A conversation on human brain clearance

Abstract

Brain clearance is essential for eliminating metabolic byproducts and macromolecules to maintain brain homeostasis. In this Q&A, four experts share their views on how the brain clears waste products, why these clearance pathways are essential for maintaining brain health, and how their dysfunction may contribute to neurological diseases. The discussion explores current imaging approaches, key measurement challenges, methodological limitations, and emerging technologies that could improve our understanding of brain clearance in humans as well as diagnosis, monitoring, and treatment of brain disorders.

What is brain clearance and how is it important in human brain health and disease?

Matthias van Osch (Professor in Radiology and director of the C.J. Gorter Center for high field MRI, Leiden University Medical Center, Leiden, The Netherlands): In general terms, brain clearance refers to the disposal of neuronal waste products from the brain and thereby includes CSF-mediated clearance (‘glymphatics’), autophagy, ubiquitination, and direct transport across the blood-brain-barrier (BBB). In recent literature, the focus is especially on CSF-mediated clearance, widely known under the term ‘glymphatics’. By many, the terms ‘brain clearance’ and ‘glymphatics’ are used interchangeably, although brain clearance should clearly be considered a broader term.

The people portraited in the pictures and their affiliations are as follows (in order left to right): Matthias van Osch, Professor in Radiology and director of the C.J. Gorter Center for high field MRI, Leiden University Medical Center, Leiden, The Netherlands. Laura Lewis, Associate Professor, Massachusetts Institute of Technology and Massachusetts General Hospital, USA. Geir Ringstad, Professor of Radiology, University of Oslo, Norway. Vesa Kiviniemi, Professor of Functional Neuroimaging, University of Oulu, Finland.

“By many, the terms ‘brain clearance’ and ‘glymphatics’ are used interchangeably, although brain clearance should clearly be considered a broader term”.

As the accumulation of waste products, like amyloid-beta and tau, is an important hallmark of many neurodegenerative diseases, it is very important to understand the relevant underlying brain clearance mechanisms and failures of these.

Laura Lewis (Associate Professor, Massachusetts Institute of Technology and Massachusetts General Hospital, USA): The brain is highly metabolically active and continuously generates waste products. Brain clearance is the process by which metabolites are removed from brain tissue, maintaining homeostasis. This clearance process is important in the healthy brain to maintain an extracellular environment that supports functioning neuronal activity. Impaired clearance is associated with neuroinflammation and neurodegeneration, so understanding the mechanisms that regulate clearance is important for multiple clinical conditions.

Geir Ringstad (Professor of Radiology, University of Oslo, Norway): Brain clearance refers to the set of biological processes by which the brain removes potentially pathogenic waste products, but also excess fluid and other metabolites from the brain tissue. It is important to emphasize that brain clearance is not a single pathway. In recent years, much attention has been given to glymphatic mechanisms, where CSF and interstitial fluid (ISF) exchange along perivascular routes. These pathways are likely important, particularly in cortical regions having the largest degree of CSF-ISF exchange, but they represent only one component of a broader clearance system.

More widely, human brain clearance also includes transport across the BBB, blood–CSF barrier, and arachnoid barriers; drainage through meningeal lymphatic and extracranial lymphatic routes; bulk movement and turnover of CSF; local cellular mechanisms, including uptake, degradation, and phagocytosis; and immune surveillance at the brain borders and meninges. These processes may be partly overlapping and may become more or less important depending on molecular size, anatomical compartment, sleep–wake state, vascular pulsatility, inflammation, ageing, and disease. This complexity is important to keep in mind when interpreting experimental or imaging-based measures of clearance.

The relevance to human brain health is that the brain has high metabolic activity but no conventional lymphatic infrastructure within the parenchyma. Efficient clearance is therefore essential for maintaining fluid homeostasis, limiting neuroinflammation, and preventing accumulation of potentially toxic proteins. Conversely, impaired clearance has been implicated in ageing, small vessel disease, hydrocephalus, traumatic brain injury, and neurodegenerative disorders such as Alzheimer’s disease, where failure to remove amyloid-β, tau, or other proteins may contribute to disease initiation or progression.

