Published online Apr 12, 2023.
https://doi.org/10.3348/kjr.2022.0980
Assessment of Meningeal Lymphatics in the Parasagittal Dural Space: A Prospective Feasibility Study Using Dynamic Contrast-Enhanced Magnetic Resonance Imaging
Abstract
Objective
Meningeal lymphatic vessels are predominantly located in the parasagittal dural space (PSD); these vessels drain interstitial fluids out of the brain and contribute to the glymphatic system. We aimed to investigate the ability of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in assessing the dynamic changes in the meningeal lymphatic vessels in PSD.
Materials and Methods
Eighteen participants (26–71 years; male:female, 10:8), without neurological or psychiatric diseases, were prospectively enrolled and underwent DCE-MRI. Three regions of interests (ROIs) were placed on the PSD, superior sagittal sinus (SSS), and cortical vein. Early and delayed enhancement patterns and six kinetic curve-derived parameters were obtained and compared between the three ROIs. Moreover, the participants were grouped into the young (< 65 years; n = 9) or older (≥ 65 years; n = 9) groups. Enhancement patterns and kinetic curve-derived parameters in the PSD were compared between the two groups.
Results
The PSD showed different enhancement patterns than the SSS and cortical veins (P < 0.001 and P < 0.001, respectively) in the early and delayed phases. The PSD showed slow early enhancement and a delayed wash-out pattern. The six kinetic curve-derived parameters of PSD was significantly different than that of the SSS and cortical vein. The PSD wash-out rate of older participants was significantly lower (median, 0.09; interquartile range [IQR], 0.01–0.15) than that of younger participants (median, 0.32; IQR, 0.07–0.45) (P = 0.040).
Conclusion
This study shows that the dynamic changes of meningeal lymphatic vessels in PSD can be assessed with DCE-MRI, and the results are different from those of the venous structures. Our finding that delayed wash-out was more pronounced in the PSD of older participants suggests that aging may disturb the meningeal lymphatic drainage.
INTRODUCTION
For several decades, the central nervous system (CNS) has been considered an immune-privileged organ as it has not been considered part of the lymphatic system [1, 2]. However, the existence of meningeal lymphatic vessels was first described in animal studies by Aspelund et al. [3] and Louveau et al. in 2015 [4], and its presence was further evidenced by human studies [5, 6, 7, 8]. The meningeal lymphatic vessels are found alongside the major venous sinus, cranial nerves, and arteries, and the meningeal lymphatic vessels drain cerebrospinal fluid (CSF) and interstitial fluid (ISF) from the CNS into the extracranial cervical lymph nodes [3, 4, 9]. In addition to its involvement in clearing brain waste, regulating CSF, and maintaining brain homeostasis, meningeal lymphatics may be an important drainage route for CSF/ISF of the brain, and it is considered to be closely related to the glymphatic system, which is a glia-dependent system of perivascular channels [3, 4, 10].
Several human studies have demonstrated the presence of meningeal lymphatics in the parasagittal dural space (PSD) around the superior sagittal sinus (SSS) using in vivo magnetic resonance imaging (MRI) with intravenous or intrathecal gadolinium contrast enhancement [6, 7]. One study showed that meningeal lymphatic drainage in the PSD was impaired in the aging human brain; the study used multiple time point fluid-attenuated inversion recovery imaging (FLAIR) acquisitions following intrathecal administration of a gadolinium contrast agent [11].
Intrathecal injection of a gadolinium contrast agent may enable direct visualization of meningeal lymphatic function [11, 12, 13]. Off-label intrathecal injection of a gadolinium contrast agent has been reported to be safe when using an appropriate low-dose injection; however, non-serious adverse events, such as headache, nausea, and dizziness, have been reported [14]. Furthermore, intrathecal administration of gadolinium contrast agents on an outpatient basis is not easily applicable, as it requires multiple time point acquisitions of up to 48 hours [14], which may limit wider clinical application and research potential. Therefore, less invasive and rapid methods of dynamic assessment and quantification of meningeal lymphatic drainage in humans are needed for widespread clinical application.
