THE GLYMPHATIC SYSTEM HYPOTHESIS

Based on the results of various animal experiments, the concept of a glymphatic system has been hypothesized.1,2 During these 10 years, various findings for a fluid and solutes clearance pathway in the brain, known as the “glymphatic system,” have accumulated.1,3 There remains considerable debate regarding this glymphatic system hypothesis4; however, numerous studies for the waste clearance function of the central nervous system (CNS) including the glymphatic system have been published.3,5–7

According to the proposed hypothesis, the glymphatic system is a highly organized fluid-transporting system. The vascular endfeet of astrocytes creates perivascular spaces around arteries (ie, periarterial spaces) through which cerebrospinal fluid (CSF) enters the brain and spreads throughout the interstitium. Cerebrospinal fluid enters these perivascular spaces from the subarachnoid space, is carried deep into the brain by arterial pulsatility, and finally permeates the brain parenchyma. The water channel aquaporin (AQP) 4 in the endfeet of astrocytes facilitates this process. Cerebrospinal fluid mixes with interstitial fluid in the extracellular space and exits the brain through the perivascular spaces around veins (ie, perivenous spaces) and along the cranial and spinal nerves. Interstitial solutes, including protein waste, are then transported by the glymphatic system and exported from the CNS via the meningeal and cervical lymphatic vessels (Fig. 1).8,9 Although there is still debate regarding the glymphatic system hypothesis, in this review, we use the term “glymphatic system” to broadly define the interstitial fluid drainage system of the CNS. One reason that the glymphatic system attracts so much attention from the scientific community is that sleep facilitates its function.2,10,11 Another reason is that malfunction of the glymphatic system is thought to be associated with neurodegenerative diseases including Alzheimer disease.8,12

FIGURE 1: Diagram of the interstitial fluid transport and the glymphatic concept. The movement of fluid and solutes in brain tissue between the brain parenchyma and the perivascular space surrounding arteries and veins is regulated by the water channel aquaporin (AQP) 4 in the endfeet of astrocytes. Green arrows indicate fluid movement. The figure shows fluid movement along the periarterial space into the brain parenchyma and out along the perivenous space. Purple indicates interstitial solutes. The solutes exit the brain parenchyma through gaps in the astrocytic endfeet into the perivenous space. This illustration was previously presented by Ray et al9 licensed under CC BY 4.0.

Reports evaluating the glymphatic system in humans have mostly used magnetic resonance imaging (MRI).5,11,13–15 For routine clinical MR examinations, scanners of various field strength are used16,17; however, 3 T MR scanners are mainly used for evaluation of the glymphatic system in humans.7,18–21 Gadolinium deposition in the brain is thought to occur when intravenously administered, linear-type gadolinium-based contrast agent (GBCA) enters the brain parenchyma via an impaired blood-brain barrier (BBB), or alternatively via the glymphatic system after passing through the blood-CSF barrier (B-CSF-B).21–24 However, the most likely explanation is that GBCAs enter the brain parenchyma via the glymphatic system.21,25–29

In this review, we will summarize the findings to date for evaluation of the glymphatic system in humans, by dividing them into studies that did or did not use administration of gadolinium tracers. We then describe the details of each imaging method and the current state of the research findings. The sequence will follow the order of fluid flow through the glymphatic system, including the subarachnoid space, perivascular spaces around arteries, brain parenchyma, and finally the perivenous drainage pathways. We will also focus on the parietal subarachnoid space, parasagittal dura, meningeal lymphatics, and cervical lymph nodes, which are thought to be downstream components of the glymphatic system. The dynamics of the glymphatic system in the eye and the endolymphatic/perilymphatic fluid of the inner ear labyrinth, which has recently attracted much attention in the research community, will also be described.

CEREBROSPINAL FLUID PRODUCTION AND DYNAMICS MRI Without Exogenous Contrast Agents

The classical theory that most of the CSF is produced by the choroid plexus has been challenged. Recent data show that the interstitial fluid and CSF are mainly formed and reabsorbed across the walls of the CNS blood capillaries, suggesting that a directed CSF circulation from the choroid plexus to the arachnoid villi is not necessary.30 For the evaluation of CSF dynamics in humans, including the production of the CSF, several noninvasive MRI techniques have been used. One of these is the time-spatial inversion pulse method. This method applies a slab of inversion pulses to the CSF and visualizes its movement using the CSF itself as the tracer. This method does not require the administration of an exogenous tracer. Therefore, CSF dynamics can be observed under physiological conditions.31 Time-spatial inversion pulse observations show almost no flow around the choroid plexus in the lateral ventricles, except near the foramen of Monro. This finding is inconsistent with the classical theory that CSF is produced primarily in the choroid plexus.