MRI can provide important in vivo information about some clearance processes, but no single MRI measure captures the entire clearance system. A major knowledge gap is the extent to which each pathway contributes to overall brain clearance, and how these contributions differ between brain regions, molecular species, physiological states, various age groups and disease conditions.

Vesa Kiviniemi (Professor of Functional Neuroimaging, University of Oulu, Finland): Prior to the discovery of the glymphatic system in 2012 by Drs. Nedergaard and Iliff, much of the mammalian brain solute clearance was not known. For a long time, it was assumed that most of the soluble waste from interstitial space was removed by slow diffusion driven by Brownian motion somehow and somewhere. The research led by Dr. Nedergaard showed that fluids in interstitial space interacted with perivascular cerebrospinal fluid (CSF) that then remove solved proteins and other metabolites by advective speeds along the perivascular CSF spaces driven by physiological pulsations. Aligning strongly with those preclinical studies, research led by Drs. Per Eide and Geir Ringstad’s showed similar findings in humans.

Original works on mechanisms driving the solutes in the subarachnoid space indicated cardiovascular pulsations along periarterial spaces in mice. Human studies have shown that the brain clearance functionality seems to be guaranteed by three different pulsation mechanisms driving solutes out of the brain, namely vasomotor waves, respiratory and cardiovascular pulsations, the perivenous respiratory pulsations are spatially the most dominant, involving also the deeper white matter outside the gray matter¹. During sleep, the vasomotor and respiratory pulsations increase in power and speed up the solute flushing from the human brain sensory areas².

Already daily living causes some wear and tear on the brain and needless to say that most (if not all) pathological brain conditions can cause damage that needs to be both cleared in order to have a fully functioning brain¹. Thus, we need to take care of these driving pulsations that clear the brain; we need to sleep 7–8 h consistently in preferably dark room and would be beneficial to follow a healthy and active lifestyle with reduced stress and sugar intake levels maintaining low blood pressure for an efficient brain cleaning to occur. Albeit the brain clearance does reduce over the years leading to solute accumulations, a healthy lifestyle can slow down this process. Hopefully, there will also be novel therapies capable of boosting the physiological slowing of brain solute efflux.

How can we assess brain clearance in humans?

MvO: Three main approaches can be identified:

1. Measurement of waste products in CSF obtained from lumbar punction or in blood plasma. Concentrations are certainly influenced by brain clearance, but certainly not exclusively, making interpretations difficult.

2. Medical imaging involving exogenous tracers. By tracking a tracer dynamically, we can monitor the journey of the tracer throughout the brain, including the brain clearance system. When injected into the CSF, i.e. intrathecally, a significant minority of the tracer (about one-fourth) will enter the brain and will subsequently be cleared. When injected intravenously, the pathways are more complicated and, for example, dependent on the condition of the BBB. For more extensive discussion on the pathways taken by intrathecal or intravenous injections³. The main modality to monitor the passage of exogenous tracers is MRI with gadolinium-based contrast agents. PET is another modality that is being used, although more limited due to the higher level of invasiveness and costs.

3. Measuring CSF dynamics. As in the glymphatic system, CSF is thought to be the main solvent that transports waste products out of the brain, measurements of the mobility of CSF are an important source of information, just as the resistance that would be experienced by CSF/ISF as measured by electrical impedance spectroscopy. CSF dynamics can be measured at many different locations within the brain, including the perivascular spaces of penetrating arteries⁴ and just below the cerebrum, like the fourth ventricle. Also, other processes influencing the mobility of CSF/ISF are areas of active research, such as the water transport from blood into brain tissue and from blood into CSF, as well as deformation of brain tissue, for example, due to the cardiac and respiratory cycle. The main modality to study CSF/ISF dynamics is MRI.

LL: Directly measuring brain clearance in humans is very challenging. In the past few years, many exciting new tools have been developed, each of which can measure a piece of the puzzle, although not the whole system. One strategy is to perform multiple PET scans to track the accumulation of endogenous metabolites such as amyloid. Alternatively, using MRI, some studies now inject contrast agents, either into the blood or into the CSF directly, and then track how that contrast agent redistributes in the brain and ultimately exit it. MRI strategies can also be used to measure fluid flow and velocity, and fluid diffusion within the tissue. Measuring these fluid dynamics can then allow modeling of where CSF goes and how it shapes transport in the brain.

