann T., Ludwin S., Prat A., Antel J., Brück W., and Lassmann H., “An Updated Histological Classification System for Multiple Sclerosis Lesions,” Acta Neuropathologica 133, no. 1 (2017): 13–24. [DOI] [PubMed] [Google Scholar]
  • 35. Daneman R. and Prat A., “The Blood–Brain Barrier,” Cold Spring Harbor Perspectives in Biology 7, no. 1 (2015): a020412. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Tagge I. J., Leppert I. R., Fetco D., et al., “Permanent Tissue Damage in Multiple Sclerosis Lesions Is Associated With Reduced Pre‐Lesion Myelin and Axon Volume Fractions,” Multiple Sclerosis 28, no. 13 (2022): 2027–2037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Abbott N. J., “Evidence for Bulk Flow of Brain Interstitial Fluid: Significance for Physiology and Pathology,” Neurochemistry International 45, no. 4 (2004): 545–552. [DOI] [PubMed] [Google Scholar]
  • 38. Saul J., Hutchins E., Reiman R., et al., “Global Alterations to the Choroid Plexus Blood–CSF Barrier in Amyotrophic Lateral Sclerosis,” Acta Neuropathologica Communications 8, no. 1 (2020): 92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Ueno M., “Elucidation of Mechanism of Blood–Brain Barrier Damage for Prevention and Treatment of Vascular Dementia,” [in Japanese], Rinsho Shinkeigaku Clinical Neurology 57, no. 3 (2017): 95–109. [DOI] [PubMed] [Google Scholar]
  • 40. Gonçalves A., Ambrósio A. F., and Fernandes R., “Regulation of Claudins in Blood‐Tissue Barriers Under Physiological and Pathological States,” Tissue Barriers 1, no. 3 (2013): e24782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Suzuki T., Yoshinaga N., and Tanabe S., “Interleukin‐6 (IL‐6) Regulates Claudin‐2 Expression and Tight Junction Permeability in Intestinal Epithelium,” Journal of Biological Chemistry 286, no. 36 (2011): 31263–31271. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Wolburg H., Wolburg‐Buchholz K., Liebner S., and Engelhardt B., “Claudin‐1, Claudin‐2 and Claudin‐11 Are Present in Tight Junctions of Choroid Plexus Epithelium of the Mouse,” Neuroscience Letters 307, no. 2 (2001): 77–80. [DOI] [PubMed] [Google Scholar]
  • 43. Kolb H., Absinta M., Beck E. S., et al., “7T MRI Differentiates Remyelinated from Demyelinated Multiple Sclerosis Lesions,” Annals of Neurology 90, no. 4 (2021): 612–626, 10.1002/ana.26194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Kober T., Granziera C., Ribes D., et al., “MP2RAGE Multiple Sclerosis Magnetic Resonance Imaging at 3 T,” Investigative Radiology 47, no. 6 (2012): 346–352, 10.1097/rli.0b013e31824600e9. [DOI] [PubMed] [Google Scholar]

Articles from Journal of Neuroimaging are provided here courtesy of Wiley

RESOURCES