Clinical-MRI paradox: Do you stop for groceries at night? 


There are three ways of demonstrating disease activity in MS. First, pwMS can experience new neurological symptoms compatible with a relapse. Second, blood or CSF neurofilament levels can be elevated which indicates brain damage because of ongoing inflammation. Third, there can be new inflammatory lesions on a brain or spinal cord MRI. From a biological point of view, there is no hierarchy between these disease activity pillars. If disease activity is demonstrated in one of these three domains, it would be a valid reason to initiate or upscale treatment. However, some pwMS and clinicians put more weight on the first clinical pillar. 

Recently, I advised two pwMS who had radiological (i.e. two enhancing brain lesions/one new brain lesion + one cord lesion) but no clinical activity to upscale towards a second line treatment. Although I discussed the relevance of these lesions, both of them decided to stay on their first line and thus less effective treatment. Undoubtedly, several factors played a role in these decisions. Not having to upscale treatment could be a coping mechanism in people who have not entirely accepted the diagnosis. For others it is difficult to grasp the impact of asymptomatic lesions, especially if they are flourishing professionally and personally. Last but not least, I failed to sufficiently clarify the concept of silent brain lesions and the clincial-MRI paradox. 

Two attempts to do it better for the readers of this blog: 

First attempt: topographical model of MS. 

This model visualises the central nervous system as a pool divided into three basic anatomical regions with increasing amounts of functional reserve: 1. spinal cord and optic nerve lesions occupy the shallow end and carry a higher permanent disability burden, 2. brainstem and cerebellum comprises the middle, 3. cerebral hemispheres constitute the deep end of the pool as many brain regions are involved in information processing. 

The water represents neurologic reserve capacity. This means the ability to compensate for damage and keep regions of damage “submerged.” Reserve might fluctuate over time during fever or concurrent illness. The water’s surface depicts the clinical threshold: peaks that cross above the threshold cause clinical relapses; peaks that remain below the surface are seen as clinically silent lesions on MRI. A peak that crosses the clinical threshold may recede again beneath the water’s surface or remain above the clinical threshold, leaving residual deficits. Progression is depicted as the slowly declining water level, representing a gradual depletion of reserve capacity, and revealing clinical symptoms or peaks who had been previously under the water surface corresponding to silent lesions or partially recovered relapses. 

The model design incorporates 5 variable factors which reflect the heterogeneity in MS disease course: 

1. Localization of relapses and causative lesions: shallow vs. deep end of the pool

2. Relapse frequency: number of peaks above the pool surface

3. Severity: height of the peak

4. Recovery: the degree to which each peak recedes under the water surface

5. Progression rate: the rate at which water level declines

Central to this model is the observation that MS progression reflects prior relapse symptoms and unmasks previously clinically silent lesions. Most importantly, the model illustrates how, although a significant brain lesion burden may appear discordant with a favourable clinical picture early in the disease course, early disease activity or the amount of clinically silent disease seen on MRI are meaningful predictors of disability in the long term. 

This model has been beautifully illustrated by the following videos: 

Early secondary progressive disease:

Relapsing MS, highly active disease:

Primary progressive MS:

Relapsing MS, Benign course:

Second attempt: the traffic analogy

The meaning of these asymptomatic MRI lesions can also be illustrated by something that we are all familiar with: car traffic. Roads in Western Europe are subject to regular maintenance, with or without a couple of days notice. If there is one road block (read: silent MRI lesion) because of construction works on your daily commute, there are often one or two diversions you can follow and it will hardly affect the amount of time spent on commuting. On the contrary, when there are two road blocks, it might become more difficult to reach your target destination and your new trajectory might not allow you to pick up some last-minute groceries from the local store. Nonetheless, you will still reach your destination, albeit at a slower pace. With three, four and five road blocks, there are less opportunities for efficient rerouting and you’d prefer working from home. 

Photo by Ben Wicks on Unsplash


Although the first attempt is more refined and granular it might not always be clinic-proof and requires a considerable amount of explaining before concepts such as silent MRI lesions sip through. The second attempt definitely does not encompass the complexity of the disease course in MS but has the beauty of simplicity. 

Any other ideas on how to explain the significance of asymptomatic brain lesions? Please share! 


Disclaimer: Please note that the opinions expressed here are those of dr. Ide Smets and do not necessarily reflect the position of the Barts and The London School of Medicine and Dentistry nor Barts Health NHS Trust.

Neurol Neuroimmunol Neuroinflamm 2016 Sep 7;3(5):e279.doi: 10.1212/NXI.0000000000000279. eCollection 2016 Oct.