GR: Brain clearance is multi-component, therefore its assessment requires a combination of methods that each interrogate different aspects of the system. Broadly, available tools to assess brain clearance in humans include MRI-based methods, PET imaging, CSF and blood biomarkers, and computational modeling.

MRI is particularly useful for studying fluid movement and solute transport in the CSF and perivascular pathways, which is where much of my own research has been focused. The most direct human approaches are tracer-based⁵. These methods assess the movement and clearance of an exogenous, intrathecally injected solute (MRI contrast agent), which may serve as an experimentally tractable surrogate for endogenously produced waste products from brain metabolism. However, such interpretation requires caution, because molecular size, charge, binding properties, and anatomical compartment may strongly influence transport and clearance.

Other MRI methods focus primarily on imaging water, the solvent, rather than solute clearance itself. These approaches can assess directional water mobility, as with diffusion-based MRI; water exchange between blood, brain tissue, and CSF; bulk CSF flow using phase-contrast MRI; brain pulsations from slow vasomotions (MREG) or visualize water-containing structures such as MRI-visible perivascular spaces. Such methods rely on the assumption that water movement or fluid-space morphology provides information relevant to clearance of larger solutes. This assumption may be plausible in some settings but should not be equated with direct measurement of solute clearance.

PET imaging may provide complementary information about molecular transport, BBB function, protein deposition, receptor systems, or neuroinflammation. However, PET markers of microglial activation, amyloid, or tau burden do not directly measure cellular clearance. Similarly, CSF or blood biomarkers can reflect downstream consequences of altered clearance, but do not necessarily reflect instantaneous clearance capacity, and can rarely identify the specific anatomical route or cellular mechanism involved.

Thus, the field currently relies on converging evidence from several imperfect but complementary tools. A useful way forward is to be precise about what each method actually measures, and equally precise about what it does not measure.

VK: The golden standard is the intrathecal gadolinium (Gd) tracer for MRI scans. Intrathecal tracer studies are invasive but safe techniques in experienced hands, technically somewhat laborious, and relatively slow for clinical routine use as they can last hours to days.

Another recently emerged method is FLAIR (or dark fluid (DF) imaging for Siemens scanners) MRI, where free water is labeled black, while water interacting with macromolecules like proteins in brain areas of edema is very bright. FLAIR can be used to detect parasagittal dural space (PSD), which is a special dural pouch in humans on the side of sagittal sinus full of protein-rich CSF. This space does not exist in mice, only in mammals with large brains.

Novel water selective phase/diffusion weighted MRI imaging techniques, such as the new EPTI and CSF STREAM MRI techniques, especially at 7T (or higher), offer high precision and quantifiable water diffusivity scans without tracers in 40 min or less. But only free CSF spaces are visualized; brain parenchyma is black in long echo time imaging enabling precise perivascular and sulcal CSF spaces tracing and flow quantification.

Ultrafast 10 Hz whole-brain functional MRI using magnetic resonance encephalography (MREG) can be used to detect pulsatile water molecule movement in both the brain parenchyma and CSF in a clinically feasible and non-invasive timescale. MREG enables detection of multiple physiological properties of brain neurofluid flow including velocity and power of vasomotor waves, respiratory fluctuations and cardiac pulsations, all affecting CSF solute advection^1,2.

However, MRI is expensive, has limited availability and requires specialized facilities. Thus, wearable, wireless brain hydrodynamic sensing technology has been developed for easier, more versatile assessment of brain solute dynamics. Such technology utilizes fNIRS, dcEEG and impedance measurements offering ways of analyzing brain hydrodynamics and pulsatility bedside, at home or even in upright position during light activity².

What aspects/routes of brain clearance can be measured?

MvO: Probably the most important observation is that there is currently not a human approach that probes the complete brain clearance system or glymphatic pathway: only sub-parts are studied. It is therefore of the utmost importance that researchers are very clear about what part of the brain clearance system their approach is probing. In my answer to question 3, I already highlighted the most important parts of the brain clearance system that can be measured.

