Offering a non-invasive window into the brain, magnetic resonance imaging (MRI) helps visualise key pathological mechanisms. Conventional MRI, routinely used in clinical practice, allows for the identification of multiple sclerosis (MS) lesions – their number, location, and activity [1]. In contrast, quantitative MRI, primarily used in research settings, can better characterise which tissue components within the central nervous system (CNS) are damaged and to what extent, as well as their spatial distribution and temporal evolution [1]. Positron emission tomography (PET) further complements MRI by enabling the in vivo assessment of molecular and cellular processes involved in MS pathology [2].
Advancements in MRI techniques
Cutting-edge imaging techniques are reshaping our understanding of MS pathophysiology [3]. Key MRI markers for diagnosis, monitoring, and treatment evaluation include central vein signs (CVS), paramagnetic rim lesions (PRLs), and slowly expanding lesions (SELs) [3]. Lesions with CVS, as identified on susceptibility-based imaging, reflect perivenous inflammatory demyelination – typical of MS lesions [4].
As Professor Jiwon Oh, from the University of Toronto, said during our recent webinar, “CVS is not only a sensitive tool that facilitates the diagnosis of MS, but it really helps prevent misdiagnosis”. A high proportion of CVS lesions has been observed in most cases of radiologically isolated syndrome (RIS), suggesting that CVS may help distinguish individuals with RIS who are at risk of developing MS from those who are not [4]. PRLs are signs of chronic active inflammation, can be detected in the early stages of the disease and are highly specific for MS, while SELs are associated with greater disease activity and may predict a more severe disease course [3]. Both CVS and PRLs will be included in the 2024 McDonald Diagnostic Criteria.
Professor Cristina Granziera – from the Department of Biomedical Engineering of the University of Basel, department of Neurology at the University Hospital Basel and Research Center for Clinical Neuroimmunology and Neuroscience Basel (RC2NB) – investigates the pathogenesis of MS, by using different MRI techniques to better detect changes in brain tissue.
MRI magnitude images and phase images are very relevant to detect CVS. Prof. Granziera tells us, “CVS will be integrated in the 2024 McDonald criteria. Therefore, my expectation is that within the next two years, these techniques will enter the clinical practice. The process works as follows: researchers develop these techniques and demonstrate their importance in showing biomarkers relevant to the disease. Once validated, these biomarkers are incorporated into diagnostic and prognostic criteria, paving the way for the techniques to be adopted in clinical practice.”
MRI: Measuring myelin and axonal integrity
“Myelin sheath has many components that we can explore and image, including proteins, water, lipids,” Prof. Granziera continues, “This is why we can use several approaches to study the integrity of myelin sheath, such as Myelin Water Imaging, which provides information about the water content within the myelin sheath. We can also use Magnetisation Transfer or Quantitative Susceptibility Mapping, which exploits the signal derived from the macromolecules in the myelin sheath and the directionality of myelin in the tissue. Additionally, we can measure axonal integrity with the neurite density index (NDI) or tractography. Tractography allows us to map the main tracts and provide insights into the functional relevance of structural changes.”
Both Meylin Water Fraction and NDI were applied to study myelin-related and axon-related properties in white matter lesions, cortical lesions, and normal appearing tissue in 91 individuals with MS (62 relapsing-remitting, 29 progressive) and 72 healthy controls [5]. The study – led by Prof. Granziera – found a reduction of myelin and axonal integrity in both white matter and cortical lesions. Diffuse reduction of myelin water fraction and NDT index was observed in normal-appearing white and grey matter of individuals with MS. Furthermore, participants with progressive MS showed a more extensive reduction in myelin water fraction and NDI in normal-appearing cortex compared to individuals with relapsing-remitting MS [5]. Individuals with clinical deficits showed a correlation between the NDI in white matter lesions and disability [5].
Using diffusion tractography, Prof. Schoonheim and his team observed a widespread reduction in fiber density – indicating microstructural axonal damage – and in fiber cross-section – indicating macrostructural tract atrophy – in individuals with MS compared to healthy controls [6]. The most severe damage was observed in individuals with secondary progressive MS [6].
Prof. Granziera specifies, “These techniques are not used in clinical practice because the scans take too long and currently, they have no direct application in diagnosis, prognosis, or treatment monitoring. Instead, we use them to deepen our understanding of MS pathology and test new therapies. There is a strong need for treatments that can repair damage – known as remyelination. To support this, we need highly specific techniques that can provide reliable quantitative measures. I believe that within the next five years, we will begin to see the benefits of this research, with these techniques gradually making their way into clinical practice.”
PET imaging: Predicting disability progression
Stopping progression remains an unmet need in MS. PET allows for the direct in vivo assessment of key processes involved in neurodegeneration in individuals with MS [7].
Prof. Laura Airas, from the University of Turku, tells us, “We know that widespread, compartmentalised inflammation within the CNS contributes to worsening of symptoms in individuals with progression independent of relapse activity (PIRA). This inflammation is mainly driven by activated glia cells – both microglia and astrocytes. However, glial activation cannot be reliably measured using conventional MRI. While certain MRI-based methods can detect microglial activation, particularly in chronic active lesions, PET is necessary to more specifically assess the underlying pathology. We can consider PET as a form of molecular imaging, as it allows us to measure the expression of a given molecule in the brain, to quantify it, and localise it to specific brain areas.”