The topographical model of multiple sclerosis: A dynamic visualization of disease course

Stephen C Krieger, Karin Cook, Scott De Nino, Madhuri Fletcher

PMID: 27648465, PMCID: PMC5015541, DOI: 10.1212/NXI.0000000000000279


Relapses and progression contribute to multiple sclerosis (MS) disease course, but neither the relationship between them nor the spectrum of clinical heterogeneity has been fully characterized. A hypothesis-driven, biologically informed model could build on the clinical phenotypes to encompass the dynamic admixture of factors underlying MS disease course. In this medical hypothesis, we put forth a dynamic model of MS disease course that incorporates localization and other drivers of disability to propose a clinical manifestation framework that visualizes MS in a clinically individualized way. The topographical model encapsulates 5 factors (localization of relapses and causative lesions; relapse frequency, severity, and recovery; and progression rate), visualized utilizing dynamic 3-dimensional renderings. The central hypothesis is that, like symptom recrudescence in Uhthoff phenomenon and pseudoexacerbations, progression clinically recapitulates prior relapse symptoms and unmasks previously silent lesions, incrementally revealing underlying lesion topography. The model uses real-time simulation software to depict disease course archetypes and illuminate several well-described but poorly reconciled phenomena including the clinical/MRI paradox and prognostic significance of lesion location and burden on disease outcomes. Utilization of this model could allow for earlier and more clinically precise identification of progressive MS and predictive implications can be empirically tested.

About the author

Ide Smets


  • I struggle to understand why any patient would disregard advice such as you were imparting, but I guess you see all types!

  • This leaking pool model of MS was introduced to me by Aaron Boster on his YouTube channel relatively early in my MS disease course, about 4 years ago. And it meant alot.
    Considering how many patients (and quite a few neurologists, to be fair) seem to think that – or act as – MS is isolated to clinically significant relapses that correspond to new lesions, this model should be known by everyone.
    In my personal opinion, it should be given as a handout to every new MS patient upon receiving their diagnosis, so that they can make good choices when considering DMTs and what risks they are willing to take, even though this may be uncomfortable and add to the initial shock of the diagnosis. The shock will wane, the MS will not.
    Probably the most important post of this year.
    Thank you, Ide!

  • Nice video, especially the link with MS drugs and why it’s important to initiate tx early.

  • Thank you so much this is a really helpful post for understanding what’s going on. I have heard variations of the traffic model used to explain neuroplasticity – if the motorway is permanently shut and everyone starts using a B road instead , then over time investment in the B road might bring it up to an A road or even a new motorway. Analogies are so helpful to me

  • Excellent post Ide thanks. I also struggle to see why pwMS might not accept urgency and importance of treatment using the pool analogy. Really like the pool depths for reserve capacity and parts of the brain, easy to visualise and explains a lot.
    I wonder if the falling water level doesn’t appear important unless reveals previous relapse damage? Maybe for the model there needs to be more ‘mild’ damage that becomes apparent as water level drops?
    The particular image shown of benign MS would never be diagnosed unless picked up on MRI by chance!

    Less sure about the road blocks, although for me, this analogy very helpful for explaining MS fatigue.

  • Yes, excellent and very helpful. I’m wondering, however, if there is ever a real-world scenario where you might feel, say, that a 50+-year-old patient should finally come off meds, rather than continue. For example, someone with definitive RRMS (i.e., not misdiagnosed) for about 20 years with a mild course having been on Copaxone, then Rebif, then Tecfidera, with low lesion load and no evidence of disease activity (NEDA) on MRI for quite some time (as opposed to someone with asymptomatic new lesions seen on MRI), and no clear evidence of progression clinically, and is now suddenly JCV mild+ with low Lymphs#? Is there ever a scenario where the risks of going on an anti CD20 med like Ocrevus, for example, is actually not worth it in a risk/benefit analysis for such a patient…particularly in the age of COVID? Or does the theoretical advice that everyone should be on the most efficacious med possible apply to everyone and outweigh all of those factors and the increased risks? Using blood or CSF neurofilament levels in practice seems not to be very mainstream yet, at least at one major MS center here in the US? Thanks!

    • No, I don’t think benefits of anti-cd20 always outweigh risks. Especially in progressive pwMS who are wheelchair bound/severe neurogenic bladder/swallowing difficulties I don’t think one or two new lesions would outweigh the risks. Then I would most certainly not initiate it as a maintenance therapy, but in discussion with the patient I could consider using it as an IRT.

  • You advised your patients to switch to a second line therapy based on radiological activity. So you are a proponent of escalation therapy. Do you not believe in the “hit hard hit early” treatment regimen proposed by prof G? Is there benefit in using the most effective DMDs from the outset?

    • There is definitely a benefit, but I was not involved in the care of these patients before.