LL: What we can do most readily in humans is: 1) measure tracer concentrations in the brain and 2) measure large-scale fluid flow. First, tracer injections allow tracking the concentration of how the tracer redistributes in brain tissue and is eventually cleared away. This enables inferences about how solutes in the CSF are transported, which is essential for understanding clearance of the contents of the CSF. This approach also enables investigation of specific transport routes, for example watching tracer spread along perivascular spaces, and exiting through meningeal lymphatic vessels. Second, MRI is a great tool for tracking water movement, and CSF is mostly water. CSF flow can be readily measured in large-scale spaces, such as the ventricles and the aqueduct⁶, where CSF travels long (> cm) distances. This large-scale movement is an important aspect of the fluid system, since the human brain is large (e.g. >15 cm long) and large-scale flow is needed to transport solutes over those long distances.

GR: Several aspects of brain clearance can be measured or inferred in humans, but with very different levels of directness. The routes that are currently most accessible are those related to the CSF compartment: the anatomy of spinal and cranial CSF spaces, subarachnoid transport, MRI-visible perivascular spaces at different levels, and elimination of tracers from the intracranial compartment, including the brain, to lymphatic efflux routes. By contrast, BBB-mediated clearance and local cellular clearance by microglia, astrocytes, or macrophages are much harder to measure directly in humans. These processes can be studied indirectly through permeability imaging, PET, biomarkers, or histopathology, but such methods usually do not provide a direct measure of waste removal.

Intrathecal contrast-enhanced MRI allows perhaps the most direct visualization of human brain clearance. After intrathecal injection, imaging can demonstrate how an exogenous solute distributes within the spinal and cranial CSF spaces, along brain surfaces, around vessels, and into selected brain regions, with delayed assessment of tracer elimination. This has made it possible to study aspects of CSF-mediated, extra-vascular solute transport and clearance in vivo. However, the method also has important limitations. Tracer pharmacokinetics in CSF and brain evolve over many hours to days, so the temporal resolution is limited. Imaging several time points over one to 2 days is demanding for patients and costly in clinical research settings. Another limitation of intrathecal MRI contrast agents is that their use remains off-label.

A consistent observation from our studies is that tracer clearance from brain tissue is strongly dependent on tracer clearance from CSF at the brain surface. Based on this, we, at Oslo University Hospital (Oslo, Norway), have developed a test that estimates CSF tracer clearance to blood⁷. A small amount of tracer is injected intrathecally, followed by repeated blood sampling during the first hours after injection. Using a pharmacokinetic model we have built based on experiences from more than a thousand blood samples, we can then estimate a CSF clearance curve for each subject. So far, we have observed striking differences between patient groups and individuals within the same group.

This approach may be more clinically translatable than serial MRI, because it is independent of brain imaging and done with other sorts of tracers than MRI contrast agents. We are therefore also validating the model using CT contrast agents, which are already approved for intrathecal enhanced imaging and could make the methodology more widely applicable. Preclinical studies suggest that CSF clearance to blood may primarily reflect dural lymphatic clearance function. An important strength of the method is that it does not depend on one specific anatomical route. It measures the net clearance capacity of a solute from CSF to blood, regardless of the underlying pathway.

VK: The cortical CSF space and cranial nerve aspects of the solute clearance can relatively well now be imaged with Gd+ tracer MRI and SPECT/PET imaging. Ultrafast MREG, and also novel transcranial HD ultrasound imaging can deliver critically sampled precision views in both animal and even human pulsation mechanisms without aliasing. MRI sequences with ultrahigh speed and/or spatial precision are making the imaging increasingly precise especially around the brain. Also, recent advances in high-sensitivity blood biomarker laboratory tests bring much easier and safer access to relevant protein biomarker changes in addition to CSF sampling for advanced analytics, if imaging data.

What are the current limitations?

MvO: The most important limitation of human brain clearance methods based on exogenous tracers is that in humans the tracer cannot be injected at the location where neuronal waste products are produced. Moreover, the employed tracers are frequently much smaller than e.g., amyloid-beta and have therefore different properties, which will probably have implications for the clearance efficiency and pathways.