However, few studies have investigated the dynamic aspect of meningeal lymphatic vessels using intravenous administration of a gadolinium contrast agent [15, 16]. One previous study applied dynamic contrast-enhanced MRI (DCE-MRI) to compare meningeal lymphatic flow in patients with idiopathic Parkinson’s disease and suggested DCE-MRI as a promising imaging biomarker for quantifying meningeal lymphatic flow [16]. However, further research is required to determine the usefulness of DCE-MRI as an assessment tool for the meningeal lymphatic vessels.
We hypothesized that a distinct dynamic enhancement pattern would be depicted in the PSD, which is thought to harbor meningeal lymphatic vessels and is distinguishable from that of adjacent venous structures. We also hypothesized that dynamic changes in PSD would occur during the aging process. Therefore, this study aimed to investigate the effectiveness of DCE-MRI with an intravenous injection of a gadolinium contrast agent in depicting dynamic patterns of PSD in neurologically normal participants from other venous structures and to evaluate the difference in dynamic contrast enhancement patterns in the PSD between older adults and their younger counterparts.
MATERIALS AND METHODS
This prospective study was approved by the Gangnam Severance Hospital Institutional Review Board (IRB No. 3-2020-0480) and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all the participants.
Study Participants
In this prospective feasibility study, 21 neurologically healthy participants were screened for recruitment between May 2021 and October 2021. Inclusion criteria for participants were as follows: 1) aged ≥ 19 years, 2) no history of neurological or psychiatric disease, 3) intact activities of daily living, 4) Mini-Mental State Examination (MMSE) scores of ≥ 26; and 5) an estimated glomerular filtration rate of ≥ 60 mL/min/1.73 m2 tested within a year. The exclusion criteria were as follows: 1) significantly abnormal findings or severe artifacts on brain MRI and 2) a physician-indicated contraindication to an MRI examination. One participant was excluded because of a history of allergy to contrast media, and two additional participants were later excluded because of severe motion artifacts in the MRI. Finally, 18 participants were enrolled in this study and were grouped into young participants (< 65 years; n = 9) and older adult participants (≥ 65 years; n = 9). A flowchart of participant inclusion in the study is shown in Figure 1.
Fig. 1
Flowchart of participant inclusion.
Clinical Evaluation
All participant’s medical conditions (including a history of hypertension, diabetes, and hyperlipidemia), MMSE, and Clinical Dementia Rating (CDR) scores were assessed.
MRI Protocol
All examinations were conducted using a 3 T scanner (MAGNETOM VIDA, Siemens Healthcare). The image sequences included sagittal three-dimensional (3D) magnetization-prepared rapid acquisition with gradient echo imaging, axial T2-weighted imaging, axial susceptibility-weighted imaging, sagittal 3D T2-weighted FLAIR, coronal DCE-MRI, and sagittal contrast-enhanced 3D T1-weighted black blood imaging (3D T1 BB). The parameters for each sequence are listed in Supplementary Table 1. DCE-MRI was acquired using the golden-angle radial sparse parallel-volumetric interpolated breath-hold (GRASP-VIBE) sequence and consisted of 92 series with time intervals of 6.2 s. Each series was composed of 30 sections of coronal images with 3.0 mm thickness and acquired in an anterior-posterior direction with the level of the corpus callosum genu being the foremost plane. The following imaging parameters were used for DCE-MRI: repetition time, 6.24 ms; echo time, 2.91 ms; field of view, 240 x 240 mm; voxel size, 0.625 x 0.625 x 3.0 mm3. At the 10th dynamic phase, a gadolinium-based contrast agent was injected (0.2 mL/kg ProHance, Bracco) at a rate of 4 mL/s, followed by 30 mL of saline bolus (Supplementary Fig. 1). The total acquisition time for DCE-MRI was 9 min 32 sec. After DCE-MRI acquisition, contrast-enhanced 3D T1 BB was performed without additional contrast injection.