Phase contrast techniques can also be used to study CSF flow. Recently, 4-dimensional phase contrast imaging, which provides a quantitative spatiotemporal velocity distribution, has been applied to CSF dynamics.32 Pulsatile “to and fro” movements are observed in the foramen of Monro and in the Sylvian fissure. On the other hand, there is less CSF movement within the lateral ventricles and in the convexity region. It has also been shown that the movement of CSF is greatly altered by both respiration and cardiac pulsation.

A method using pseudo-continuous arterial spin labeling with varying the labeling duration as well as the postlabeling delay and with various echo times for a long, echo train readout was reported, allowing the separation of a very long T2 CSF signal. A 2-compartment dynamic model was used to describe the exchange of water from the blood to the CSF. Despite the expectation that the CSF arterial spin labeling signal would be concentrated in the choroid plexus region based on the classical CSF circulation theory, the study found that it was more distributed around cortical regions. This report supports the view that the classical theory of CSF secretion exclusively in the choroid plexus is an inadequate explanation of CSF dynamics.33

Low b-value diffusion imaging is a simple method to observe CSF dynamics over a large area in a short time. Diffusion-weighted images with low b-values such as b = 500 s/mm2 can evaluate CSF dynamics as the stationary CSF signal remains to some extent, and the signal decrease reflects the magnitude of the water movement.34,35 Even with this simple method, there is almost no signal reduction in the bilateral lateral ventricles, including the region surrounding the choroid plexus.

Low-frequency CSF oscillations have been measured using ultrafast MRI. In one study, BOLD fMRI dynamics, electroencephalogram, and CSF flow were measured simultaneously in humans during sleep.36 The results showed that the sleeping brain exhibits waves of CSF flow on a macroscopic scale, and that CSF dynamics are related to neural and hemodynamic rhythms.

MRI With the Administration of Exogenous Contrast Agents

For the evaluation of CSF dynamics in humans using MRI, GBCAs and 17O-labeled water are the candidate tracers.5,37,38 The intrathecal administration of GBCA is basically off-label use. The administration of large amounts of GBCA into CSF space can cause severe adverse effects.39–44 Intrathecally administered GBCA (IT-GBCA) permeates the brain parenchyma. Results with administration of a small amount of IT-GBCA have been reported in many studies.27,37,39 Penetration of GBCA to the brain parenchyma is first seen along large arteries, indicating a pivotal role of intracranial pulsations in glymphatic function.37 The distribution pattern of GBCA is different between idiopathic normal pressure hydrocephalus (iNPH) patients and controls. Washout of IT-GBCA was impaired in iNPH.13,45 The time course of GBCA concentration in blood after IT administration varies among various diseases.46

After IT-GBCA in animals, penetration of GBCA into the brain is inhibited by subarachnoid hemorrhage. Subsequent infusion of tissue plasminogen activator into the ventricles restores GBCA penetration into the brain.47 Thus, because the penetration of IT-GBCA into the brain is affected by diseases and therapeutic conditions, IT-GBCA may be established in the future as a new method for testing the function of the glymphatic system if its safety can be ensured in humans as well.39,40,42

The tracer 17O-labeled water is a water molecule containing a stable isotope of oxygen and is an extremely safe agent, which behaves physiologically similar to normal 16O-water.38,48 Proton MRI with a T2-shortening effect by 17O-labeled water is easy to use and does not require special equipment or pulse sequences.38,48,49

Initial human studies with the intravenous (IV) administration of 17O-water showed a rapid decrease in the cortical signal with a rapid decrease in choroid plexus signal after administration, but the CSF signal in the ventricles was delayed compared with CSF signal decrease in the subarachnoid space.38 These results do not support the classical theory that CSF is produced in the choroid plexus. More recently, it was proposed that CSF is produced not only in the choroid plexus but also from capillaries within the brain parenchyma.30 However, 17O-labeled water is very expensive and can cost several hundred US dollars per milliliter. Intrathecal administration requires a lower dose as 10 mL and is more practical than IV administration, which can require more than 100 mL for a single dose.50

Recently, there have been reports of CSF and GBCA being absorbed from the choroid plexus.51,52 The absorption of IT-GBCA from the choroid plexus was reported to be delayed in patients with iNPH compared with controls.52 Serial MRI showed that the T1 signal within the choroid plexus and CSF of the lateral ventricles reached maximum values at 6–9 hours after intrathecal administration of gadobutrol in control subjects and in the patients with iNPH. The results suggest that gadobutrol is absorbed from CSF by the choroid plexus. A larger volume of the choroid plexus was associated with the severity of cognitive impairment across the spectrum of Alzheimer disease.53 The function of the choroid plexus could be a new research focus in the field of neurofluid dynamics.