The single most important limitation of imaging methods that measure CSF dynamics is that it doesn’t inform us on how it relates to the efflux of waste materials: at best it just measures the dynamics of the solvent. Other limitations are that many MRI methods do not exclusively measure signal from CSF or ISF, which makes the interpretation complicated. Moreover, the location where the measurements are performed will limit which part of the brain clearance system is being measured. For example, measurements of CSF-velocities at the base of the brain tells us on the communication of CSF between the cranium and the spine but is not providing information on how the CSF is moving at the tissue level.

Of course, contrast agent-based methods are limited by a certain level of invasiveness, which will also limit their ability to perform repeated measurements. Intrathecal studies involve off-label use of MR contrast agent and can show side effects common to lumbar punctions, although the method should still be considered a safe procedure.

Moreover, it is important to realize that human brain clearance methods differ with respect to the temporal footprint. For example, intrathecal measurements measure the in- and outflow of tracer over 24–72 h and will therefore be blind to how for example the cardiac or respiratory cycle influences the clearance of the tracer. The same holds true for liquid biomarkers, such as amyloid levels in CSF or blood. On the other hand, many non-invasive MRI techniques that measure CSF mobility allow to be triggered (prospectively or retrospectively) with respect to physiological processes and are therefore better suited to study the driving forces of CSF-mediated clearance.

Finally, an important limitation of the field of human brain clearance imaging is the very high level of interest fueled by the important implications for understanding pathological processes underlying neurodegenerative diseases. This sounds strange, but in my opinion the field is vulnerable to hypes and overinterpretation of results, as for example also noticed by the first author of the DTI-ALPS technique⁸. The many publications employing DTI-ALPS as a measure of glymphatic activity has maybe more clouded our knowledge than helped in understanding the complexity of the brain clearance system and its involvement in brain diseases.

LL: A key challenge is obtaining measures that are clearly physically interpretable—that is, they measure a specific physiological process, and not just a proxy metric that can be influenced by multiple factors. For example, some metrics from diffusion MRI might reflect changes in how water moves in the extracellular space, but they could also be influenced by white matter anatomy, blood vessels, or motion, meaning that multiple factors could change this metric⁸. What we ideally need is methods that can precisely report quantitative values about diffusion and flow, with high spatial and temporal precision, and without being influenced by confounding signals from the blood, so that we can build accurate models of how fluid transport takes place across multiple spatial scales in the brain. Much ongoing work in MRI is focusing on this challenge, aiming to make imaging methods more biologically precise⁶.

“A key challenge is obtaining measures that are clearly physically interpretable—that is, they measure a specific physiological process, and not just a proxy metric that can be influenced by multiple factors”.

A second limitation is that we can’t separately measure entry and exit from the brain in humans. For example, with tracer injections, the tracer first moves into brain tissue before eventually being cleared out. Ideally, we would be able to measure how molecules in the tissue are removed from the brain; but this can’t be isolated with current methods, since we have to get the tracer into the brain first. It’s hard to envision how to do that noninvasively, so this has been a longstanding limitation in human imaging.

Finally, another important need is high-resolution imaging. Many of the key structures involved in clearance are too small to be visible on conventional MRI scans. Recent studies have aimed to push the boundaries of spatial resolution to address this, for example, by developing long, ultra-high field MRI scans that can resolve perivascular spaces. Temporal resolution is also critical. Many MRI scans construct an image using measurements that last 10 minutes or longer. However, fluid systems in the brain are highly dynamic: flow is constantly oscillating and switching directions, and flow patterns can change completely within a few seconds, for example, if the person falls asleep or becomes drowsy⁹. We need imaging techniques that have high temporal resolution, in order to understand this very dynamic fluid system.

GR: As mentioned above, all current methods have limitations, and they measure different aspects of brain physiology with variable relevance to brain clearance. A major limitation of the field is that most studies have focused on transport along perivascular conduits, often framed as glymphatic clearance, whereas local degradation, cellular uptake, BBB transport, and clearance of specific disease-relevant solutes remain poorly accessible in humans.