Imaging Analysis
To investigate the distinct enhancement pattern of the meningeal lymphatic vessels, three circular regions of interests (ROIs) were manually placed by a neuroradiologist with 9 years of experience in neuroradiology, with reference to the contrast-enhanced 3D T1 BB and 3D T2 FLAIR images in the following areas: PSD, SSS, and cortical vein (CV; Fig. 2). PSD was defined as the area along the middle to posterior segment of the SSS showing intermediate or high signals on 3D T1 BB (not vascular structures) and signals not suppressed on 3D T2 FLAIR images (not CSF space) [12]. The ROI on the PSD was placed at the superolateral aspect of the SSS, which is known to harbor meningeal lymphatic vessels [6, 7, 8, 12]. The ROI size ranges for PSD, SSS, and CV were 0.39–5.08 mm2, 0.78–15.62 mm2, and 0.39–5.86 mm2, respectively.
Fig. 2
An example of regions of interest (ROI) placement. An example of ROI placement around the posterior superior sagittal sinus (SSS) is shown in the images of a 71-year-old male participant. Referencing the contrast-enhanced 3D T1 BB (A) and 3D T2 FLAIR (B) images, three circular ROIs were manually placed on the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) images (C) by an experienced neuroradiologist. The red, yellow, and blue circles represent ROIs placed on the parasagittal dural space (PSD), SSS, and cortical vein, respectively. D: The PSD is depicted as the area at the superolateral aspect of the SSS outlined by red lines on a late-phase DCE-MRI image. After these three ROIs were transferred to the processed DCE-MRI maps (an example of the peak enhancement ratio map is depicted) (E), the final time-intensity curve for each ROI was generated (F). The timing of the contrast injection is indicated by the gray dashed line. 3D T1 BB = three-dimensional T1-weighted black blood imaging, 3D T2 FLAIR = three-dimensional T2 fluid-attenuated inversion recovery imaging
For each ROI, DCE-MRI results were automatically analyzed qualitatively and semi-quantitatively using time-intensity curve analysis in the Olea Sphere 2.2 software (Olea Medical). For qualitative analysis, the enhancement phase was divided into early and delayed phases based on 2 min post-contrast agent injection [17]. The enhancement pattern of the early phase was then classified based on the percentage increase in signal intensity from the precontrast levels, with increases of < 50%, 50%–100%, and > 100% considered as slow, medium, and fast, respectively. The enhancement pattern of the delayed phase was classified as persistent (defined as a continuous increase in the enhancement of > 10% early enhancement), plateau (constant signal intensity once the peak reaches ±10% early enhancement), and wash-out (decreasing signal intensity after peak enhancement of > 10% early enhancement) [18].
For semi-quantitative analysis, the following six kinetic curve-derived parameters were obtained: wash-in rate; wash-out rate; peak enhancement ratio, defined as (maximum signal intensity – precontrast signal intensity [PER])/PER; time to peak enhancement (TTP); relative wash-out, defined as (maximum signal intensity – end signal intensity)/(maximum signal intensity – PER), and area under the curve (AUC) [19, 20].
To measure the maximal thickness of the PSD, sagittal contrast-enhanced 3D T1 BB images were reformatted into coronal and axial 1 mm slices. We noted the presence of hyperintense signals around the SSS; based on methods detailed in a previous study [21], an experienced neuroradiologist measured its maximum thickness on the coronal images of the anterior, middle, and posterior sagittal sinus, perpendicular to the posterolateral wall of the PSD to the SSS.
For the analysis of inter-rater reliability, another neuroradiologist with 4 years of experience in neuroradiology independently placed ROIs in the PSD, and kinetic curve-derived parameters were obtained for all participants.
Statistical Analysis
Categorical data are shown as the number of observations (percentage), and continuous data are shown as the median (interquartile range [IQR]), according to normality tests using the Kolmogorov-Smirnov test. Demographic and clinical characteristics were compared between the young and older adult participants using the Mann–Whitney U test or Fisher’s exact test.
Intraclass correlation coefficients (ICC) were obtained using a two-way random model and an absolute agreement type for each parameter to assess inter-rater reliability. ICC estimate values of < 0.5, 0.5–0.75, 0.75–0.9, and > 0.90 were indicative of poor, moderate, good, and excellent reliability, respectively [22].
Enhancement patterns in the early and delayed phases were dichotomized into non-fast versus fast patterns and non-wash-out versus wash-out patterns, respectively. Comparisons of the dichotomized enhancement patterns of the early and delayed phases in the three ROIs (PSD, SSS, and CV) were analyzed using the chi-square test. The kinetic curve-derived parameters of each ROI were compared using the Kruskal-Wallis test with post-hoc Bonferroni corrections.