Another important point in elucidation of the waste clearance system is that the movement of a solvent does not necessarily coincide with the movement of a solute. In animal studies, 17O-water enters the brain much faster than the GBCA.49 This discrepancy between water movement and solute movement is an important element when considering the waste clearance system.

PERIVASCULAR SPACE

An electron microscope study has shown that the structure of the perivascular space varies by location as well as by specific arteries or veins. The arteries in the basal ganglia are surrounded by 2 distinct leptomeninges, which are separated by a perivascular space. The inner layer of the leptomeninges adheres to the adventitia, and the outer layer is continuous with the pia mater (Fig. 2). The veins in the basal ganglia do not have an outer layer of leptomeninges, and thus, the perivascular space around the vein is continuous with the subpial space.54 The periarterial spaces in the cerebral cortex have only a single periarterial layer of leptomeninges.55

FIGURE 2: Schematic diagram of Virchow-Robin space (VRS). VRS or perivascular space is delineated by basal membranes of glia, pia, and endothelium. The VRS corresponds to the space surrounding vessels penetrating into the parenchyma. The VRS is obliterated at the capillaries where the basement membranes of glia and the capillary endothelium join. The invagination of both cortical and vessel pia into the VRS makes pial funnel. The pial sheath around arteries extends into the VRS but becomes more fenestrated and eventually disappears at the precapillary section of the vessel. Unlike arteries, veins do not possess a pial sheath inside the VRS. Interstitial fluid may drain through an intramural pathway along the basement membranes of capillaries and arterioles into the lymphatics (green arrows). SAS, subarachnoid space. This illustration was previously presented by Brinker et al30 licensed under CC BY 2.0.

Several studies have been conducted using MRI-based imaging in humans. In a longitudinal study of the association between dilated perivascular space (dPVS) and the risk of ischemic stroke and intracerebral hemorrhage in elderly subjects, an increased global dPVS burden was associated with a higher risk of any incidence of stroke and intracranial hemorrhage.56 Several studies have shown that basal ganglia dPVS is more strongly associated with hypertension and blood pressure variability, whereas white matter dPVS is more strongly associated with Alzheimer disease and cerebral amyloid angiopathy.57–59 In a retrospective MRI study of 1789 healthy participants (age range, 8–100 years), dPVS in the basal ganglia and white matter showed a biphasic volume pattern with age (lower volumes until the mid-30s, and higher volumes in older individuals).60

In the field of neuroradiology, it was long assumed that the dPVS shows the same signal intensity as CSF for any pulse sequence and does not have contrast enhancement after IV-GBCA.61 However, the distribution of GBCA into the perivascular space of the basal ganglia was shown by heavily T2-weighted 3D fluid-attenuated inversion recovery (FLAIR) imaging at 4 hours after IV-GBCA (Fig. 3).62 Subsequently, it was shown that the perivascular space of the basal ganglia had contrast enhancement after IV-GBCA, but the perivascular space of the white matter did not.63 The signal of the dPVS in white matter on precontrast heavily T2-weighted 3D-FLAIR imaging showed greater intensity than that of the dPVS in the basal ganglia. This finding suggests that the content of the dPVS in white matter is different from that of the dPVS in the basal ganglia. The signal intensity of the dPVS in the white matter on precontrast heavily T2-weighted 3D-FLAIR imaging has been suggested as a biomarker of waste clearance function in the brain.63 Alternatively, the distribution of GBCA in the perivascular space of the white matter has been demonstrated in accidental cases where high doses of GBCA have been administered intrathecally.41 A significant positive correlation between Fazekas scoring and the signal intensity increase in the perivascular spaces of the basal ganglia after IV-GBCA was reported.21 Further studies are needed to determine whether the greater degree of deep white matter hyperintensity (WMH) and the greater distribution of GBCA in the perivascular space indicate increased BBB permeability or glymphatic system congestion.