  • Very useful post Ide, thanks. I was lass enamoured of the leaky pool analogy before, but I think I’ve seen it more clearly diagrammatically than just seeing it explained in a video by Dr B.It rushed past me a bit but I was able to appreciate it more this more static way.It now makes more sense. Having the four pairs of “pool” for each sort of MS is more useful to me.

  • I wonder how your analysis fits with NICE guidance?
    I am aware that some centres are prepared to game patients experience to obtain the “right” result for treatment.
    I have not found the Statistics for any of the DMTs miraculously good and those who take the best of class do not really accept the consequences as has been shown by the concern over COVID. The aftermath of the consequences of Lemtrada has left some prescribers more cautious.
    The model has so many variables and uncertainty, it has value in helping patients to think about what may happen but the variability in untreated outcome seems to be poorly characterised. There are people with MS in their 80s who have kept themselves active and reasonably well without ever having received DMTs. Would they really have been better off if they had been treated and suffering the side effects of the very powerful drugs?
    I have looked for but never found an explanation of the trajectory of untreated MS though there must be considerable history available. When we are able to predict the untreated fate of an individual, we will be better able to make an informed choice. At the moment the pwMS is making the choice based on hoping for the best or fearing the worst.

    • Good outcomes in untreated pwMS are highly biased towards EDSS, but dismiss other variables such as employment, cognition, etc.

  • for the record, I use both analogies — leaky pool, and jammed traffic — in my practice. I hope that people find them both helpful in understanding this disease.
    thank you, dr. smets, for this lovely and thoughtful post.

  • Hello Dr. Smets- I suppose I have been grappling with one form or another of this question “recognizing fact and making best choice” for some time (with MS and everything else). I think the topographical model is splendid and the threshold water level works well. But I don’t think there’ll be any model that fixes the problem of understanding, acceptance and action.
    After fining the best model and having clinicians explain exactly what is needed, there are other factors that will often prevail. I would suggest the bigger two of those is denial, and yes, intelligence (and of course social factors and upbringing). Not being bright has its advantages in the world of believing you are safe and seemingly moving forward. Keep in mind I am referring to one half of us. At what point of intelligence (generally) does understanding and motivation for action “kick out”? 85 IQ, 100 IQ?
    “Back in the day” when I was diagnosed (1992 ish), I read up as much as possible at the library (U of Il). I watched worse case scenario video from MS Soc.(man dies motionless in bed), went to a support group with a couple fairly handicapped people you could see were no longer doing well. When the first treatment came out in 1994, I jumped on it. Granted, things have become so much more complicated, but the point is, that for me, one of the important acceptance factors was diving in the pool, so to speak.
    I like the topographical representation, but believe it’ll be a bit complicated for many to grasp. Others, still, will not care, as they are bullet-proof.

    • We don’t know much about coping mechanisms in pwMS but they are definitely a dealbreaker when it comes to DMT management and willingness to understand the topographical model.

  • I really wish that everyone newly diagnosed were shown both the pool and road analogies for the damage ms does silently as well as loudly via the distinct symptoms of relapses. Even if they were to be guaranteed this information, here in the UK there is still the likelihood of seeing a neuro who applies the escalation model and quite possibly describes the more effective treatments in scary terms or as cost prohibitive to the local health care service, as did the first two I saw.

    I was fortunate to receive treatment with Alem, having pushed for it back in 2016. Now it’s happens that my diagnosis and initial MRI was done privately and that MRI is the most comprehensive one I’ve received. It showed numerous lesions, even including my thoracic spine and a ‘black hole’ spinal lesion. However, I’ve not been provided with a single spinal scan since and when the NEDA report comes back from my annual brain scans, I feel I’m having to keep fingers crossed regards what’s happening with my spine.

    Without the sense of an urgent or proactive culture from the various neurology depts and clinicians, is it any wonder so many PwMS have internalised a more cautious or insouciant attitude/outlook? If only the culture were similar to that shown with cancer, or say Sepsis. Another good example is the recent awareness raising that’s been done regards the menopause and HRT and look at how dynamic self advocates women have become there! I’m sure if education were provided using the pool/roads analogies and the clinical culture was more dynamic then PwMS would respond and act in their own best interests.

    Me = diagnosed Nov 2015 and referred to the NHS .
    First apt with a neuro 6 months later, before which I had another relapse and was lucky that the private neuro was ok to see me again!
    Pushed for Alem at the first NHS apt and so was referred to a London centre.
    First apt at the London centre four months later.
    It was a year post diagnosis before I was treated as someone with ‘severe, rapidly evolving RRMS’.
    That was sufficiently long enough for me to go from being able to walk, albeit slowly, with risk of foot drop, for a few miles to not since being able to walk more than 1.5 miles and two miles on a good day…

    Education, urgency, dynamic and quality of ongoing monitoring.



Recent Posts

Recent Comments