We investigated the differences in the enhancement patterns or the kinetic curve-derived parameters of the three ROIs (PSD, SSS, and CV) between the young and older adult participants using Fisher’s exact test for pattern comparison and Mann–Whitney U test for kinetic curve-derived parameter comparison. To compare the maximal thickness of the anterior, middle, and posterior PSD between young and older adult participants, the Mann–Whitney U test was performed for each area. Statistical analyses were performed using SPSS (version 25.0; IBM Corporation) and MedCalc (version 9.3.6.0; MedCalc Software). Statistical significance was set at a two-tailed level of 0.05.
RESULTS
Participant Characteristics
The participants’ demographic and clinical characteristics are summarized in Table 1. The younger participants were significantly younger than the older adult participants (median 28.0 [IQR 26.8–35.3] years vs. 66.5 [IQR 65.8–69.5] years; P < 0.001). The MMSE score was significantly higher in young participants than in older adult participants (median 30.0 [IQR 29.8–30.0] vs. 28.5 [IQR 28.0–30.0]; P = 0.040]. There were no significant differences between the two groups in several factors including: sex; CDR score; and the presence of hypertension, diabetes, and hyperlipidemia.
Table 1
Participant Characteristics
Inter-Rater Reliability of Kinetic Curve-Derived Parameters of PSD
ICC analysis showed moderate or good inter-rater reliability for the following kinetic curve-derived parameters of the PSD: wash-in rate, 0.700 (95% confidence interval [CI], 0.209–0.887); wash-out rate, 0.769 (95% CI, 0.389–0.913); PER, 0.683 (95% CI, 0.150–0.881); TTP, 0.697 (95% CI, 0.197–0.887); relative wash-out, 0.755 (95% CI, 0.364–0.907); and AUC, 0.605 (95% CI, 0.023–0.848).
Comparisons of Dynamic Enhancement among PSD, SSS, and CV
In the early phase, two (11.1%) of the 18 participants showed fast enhancement patterns in the PSD. Among the 18 participants, a fast enhancement pattern was observed in the SSS of all (100%) participants and CV of 16 (88.9%) participants (P < 0.001). In the delayed phase, a wash-out pattern was only observed in four (22.2%) PSDs of 18 participants; whereas a wash-out pattern was observed in all (100%) SSSs and 17 (94.4%) CVs (P < 0.001) (Table 2).
Table 2
Enhancement Pattern Comparisons among PSD, SSS, and CV
All six kinetic curve-derived parameters, including wash-in rate, wash-out rate, PER, TTP, relative wash-out, and AUC, differed significantly among the three ROIs (all P < 0.001). In the post-hoc analysis with Bonferroni corrections, PSD showed distinct values than the other two structures (Table 3).
Table 3
Kinetic Curve-Derived Parameter Comparison among PSD, SSS, and CV
Age-Associated Differences in the Dynamic Enhancement of PSD
In the early phase, no participants in the older adult group demonstrated a fast enhancement pattern (n = 0, 0%), whereas two participants in the young participant group demonstrated a fast enhancement pattern (n = 2, 22.2%); however, this difference was not significant (P = 0.470). In the delayed phase, there was only one participant who showed a wash-out pattern in the older adult group (n = 1, 11.1%), whereas three participants showed wash-out patterns in the young group (n = 3, 33.3%); however, this difference was not significant (P = 0.576) (Table 4).
Table 4
Comparison of Enhancement Patterns in the Parasagittal Dural Space between Young and Older Adult Participants
Concerning the kinetic curve-derived parameters, the median wash-out rate in the young participant group (0.32, IQR 0.07–0.45) was significantly higher than that in the older adult participant group (0.09, IQR 0.01–0.15; P = 0.04). However, no significant differences were found in the other five kinetic curve-derived parameters between the two groups (Table 5, Fig. 3). In addition, no significant differences were found in any of the kinetic curve-derived parameters for SSS and CV between the two groups (Supplementary Table 2). Representative time-intensity curves for young and older adult participants are shown in Figure 4.