FIGURE 3:

Heavily T2-weighted 3D fluid-attenuated inversion recovery (FLAIR) images were obtained before (A) and 4 hours after intravenous administration of a single dose of gadolinium-based contrast agent (IV-SD-GBCA) (B) in a female patient in her 70s with a suspicion of endolymphatic hydrops. Cerebrospinal fluid in the subarachnoid space (short arrows, B) and perivascular space in the basal ganglia (long arrows, B) show increased signal intensity in the image obtained at 4 hours after the IV-SD-GBCA. The high sensitivity of the heavily T2-weighted 3D-FLAIR imaging to low concentrations of GBCA in fluid may facilitate investigation of the waste clearance system by the clinically feasible method of imaging with IV-SD-GBCA.

Although the perivascular space around the arteries is described as an entry point in the conceptual diagram of the glymphatic system by Iliff et al,64 there are various ideas about the direction of fluid flow. Some investigators hypothesize that the perivascular space around the arteries is an outflow tract, and alternatively, others propose that it can change direction depending on the situation.65 Therefore, the perivascular space in the basal ganglia may also function as an outlet for solutes and CSF from the brain parenchyma. Metabolic waste products, such as amyloid-β, are thought to accumulate around the blood vessel, possibly leading to perivascular blockage and enlargement of the space.66

BRAIN PARENCHYMA MRI Without the Administration of Exogenous Contrast Agents

Diffusion imaging is a powerful method to noninvasively assess interstitial fluid dynamics within the brain parenchyma. However, water diffusion in the brain parenchyma is affected by large white matter fibers, and it is difficult to evaluate fine water movement in the direction of the perivascular space. To overcome this problem, an evaluation method called DTI-ALPS (diffusion tensor image analysis along the perivascular space) has been proposed.15 This method assumes that water diffusivity within the perivascular space correlates with interstitial fluid dynamics. In the white matter outside the lateral ventricular body, the perivascular space is mainly perpendicular in the left-right (x axis) direction, with the projection fibers in the vertical (z axis) direction and the association fibers in the anteroposterior (y axis) direction. Therefore, the effects of diffusivity in the large white matter fibers can be vectorially separated from the diffusion in the perivascular space. The ALPS index was used to evaluate the degree of diffusion along the perivascular space (Fig. 4). The ALPS index is calculated by the ratio of 2 sets of diffusivity values, which are perpendicular to the dominant fibers in the white matter tissue. It is the ratio of the average of the x axis diffusivity in the projection fiber area (Dxxproj) and the x axis diffusivity in the association fiber area (Dxxassoc) to the average of y axis diffusivity in the projection fiber area (Dyyproj) and z axis diffusivity in the association fiber area (Dzzaccoc) as follows.

FIGURE 4: The concept of the DTI-ALPS (diffusion tensor image analysis along the perivascular space) method. In the white matter outside of the lateral ventricular body, the medullary vessels run in the left-right direction (A and B), the projection fibers run in the vertical direction, and the association fibers run in the anteroposterior direction (C and D). Thus, in this region, diffusion in the direction of the perivascular space of the medullary arterioles can be assessed separately from the diffusion effects of the large white matter fibers. This illustration was adapted with permission from the previous publication by Taoka et al.15

ALPS index = mean (Dxxproj, Dxxassoc)/mean (Dyyproj, Dzzassoc)

The underlying hypothesis is that a higher ALPS index indicates more efficient clearance of waste through the glymphatic system.

A study for the reproducibility of DTI-APLS showed very high reproducibility for the same imaging sequence and head position.18 In addition, as an alternative to diffusion tensor imaging, a method using 3-axis diffusion-weighted images has also been proposed to reduce the scan time.7

In the evaluation of Alzheimer disease, and mild cognitive impairment versus normal controls, the ALPS index was correlated significantly with MMSE score and inversely correlated with age.15 A recent study of Alzheimer disease and mild cognitive impairment showed that patients with Alzheimer disease had a lower ALPS index compared with controls. A lower ALPS index was also associated with lower CSF amyloid-β and lower FDG-PET uptake.14 In a comparative study of amyloid PET and ALPS, a significant negative correlation was found between the ALPS index and the standard uptake value ratios of 11C-PiB and 18F-THK5351.67 In a study of Parkinson disease, abnormal interstitial fluid dynamics has been suggested. The Parkinson disease group with mild cognitive impairment and dementia had a significantly lower ALPS index than normal controls, and the ALPS index was inversely correlated with the degree of oxidative stress as assessed by plasma DNA levels.68