I have already described limitations with intrathecal contrast-enhanced MRI. The on-label alternative for analyzing exogenous tracer transport is intravenous contrast-enhanced MRI. However, the BBB is not completely impermeable and intravenously administered contrast agents are known to leak from blood into brain tissue. Although this approach has been proposed to exploit contrast leakage into CSF, and assess the entry of leaked contrast into brain, it creates a fundamental compartmentalization problem: contrast may already be present in brain tissue after leakage from the vascular side. From a patient safety perspective, an intravenous dose also gives a substantially higher total body contrast exposure than low-dose intrathecal administration, while brain tissue enhancement may be in the same order of magnitude.

Non-invasive, water-based clearance imaging has generated many interesting results, but some interpretations of what these methods can tell us about brain clearance function have, in my opinion, been overstated. For instance, the DTI-ALPS method is based on directional measurements of water diffusivity in deep cerebral white matter on the left side, yet it is still widely considered a surrogate marker of CSF-driven, brain-wide clearance of amyloid-β and tau from the cortex. This is a major conceptual leap. In a recent paper, our group showed that CSF tracer influx to the ALPS region is small or absent, and that the DTI-ALPS index did not reflect tracer clearance¹⁰. This does not mean that DTI-ALPS is uninformative, as it may measure a feature of cerebral white matter in a new way, but it should not be equated with cortical glymphatic clearance.

A similar problem applies to counts or volumetric assessments of MRI-visible perivascular spaces. These spaces are not visible in the cortex, but primarily in cerebral white matter, the basal ganglia, and the mesencephalon. Yet the main reason brain clearance is of such interest is its potential relevance to neurodegenerative diseases, particularly Alzheimer’s disease, where pathology is predominantly cortical. One proposed rationale is that MRI-visible enlarged PVS in cerebral white matter could reflect upstream obstruction of perivascular clearance in the cortex. However, white matter PVS surround medullary arteries, which do not simply represent downstream continuations of cortical arteries, and most cortical arteries do not connect with white matter arteries. It is therefore likely that enlarged white matter PVSs often reflect mechanisms other than upstream obstruction of cortical perivascular flow.

Given the limitations of both non-invasive approaches and intravenous contrast-enhanced MRI, we experience there is increasing willingness at several centers to consider intrathecal tracer administration, and some have already started. At the same time, non-invasive methods can provide important complementary information when interpreted within their biological limits. For example, CSF STREAM has helped characterize CSF flow mobility within the perivascular subarachnoid space, a compartment that appears to have a functional role in propagating subarachnoid water and solutes toward the brain surface and tissue. The broader point is that the field must distinguish clearly between direct tracer-based measures of solute movement, indirect measures of water dynamics or anatomy, and claims about clearance of disease-relevant macromolecules.

VK: The previous focus of neurofluidics clearance has been on the brain surface, as most findings regarding the glymphatic mechanism involve the cortex. Deep structures, like most of the white matter and basal ganglia, are still a bit of a mystery when it comes to clearance. Spinal canal and their perinervous solute transport are also another dark spot due to lack of accurate spatiotemporal MRI or any other imaging methods for that matter at that depth.

“Deep structures like most of the white matter and basal ganglia are still a bit of a mystery when it comes to clearance”.

Also, the solute efflux routes are now clear; as the BBB is always there to prevent removal of water-soluble molecules in/out of the interstitium, especially the exit routes are not clear. The inter-astrocytic cleft opening has not been imaged, only it’s hydrodynamics traces have been detected. Cranial nerves are one efflux route and although active, they may not be the only route of waste efflux in humans according to some research findings from spinal efflux.

Recent work suggests that the upright spinal canal position with the relatively elevated (up to 20–30 cmH₂O) hydrostatic pressure in humans seems to facilitate the entering of CSF solutes also along spinal nerve roots into peripheral tissues¹¹. While upright high spatiotemporal resolution MRI for spinal nerve roots is not yet feasible, the recently developed HD ultrasound imaging could in principle be used to investigate neurofluidic efflux at least in the periphery.

Several questions are open: Deep brain structures, how does solute transport happen there?

How do individual pulsation mechanisms interact with moving proteins and other solutes?

How to directly measure human protein synthesis and removal; water can be imaged, but how about proteins? Spinal canal seems to be the CSF sink but we have no mechanistic imaging of CSF flow or pulsations in any species. How do the solutes enter the perivenous spaces and veins?