A recent study analysed brain tissue from donors with MS with slow versus rapid progression [8]. Presence of lesions with a broad rim of activated myeloid cells was linked to rapid progression. These lesions exhibited cellular and transcriptional signatures of innate immune activation. To validate these findings in vivo, researchers ran an independent study with the mitochondrial 18 kDa translocator protein (TSPO) PET imaging. The study confirmed the link between these lesions and disease progression [8].
Prof. Airas says, “TSPO is expressed on the outer mitochondria membrane of microglia cells, and its expression increases when these cells become activated. In MS, TSPO availability in the brain increases as the disease advances. In progressive MS, we can see significantly higher TSPO binding across many brain areas compared to relapsing-remitting patients or healthy controls. Notably, we can detect an increased TSPO binding in the normal appearing white matter. These areas look normal with MRI, but with PET we can see that they are not normal. An increase of TSPO binding in these regions is associated with a greater likelihood of clinical progression in the following years.”
In vivo TSPO-PET studies demonstrate that the activation of innate immune cells in normal appearing white matter predicts later disability progression – independent of clinical relapses – in the following 4 years [9]. Recently, a third-generation TSPO-PET radioligand – 11C-ER176 – was used to quantify the density of innate immune activity in the brain [10]. The study showed that the thalamus of individuals with MS shows higher density of innate immune activity compared to controls. Furthermore, higher density of innate immune activity in the thalamus was linked to progressive MS and higher disability [10].
New tools to personalise MS therapies
Prof. Airas continues, “TSPO-PET can be a very valuable tool for identifying patients who may benefit from therapies targeting microglia activation. It enables more personalised treatment strategies and is particularly useful in clinical trials aimed at modulating microglial function. We strongly believe that if we can reverse this proinflammatory activated microglia phenotype back to a more homeostatic state, we could slow or prevent further disease progression. And we could measure whether the drug really targets microglia in early trials in vivo. Importantly, it has been demonstrated that TSPO-PET imaging can be reliably used in longitudinal settings to measure change in microglial activation. TSPO-PET is currently the only available method that can capture the widespread microglial activation in vivo. We also have to say that TSPO ligands are not 100% specific for microglia – they also bind to activated macrophages and a subset of astrocytes. However, since MS is an inflammatory disease, we know in which brain areas glial activity is prevalent and detrimental. Therefore, we can focus our imaging study findings on these key brain regions. Despite this lack of complete specificity, increased TSPO binding has been shown to reliably predict future disease progression, making it a useful biomarker for identifying high-risk patients and testing treatments.”
Another way to depict specific patterns of brain morphology in individuals with MS is by representing the brain as a network with nodes and edges. Network studies based on structural similarity – e.g. gray matter (GM) morphology – have detected more segregated and less integrated structural networks at various MS stages. For example, subcortical networks have been linked to specific MS-related symptoms and clinical progression, in some cases more effectively than conventional volumetric analysis [11]. However, most of these studies have examined GM networks at group level, neglecting inter-individual variability thus limiting their application to individual predictions of clinical courses.
The so-called single-subject GM network approach extends the group-level analysis by providing a reliable analytical framework of the individual GM morphology. In a recent longitudinal multi-center study within the MAGNIMS consortium, researchers from the University Medical Center Mainz and their European colleagues introduced a novel methodological approach for extracting one GM network from one patient to predict disease progression in MS [12]. This study showed that individual GM networks are sensitive to an underlying progressive disease course and largely independent of relapse activity. The observed network changes towards a less efficient network preceded disability worsening and outperformed conventional MRI predictors [12].
These observations suggest that single-subject GM networks may capture crucial information explaining variance beyond conventional volumetric measures, with promising potential for personalised prediction of clinical courses.
***
Written by Stefania de Vito
Special thanks to Professor Laura Airas (University of Turku, Finland) and to Professor Cristina Granziera (University Hospital of Basel, Switzerland) for their insights.
References
[1] Granziera C et al. Brain 2021; 144(5): 1296-1311.
[2] Sucksdorff M et al. Brain 2020; 143(11): 3318-3330.
[3] Rocca MA, Preziosa P, and Filippi M Curr. Opin. Neurol. 2025; 10-1097.
[4] Sati P et al Nat. Rev. Neurol. 2016; 12(12): 714-722.
[5] Rahmanzadeh R et al. Brain 2021; 144(6): 1684-1696.
[6] Koubiyr I et al Brain Comm. 2024; (6)1: fcae018.
[7] Bodini B et al Nat. Rev. Neurol. 2021; 17(11): 663-675.
[8] Klotz L et al Nat. Med. 2025; 1-11.
[9] Sucksdorff M et al Brain 2020; 143(11): 3318-3330.
[10] Zeydan B et al. Brain Comm. 2025; 7(3): fcaf141.
[11] Fleischer V. et al. Ann. Neurol. 2022; 91: 192-202.
[12] Fleischer V. et al. Brain 2024; 147: 135-146.