In a study comparing normal controls with iNPH cases and with ventricular dilatation in non-iNPH cases, the iNPH group had a significantly lower ALPS index than that of the non-iNPH ventricular dilatation group.69 In an evaluation of the ALPS index after lumboperitoneal shunt surgery for iNPH, the mean postoperative ALPS index in the responder group was significantly higher than that found preoperatively. In the nonresponder group, the mean postoperative ALPS index was not significantly different from that found preoperatively.70

The ALPS method has several limitations. One is that the ALPS index measurements are affected by the imaging plane and head position.18 In this regard, a method to retrospectively correct the angle has been proposed.71 In addition, the manual ROI setting is unstable and may affect the ALPS index measurements. A solution by standardizing the images to Montreal Neurologic Institute coordinates has been proposed.14 Although the ALPS method has been criticized for lack of validation, a good correlation with intrathecal administration of GBCA has been reported.72

Other diffusion imaging methods for the evaluation of interstitial fluid have been reported, including free water imaging.73 One study investigated if increased interstitial fluid mediates the association between dPVS and WMH volumes. The free water fraction mediated this association over all subjects and particularly in subjects with a relatively high WMH load.74 In a report comparing standard DTI and free water imaging indices in Alzheimer disease, improved consistency was observed with fractional anisotropy, as well as axial and radial diffusivity.75 Free water DTI better reflects the underlying pathology of Alzheimer disease and improves the accuracy of DTI metrics related to white matter integrity in Alzheimer disease.

MRI With the Administration of Exogenous Contrast Agents

When contrast agents are used to observe glymphatic function in the brain parenchyma of humans, IT-GBCA is most commonly reported.37,45,46,76,77 Many studies using IT-GBCA have shown delayed elimination of GBCA from the brain parenchyma in iNPH.13,52,78 Delayed excretion of GBCA from the brain in humans with sleep deprivation has also been confirmed using IT-GBCA.10,79

The transfer of GBCA into the brain parenchyma after IV administration is still under investigation in humans because the amount transferred is very small, and also it is affected by BBB integrity and renal function.5,80 In animals, both linear and macrocyclic agents have been shown to enter the brain after IV administration.29 In humans, it has been shown from both imaging and CSF samples that IV-GBCA can distribute into CSF space, even in subjects without disruption of the BBB.21,62,81,82 In a study in which CSF samples were obtained after IV administration, higher GBCA concentrations in the CSF were associated with the subject's age of 18 years or older.82 The analysis of GBCA kinetics after IV administration is complicated because some amounts of GBCA, which leak into the CSF, may enter the brain via the glymphatic system, whereas others may be eliminated by the meningeal lymphatics.5,25

One report compared the elimination of GBCA from the brain parenchyma between sleep and wakefulness from MP2RAGE (magnetization-prepared 2 rapid acquisition gradient echo) T1 maps obtained at multiple times before and after IV-GBCA. A comparison of the slope of the T1 values between 2 and 12 hours post–IV-GBCA showed that the clearance of GBCA was greater during sleep compared with wakefulness.11 In the future, quantitative methods such as MR fingerprinting and synthetic MR may be used for such studies.83,84

The GBCA permeability through the BBB increases with age.85–87 Leakage of IV-GBCA into the CSF is also more pronounced by aging.26,88 If IV-GBCA is used to evaluate the glymphatic system function in the brain parenchyma, it is necessary to exclude confounding factors such as permeability of the BBB and B-CSF-B, aging, and renal function. However, if we only want to estimate how much waste is present in the brain parenchyma, we might be able to do so by evaluating the GBCA concentration in the CSF after IV-GBCA. The GBCA concentration in CSF after IV administration might be a practical biomarker to evaluate the BBB, B-CSF-B, aging, renal function, and glymphatic system function collectively as a whole.5

DOWNSTREAM COMPONENTS OF THE GLYMPHATIC SYSTEM Evaluation of the Perivenous Drainage Pathway by MRI Without the Administration of Contrast Agents

For the visualization of the perisinus drainage route without the administration of contrast agents, 3D-FLAIR sequences have been used to image the protein-rich meningeal lymphatics along the superior sagittal sinus.89 However, 3D-FLAIR imaging with a T2-selective inversion recovery pulse experiences an incomplete suppression of CSF near bone or air.90 Thus, the areas of high signal intensity near the skull base seen with this imaging modality might include the depiction of artifacts.