Furthermore, most of the methods are not usable in the clinic, yet. We need quantitative and predictive imaging biometrics that can yield more causative relations rather than correlative inferences regarding the neuropathology behind the patients’ symptoms. Recently developed methods could be optimized for such purposes easily given the appropriate means and resources. One way would be to somehow join technical aspects and fuse some of the spatiotemporal aspects of the imaging technologies. Direct imaging of endogenous brain-derived protein passage would be a nice thing to do at macroscopical level, too.

What (conceptual and/or technological) developments do you wish to see/look forward to seeing in human brain clearance?

MvO: In my opinion, it is essential to understand whether or how well CSF/ISF mobility can reflect the functioning of the CSF-mediated brain clearance system. The validity of this general assumption underlying many of the human MRI studies is, however, difficult to prove without a clear gold standard. Comparisons between intrathecal MRI and non-invasive MRI can be an important step in this, but are certainly not providing the full picture. This is especially important as non-invasive MRI techniques can measure the dynamics of the CSF-mediated brain clearance system at a relatively high temporal resolution and allow repeated measurements. It is therefore well suited for studying the driving forces as well as in intervention studies, for example, studying the influence of the different sleep stages on brain clearance.

As already mentioned before, I do think it is essential for the field to report more clearly what part of the brain clearance system is being probed by the employed techniques, as well as what is not probed.

LL: I would love to see a technology that can directly measure CSF flow (both its velocity and direction) in tiny structures like perivascular spaces. Since perivascular spaces are the channels in brain tissue, understanding fluid movement through these spaces would be particularly informative for understanding transport in the human brain. In addition, it would be exciting to have a technology that could noninvasively measure metabolite concentrations in brain tissue with high sensitivity. PET and MRS are appealing methods for tracking specific molecules, but higher sensitivity may be needed to understand the dynamics of clearance. Finally, we would really benefit from a tool that could label molecules inside the brain noninvasively and then track their movement. This would let us measure what is going out, and not only what is going into the system, which is important since exit routes can be quite different from entry routes.

GR: I would first like to see greater consistency in the use of terminology. The field would benefit from converging around common definitions of brain clearance. Without clearer definitions, there is a risk that very different biological processes and imaging readouts are discussed as if they measure the same phenomenon.

Technologically, I hope the field will move from semi-quantitative imaging measures toward more quantitative approaches. In MRI, this means greater use of quantitative T1 mapping so that absolute concentrations of exogeneous tracers and even endogenous proteins can be estimated, rather than relying mainly on semi-quantitative signal intensity changes, and preferentially at higher field strengths. Quantitative imaging would improve comparability across scanners, protocols, centers, and patient groups.

Ultimately, the goal is to develop and validate reliable non-invasive methods, but such methods need validation. An important limitation considering validation with exogeneous tracers concerns molecular size and biological specificity. Most imaging tracers are small molecules, whereas endogenous waste products, protein aggregates, and macromolecular complexes differ in size, charge, binding properties, and cellular handling. Development of clinically safe and more relevant tracers that better reflect clearance of the waste solutes they are intended to mirror, would be a big step forward.

Another development I look forward to is a broader clinical perspective on CSF pathways. Brain clearance research is often framed only as waste removal, but CSF may also be exploited as a carrier route for drug delivery to the brain. In a recent study, we used intrathecal contrast-enhanced MRI to demonstrate an alternative route for drug delivery and brain tumor visualization in glioma, illustrating how clearance research may also inform therapeutic delivery strategies.

Overall, I hope the field will become more quantitative, more mechanistically precise, and more clinically useful: not only describing fluid movement, but determining when altered clearance matters for disease, diagnosis, treatment selection, and drug delivery.

“I hope the field will become more quantitative, more mechanistically precise, and more clinically useful: not only describing fluid movement, but determining when altered clearance matters for disease, diagnosis, treatment selection, and drug delivery”.

VK: Joined multimodal verification of tracerless neurofluidic MRI scanning and wearable sensing metrics against golden standard tracer kinetic investigations will be a very important step for clinical usage for advanced brain and blood solute quantitation in different diseases. We need clinically feasible, fast and repeatable measures to quantify human brain solute clearance driving the neurofluidics. Moreover, a baseline normative data for both brain sensing and imaging data over a lifespan in healthy individuals as a joint international effort would be very highly appreciated as it will enable individual diagnostics and predictive health monitoring.