Fig. 3
Comparisons of kinetic curve-derived parameters in parasagittal dural space between young and older adult participants. (A-F) The central red dots and whiskers represent the medians and interquartile ranges, respectively. *Statistically significant difference. AUC = area under the curve, PER = peak enhancement ratio, TTP = time-to-peak enhancement
Fig. 4
Representative time-intensity curves of young (A) and older adult (B) participants. The dashed lines representing 2 min after contrast injection, divide the early and delayed phases. CV = cortical vein, PSD = parasagittal dural space, SSS = superior sagittal sinus, A.U. = arbitrary units
Table 5
A Kinetic Curve-Derived Parameter Comparison of the Parasagittal Dural Space between Young and Older Adult Participants
Age-Associated Differences in the PSD Maximal Thickness
The older adult participants had a higher median thickness of anterior (4.4 mm [IQR 3.2–5.1] vs. 3.2 mm [IQR 2.2–3.6]; P = 0.008), middle (5.9 mm [IQR 5.0–7.8] vs. 4.4 mm [IQR 3.5–5.0]; P = 0.031), and posterior (6.7 mm [IQR 6.6–7.0] vs. 2.8 mm [IQR 2.7–3.3]; P < 0.001) PSD.
DISCUSSION
This study investigated in vivo dynamic assessment of meningeal lymphatic vessels in neurologically normal young and older participants using DCE-MRI. Meningeal lymphatic enhancement in the PSD around the SSS showed a distinct dynamic enhancement pattern from adjacent venous structures, confirming that PSD is not a constituent of the venous system but of the meningeal lymphatic system itself, as previously suggested [6, 12]. Furthermore, on semi-quantitative analysis, delayed wash-out of PSD was found in older adult participants than in young participants, which may represent meningeal lymphatic stasis and, consequently, meningeal lymphatic drainage dysfunction.
Recently, both animal and human studies reported the presence of a lymphatic system within the meninges [3, 4, 5, 6, 7, 8, 11, 12, 13, 15, 16, 21, 23, 24, 25]. However, most prior in vivo studies have either assessed only the morphological features of meningeal lymphatic vessels [21, 26] or required the invasive off-label use of a gadolinium agent with a very long scan time of > 24 h, which may limit its wider clinical application [11, 12, 13, 14]. Only a few studies have utilized intravenous injection of a contrast agent to assess the function of the meningeal lymphatic system. A recent study reported that meningeal lymphatic vessels around the SSS showed distinct enhancement on T1 BB or T2 FLAIR images after intravenous injection of gadolinium contrast in healthy participants [6]. Another study applied DCE-MRI with an intravenous injection of gadolinium contrast in patients with various Parkinsonian disorders and reported that the disease diagnosis could affect meningeal lymphatic drainage differently [16]. The main strength of our study is that it provides an easily applicable method that can be performed within an acceptable scan time in daily practice. Moreover, since meningeal lymphatic dysfunction has been proposed as an aggravating factor in Alzheimer’s disease pathology and age-associated cognitive decline [23], this in vivo monitoring tool using DCE-MRI may serve as a valuable imaging biomarker for assessing the functional capability of meningeal lymphatics.
Initially, the enhancing structure visible adjacent to the major intracranial venous sinuses on MRI had, been considered as a meningeal lymphatic vessel itself [6]; however, rather than being the meningeal lymphatic vessel itself, this has now been considered to be the parasinus dural space which consists of prominent intradural channels that are supposed to work as a barrier for CSF drainage into the venous or lymphatic systems [12, 27, 28, 29, 30]. Considering the absence of blood-meningeal barriers in dural blood vessels, the enhancement detected in PSD after intravenous injection of gadolinium agent may reflect that the gadolinium contrast agent in the dura mater migrates to collecting lymphatics in PSD, through which it finally drains into the extracranial cervical lymph nodes [8, 15, 16]. Given that PSD harbors meningeal lymphatic vessels [23, 31], it appears reasonable to suggest that our DCE-MRI technique, which focuses on PSD, is a practical method for the assessment of meningeal lymphatic drainage.