Another contrast independent method used multicomponent T2 analysis.91 The authors speculated that the perivenous drainage route contains protein-rich interstitial fluid, which has a shorter T2-value compared with CSF. The voxels with a shorter T2-value could be discriminated from voxels with a longer T2-value using a CPMG (Carr-Purcell-Meiboom-Gill) imaging sequence with a signal decomposition into 25 T2 components. Shorter T2 components were visualized along the cortical and bridging veins, as well as along the superior sagittal sinus.91

Morphologically, cystic structures in the subarachnoid space along the superior sagittal sinus and bridging veins in heavily T2-weighted MR cisternography have been reported.19 These parasagittal sinus cysts or parasagittal perivenous cysts have been reported to be continuous to the perivenous drainage route around the cortical veins with noncontrast enhanced heavily T2-weighted MR cisternography.92 The number of cysts contacting the bridging veins in non–contrast-enhanced heavily T2-weighted MR cisternography was reported to be a potential biomarker for impaired drainage function.19

Evaluation of the Perivenous Drainage Pathway to the Parasagittal Dura by MRI With IV-GBCAs

For the most part, perivenous contrast enhancement after IV-GBCA has been reported using FLAIR sequences.93,94 The perivenous CSF enhancement at 4 hours after IV-GBCA was shown to be age dependent, which suggests that this enhancement correlates with perivenous drainage.88,95 A direct connection of the perivenous drainage pathway to the meningeal lymphatics along the superior sagittal sinus was suggested visually using MRI with IV-GBCA.96

Meningeal Lymphatics

The discovery of the meningeal lymphatic vessels, which were previously thought to be absent from CNS, has attracted great interest from the scientific community.97,98 Recent characterization of the glymphatic and meningeal lymphatic systems in rodents and humans has facilitated the reevaluation of the anatomical routes for CSF-interstitial fluid flow and the physiological role that these pathways play in CNS health. In humans, the meningeal lymphatics along the superior sagittal sinus have been visualized with the IV administration of regular extracellular-type GBCA, but not with the IV administration of intravascular-type GBCA.99 Intravascular-type GBCA is a serum albumin-binding contrast agent that remains largely intravascular. The lack of enhancement of tubular structures along the superior sagittal sinus by intravascular-type GBCA means that these structures are not blood vessels.

Dynamic change of signal intensity for the glymphatic-lymphatic system along superior sagittal sinus has been measured before and after IV administration of regular extracellular-type GBCA in 20 volunteers.100 Signal intensity of glymphatic-lymphatic system increases immediately after IV administration of GBCA and decreases 1 hour after the administration. This might reflect the clearance function of the glymphatic system.

Drainage of IV-GBCA through the meningeal lymphatics was delayed in patients with idiopathic Parkinson disease.101 Recently, perisinus lymphatic tissue was visualized in humans, not only along the superior sagittal sinus, but also along the transverse and sigmoid sinuses using IV-GBCA.102 Dilated meningeal lymphatics along the sigmoid sinus were associated with delayed clearance of GBCA from the CSF. Dilated meningeal lymphatics along the sigmoid sinus might be a sign of impaired waste drainage (Fig. 5).103

FIGURE 5:

Band-like structures with high signal intensity can be visualized between the sigmoid sinus and cerebellar hemisphere from a male patient in his 60s. This patient underwent MRI for the suspicion of endolymphatic hydrops in the inner ear. The 3D real inversion recovery (IR) images were obtained before (A), 10 minutes after (B), 4 hours after (C), and 24 hours after (E) the intravenous administration of a single dose of gadolinium-based contrast agent (IV-SD-GBCA). An MPRAGE image was obtained immediately after the IV-GBCA (D). The signal of the band-like structures shows the highest intensity in the 3D real IR image obtained at 4 hours after the IV-GBCA (arrows in A, B, C, E). On the MPRAGE image, the band-like structures have a lower signal intensity, which is different from the vascular signal (arrow in D). The presence of this band-like structure was reported to be associated with the delayed clearance of GBCA from the CSF and the presence of contrast enhancement in the perivascular space of the basal ganglia. The band-like structures are suggested to be a sign of impaired waste clearance. This structure is presumed to be dilated meningeal lymphatics along the sigmoid sinus wall.