What (pharmacological or clinical) developments do you wish to see/look forward to seeing in human brain clearance?

MvO: It would be great to have a non-invasive and comfortable method to influence the brain clearance system that could be used for validation and comparison of human brain clearance methods. Sleep is currently the most reliable candidate, but sleep studies in e.g. the MRI scanner are very challenging and probably also not reflecting natural sleep. And of course, the holy grail would be to prove whether impaired brain clearance has a causative effect in neurodegenerative diseases. If this was true, this would also open-up new treatment as well as preventive avenues that are so desperately needed.

LL: One core clinical need is to determine whether clearance systems are a contributing mechanism in a narrow or broad set of diseases. A large number of studies have shown associations between MRI metrics associated with brain fluid systems, and several different neurological and psychiatric disorders. In which cases are these MRI metrics reflecting that clearance impairments caused the disorder, and when are they instead perhaps a readout of physiological dysfunction without being the initial cause? One strategy to answer this could be leveraging longitudinal imaging studies, where a single person is imaged over time, to understand which brain changes develop before clinical symptoms.

I would also love to see a better understanding of which sleep medications affect clearance, and how they do so. Different sleep medications act through very different molecular and circuit mechanisms, and it’s likely that they have very different consequences for brain clearance. This hasn’t yet been measured directly in humans and would be very valuable to understand, for the millions of people taking sleep medications, and for developing targeted pharmacological approaches to improving this process.

GR: Clinically, I would like to see brain clearance move from a research concept to an individual functional measure. In nephrology, serum creatinine and estimated glomerular filtration rate are used routinely to assess kidney function and guide clinical decisions. I think we need an analogous approach for the brain: a clinically applicable test that provides an individual measure of brain fluid clearance capacity. Our experiences with CSF tracer-to-blood clearance are promising in this regard.

Such a test could have several applications. One is dose adjustment for drugs administered intrathecally. Today, intrathecal drug dosing rarely accounts for individual differences in CSF clearance, although such differences may influence exposure, efficacy, and toxicity. Another potential application is risk stratification in neurodegenerative diseases. If impaired brain fluid clearance contributes to accumulation of amyloid-β, tau, inflammatory mediators, or other toxic solutes, then measuring clearance capacity could become relevant for assessing dementia risk, selecting patients for trials, or monitoring treatment effects.

I am also interested in whether MRI and CSF clearance measurements can be used to evaluate interventions that aim to improve lymphatic drainage of solutes from the craniospinal compartment. Lymphovenous anastomosis, a microsurgical technique used in lymphedema treatment, has recently been proposed as a way to enhance cervical lymphatic outflow and thereby support brain waste clearance in Alzheimer’s disease. Early reports are intriguing, and several clinical studies are ongoing or registered, but the evidence is still preliminary and the approach remains controversial. I think intrathecal enhanced MRI is needed to characterize which groups of neck lymph nodes that drain CSF and therefore could be targeted surgically. Robust controlled studies of such surgery are needed to determine safety, patient selection, biological mechanism, and clinical efficacy.

VK: Developing an easy-to-use monitoring along with an inexpensive and safe treatment cascade for preventing/treating brain proteinopathies prior to irreversible neurodegenerative stage develops. This could be done by increasing neurofluidic solute transport within and out of the CNS by a following repertoire of interacting tools and therapeutic approaches: 1. Use non-invasive wearable brain sensing technology to detect early signs of pulsatile neurofluidic dysfunction prior to cognitive decline 2. Increase neurofluidic brain efflux with recently developed, individually tailored treatments guided by brain-sensing biofeedback 3. In advanced dementia, localize solute transport deficiency using water flow sensitive techniques like CSF-STREAM, MREG or EPTI, and 4. treat it with focused ultrasound opening targeted to jammed perivascular spaces for restoration of cognitive and memory performance and glymphatic flow. The previous therapies could, in theory, also be joined together to further increase therapeutic efficacy.

This interview was conducted by Elisa Floriddia, Senior Editor, Nature Communications.