Previous animal studies have highlighted decreased meningeal lymphatic function in addition to different genetic expressions in the endothelial cells of meningeal lymphatic vessels in aged mice, which may explain age-associated cognitive decline or neurological conditions [23, 32]. A recent human MRI study also reported meningeal lymphatic dysfunction in older adult participants using intrathecal gadolinium injection, which exhibited delayed clearance of meningeal lymphatic vessels in this population [11]. Additionally, researchers have described morphological changes in the PSD in relation to aging and reported an age-related increase in the volume/thickness of the PSD [21, 26, 33]. These findings of age-related functional and morphologic changes in the PSD are consistent with our observation that older adult participants had increased PSD thickness and delayed PSD contrast wash-out compared to that of younger participants, which may represent meningeal lymphatic system stasis. Furthermore, these findings seem to be consistent with general age-related dilatation and contractile dysfunction of lymphatic vessels [34, 35].
This feasibility study had some limitations. First, only a small number of participants were enrolled. Further studies with a larger number of subjects are needed to implement DCE-MRI as a tool to evaluate the meningeal lymphatic system in clinical practice. Second, all participants were neurologically healthy individuals. Although the aging process is known to cause meningeal lymphatic dysfunction, participants with various diseases that may induce meningeal lymphatic dysfunction, such as Alzheimer’s disease or traumatic brain injury, may show clearer differences in the dynamic changes in the PSD. Third, some participants had vascular risk factors such as hypertension, diabetes, and hyperlipidemia, which have been proven by both animal and human studies to be associated with glymphatic dysfunction [36, 37, 38, 39, 40]. However, to date, these vascular risk factors have not been proven to have an effect on meningeal lymphatic function. Nevertheless, these common vascular risk factors may potentially affect meningeal lymphatic function and the glymphatic system. Further investigations are required to address this issue. Fourth, although there was moderate to good inter-rater reliability for dynamic parameters showing ICC values > 0.6 in all parameters, the manual placement of ROIs on a small structure may lead to possible reproducibility issues. To address this issue, we plan to automate the positioning of the ROI on the PSD using an artificial intelligence-based approach. Finally, test-retest reliability was not evaluated in this study. Future studies should address this issue to validate the applicability of DCE-MRI in the evaluation of meningeal lymphatic function.
In conclusion, the findings of this feasibility study demonstrated the in vivo dynamic nature of the meningeal lymphatic system in PSD using DCE-MRI with intravenous injection of a contrast agent, which is different from that of adjacent venous structures. Delayed wash-out was more pronounced in the PSD of older adult participants, suggesting that the aging process may disturb meningeal lymphatic drainage. Future studies that include patients with various neurological diseases are warranted to determine whether this quantitative DCE-MRI technique can facilitate the assessment of diseases and the associated dysfunction of the meningeal lymphatics.
Supplement
The Supplement is available with this article at https://doi.org/10.3348/kjr.2022.0980.
Click here to view.(2M, pdf)
Conflicts of Interest:Sang Hyun Suh, a contributing editor of the Korean Journal of Radiology, was not involved in the editorial evaluation or decision to publish this article. All remaining authors have declared no conflicts of interest.
Author Contributions:
Conceptualization: Mina Park.
Data curation: Bio Joo, Sung Jun Ahn.
Formal analysis: Bio Joo.
Funding acquisition: Mina Park.
Investigation: Mina Park.
Methodology: Mina Park.
Resources: Mina Park.
Software: Sung Jun Ahn.
Supervision: Sang Hyun Suh.
Validation: Mina Park.
Visualization: Bio Joo.
Writing—original draft: Bio Joo.
Writing—review & editing: Mina Park.
Funding Statement:This study was supported by a faculty research grant of Yonsei University College of Medicine (6-2019-0098) and by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF-2020R1C1C1005724).
Availability of Data and Material
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
Acknowledgments
We would like to thank In-seong Kim, PhD of Siemens Korea for his contribution in the protocol setting.
References
-
Sung YS, Park B, Choi Y, Lim HS, Woo DC, Kim KW, et al. Dynamic contrast-enhanced MRI for oncology drug development. J Magn Reson Imagin 2016;44:251–264.
-
MeSH Terms
Figures
1 / 4
Tables
1 / 5
ORCID IDs
Funding Information
-
Yonsei University College of Medicine
6-2019-0098
-
National Research Foundation of Korea
NRF-2020R1C1C1005724