Parasagittal Dura and Surrounding Structures

Ringstad and Eide104 have reported that IT-GBCA is absorbed from CSF by the parasagittal dura.104,105 The parasagittal dura contains many fluid-filled channels. The meningeal lymphatics exist between the wall of the superior sagittal sinus and the parasagittal dura.104,106 The parasagittal dura is speculated to be a reservoir for CSF absorption.106 Contrast-enhanced T1-weighted black blood imaging can visualize the parasagittal dura or perisinus lymphatic tissue in humans, and the volume of this structure has been demonstrated to increase with aging.107 IT-GBCA was shown to distribute in the bone marrow of the skull near the parasagittal dura.108 This overlap of meningeal and skull barriers suggests that bone marrow may contribute to brain immune surveillance in humans. From the results of a histological analysis of human cerebral arachnoid granulations near the parasagittal dura, the arachnoid granulations were suggested to be lymphatic conduits, which communicate with bone marrow and dura-arachnoid stroma.109

Cystic structures near the parasagittal dura and bridging veins have been reported in delayed contrast-enhanced images after IV-GBCA.5,20,92,110 The cystic structures were associated with increased GBCA leakage into the CSF from the perivenous subpial space (Fig. 6).110 These cystic structures were speculated to be a part of a fluid drainage route from the brain.20 Histopathological analysis of the cystic structures is necessary to better understand their function. A case report indicating the coexistence of parasagittal sinus cysts with prominent dPVS in white matter has been also published.111 Parasagittal sinus cysts were suggested to be a cause or a result of subpial fluid flow obstruction.

FIGURE 6:

3D real IR images obtained at 4 hours after the intravenous administration of a single dose of gadolinium-based contrast agent (IV-SD-GBCA). The GBCA has leaked into the CSF in the subarachnoid space around the cortical veins (arrows, A and B). Cystic structures (short arrows in B, C, and D) can be seen along the superior sagittal sinus and cortical veins. The meningeal lymphatics are visualized along the superior sagittal sinus as linear structures with high signal intensity (long arrows in C and D). The subpial perivenous drainage route (dotted arrow in C) can be visualized along the cortical vein.

The Role of Cervical Lymphatic Ducts

For the evaluation of lymphatic drainage from the brain parenchyma, changes in the signal from chemical exchange saturation transfer (CEST) MRI of the brain were assessed after deep cervical lymph node ligation in animals and were correlated to behavior. In a pilot in vitro study, it was shown that the CEST effect of the lymph fluid is significantly greater than that of blood, CSF, or distilled water.112 The results of an in vivo study showed that the intensity of the CEST effect was significantly higher in the hippocampus of the ipsilateral side of the deep cervical lymph node ligation than in the contralateral hippocampus.112 The correlation between the signal abnormality and the behavioral score was significant. These results support the use of CEST MRI as a tool to assess the brain's glymphatic system and to predict glymphatic system dysfunction.112 In humans, it was shown that IT-GBCA drains into cervical lymph nodes using serially obtained MRI.113 It is expected that CEST MRI will be used in the future to investigate how neck lymphadenectomy affects the waste excretion from the brain.

OCULAR GLYMPHATIC

It has long been hypothesized that glaucoma, Meniere disease, and iNPH may be diseases in which there is an excess of local water.114 Many articles have been published on iNPH suggesting a relationship to the glymphatic system.12,13,69,78 In an animal study, it was shown that drainage along the optic nerve of an amyloid-β tracer injected into the vitreous was facilitated by light stimulation and associated constriction of the pupil.115 Due in part to this result, the previously proposed concept of an ocular glymphatic system became more widely accepted.115–118 In humans, MR studies using IV-GBCA have indicated penetration of the GBCA into the anterior chamber of the eye and the subarachnoid space around the optic nerve in healthy subjects.21,81,119 It was also shown in humans that IV-GBCA often leaks into the vitreous from the ora serrata in the inferior temporal side of the eyes (Fig. 7).120 This leakage may be related to aging, but the reason for the predominance on the inferior temporal side is unclear. It has also been reported in some patients that retinoblastoma extending into the vicinity of the optic nerve papilla inhibits the ocular glymphatic system and delays GBCA efflux from the anterior eye segment.121

FIGURE 7:

A healthy male volunteer in his 50s. Heavily T2-weighted 3D-FLAIR images were obtained before, and 10 minutes, 1.5 hours, 3 hours, 4.5 hours, and 6 hours after an intravenous administration of a single dose of gadolinium-based contrast agent (IV-SD-GBCA). Punctate enhancement near the ora serrata of the lateral side in both eyes appears at 10 minutes after the IV-SD-GBCA (arrows). This enhancement gradually spreads to the deep area of vitreous (short arrows in later phase images). In another male volunteer of a similar age, no enhancement near the ora serrata was seen (not shown). IV-GBCA might have the potential to evaluate the integrity of the blood-retina barrier and the ocular glymphatic system.

Changes in the ocular signal after the administration of 17O-water eye drops have also been reported in humans. Currently, this research is accomplished by proton detection MRI using an indirect effect of T2-shortening by 17O-water. Signal changes in the anterior chamber of the eye have been observed consistently, and signal changes in the vitreous have not yet been seen.122 In the future, it is expected that deeper tissue depths will be observable with improvements in 17O-water dosing methods and MR detection sensitivity. The visual impairment caused by glaucoma, including glaucoma associated with severe myopia, is a great burden to mankind, and it is expected that the development of research from the viewpoint of the ocular glymphatic system will help overcome glaucoma.117

LYMPH FLUID IN THE INNER EAR

Meniere disease is a disorder of unknown cause characterized by recurrent attacks of rotational vertigo, hearing loss, and a sense of ear fullness. The pathology is characterized by an enlargement of the endolymphatic space in the inner ear (ie, endolymphatic hydrops). In most cases, the disease is progressive and eventually leads to severe hearing loss.123,124 Endolymphatic hydrops in a patient with Meniere disease on MRI was depicted first by 3D-FLAIR imaging at 24 hours after an intratympanic GBCA administration.125 Subsequently, endolymphatic hydrops in patients was successfully depicted by imaging at 4 hours after IV double-dose GBCA administration, which did not require puncture of the tympanic membrane.126 In the same year, endolymphatic hydrops was visualized in patients by imaging at 4 hours after the IV administration of a single dose of GBCA, which has the potential for use in daily clinical practice.127 With these developments in MRI, the visualization of endolymphatic hydrops is now clinically feasible. Delayed contrast-enhanced MRI of endolymphatic hydrops is now included in the Japanese diagnostic criteria for Meniere disease.128

Subsequently, heavily T2-weighted 3D-FLAIR and 3D real inversion recovery imaging was performed in thousands of cases to detect trace amounts of GBCA in a super delayed phase, at 4 hours after IV-GBCA for the purpose of endolymphatic hydrops diagnosis.129–131 The clear images obtained by HYDROPS processing132 and HYDROPS-Mi2 processing,133 in which the MR cisternography was multiplied with the HYDROPS image, were further augmented by noise reduction techniques using deep learning reconstruction methods.134,135 The combination of these techniques produced images with an extremely high signal-to-noise ratio, which were easily obtained within a short time (Fig. 8).

FIGURE 8:

Endolymphatic hydrops in the cochlea and vestibule (arrows) was visualized in a patient with Meniere disease. The black areas are the enlarged endolymphatic spaces (arrows). This image utilized HYDROPS-Mi2 processing and a deep learning denoising technique with 3 T MRI. The image was obtained at 4 hours after the intravenous administration of a single dose of gadolinium-based contrast agent (IV-SD-GBCA). The GBCA was distributed mainly in the perilymphatic space, not in the endolymphatic space. The total scan time to obtain the data for this processed image was 20 minutes. With the development of these MRI techniques, the visualization of endolymphatic hydrops is now clinically feasible.

The aforementioned detection techniques for low concentrations of GBCA in fluids made it possible to detect contrast effects in the CSF, subarachnoid space, anterior eye segment, perivascular space in the basal ganglia, and so on.5,20,26,62,81,88 These findings have led to studies for the evaluation of waste clearance, including the glymphatic system, in humans using MRI with IV-GBCA.11,21 Results have been reported on the serial observations of the influx and efflux of GBCA into and out of the lymphatic fluid of the inner ear labyrinth from preadministration to 24 hours after IV administration.136

With an intratympanic injection of 17O-water, the distribution of the injected tracer in the labyrinth can be observed over time using the T2-shortening effect from the indirect effect of 17O-water on proton MRI.137 The tracer 17O-water is a stable isotope and does not emit radiation.38 Unlike GBCA,