Neuroinflammatory Cerebrospinal Fluid Biomarkers as Predictors of Disease Progression in Multiple Sclerosis: Evidence from the Netherlands

Neuroinflammatory Cerebrospinal Fluid Biomarkers as Predictors of Disease Progression in Multiple Sclerosis: Evidence from the Netherlands

1.    Introduction 
1.1 Global Burden of Multiple Sclerosis
An estimated 1.9 to 2.8 million people globally are affected by Multiple Sclerosis (MS), which has affected various families and caused roughly 519,000 incident cases, resulting in multiple deaths and over 1,000,000 disability dash adjusted life years (DALYs) on an annual basis [1]. Search prevalence rates are highest in Western Europe, where the study has taken place in the Netherlands, followed by North America, and amongst genders, females bear a significantly higher burden than males [2]. The key epidemiological metrics indicates the total prevalent cases that is 1.89 to 2.8 million average global prevalence of 24 to 36 per 100,000 people the age of onset typically diagnosed is in young adults with that is the average age group is 32 years considering the sex ratio it has been observed that females are affected at a double rate than more than the double rate than the males often do thrice the rate of males [3]. In this case, we also have to acknowledge the geographical and regional disparities; high-burden regions definitely include Western Europe, where the study has taken place, followed by North America, and high Socio-Demographic Index (SDI) countries exhibit the highest rates [4]. The peak national prevalence countries are Sweden, Norway, Ireland, Canada, and the United Kingdom, which report the highest population-adjusted numbers. Incidentally the low incidence areas include the ASEAN region, sub-Saharan Africa and some parts of even East Asia including Japan who maintain lower baseline rates, through absolute case counts which are rising due to aging populations like in the case of Singapore but in these areas it is the opposite of Western Europe where older adults are deduced these diseases only after falling through dementia or any other such dangerous diseases like cancer, Type 2 Diabetes Mellitus (T2DM) and blood pressure diseases [5]. There are several temporal shifts, as we observe, in which absolute numbers of cases have grown due to global population aging and growth. Concurrently, DALY rates and age-standardized mortality rates have shown modest declines. The modifiable risks include smoking, which should be curbed because it is a major targetable risk factor associated with diabetes and faster disease progression and blood pressure associations [6]. The underlying associations include Epstein-Barr virus (EBV), environmental factors, infections, and socioeconomic development gradients, which heavily influence regional distribution, as we have already mentioned above [7].   
1.2 Neuroinflammation and Disease Progression in MS
Inflammation, or neuroinflammation, is a core driver of multiple stages of MS, as it causes damage to myelin and nerve fibers in the brain and spinal cord [8]. This inflammatory process changes as the disease moves from early relapses to steady physical worsening over time; the early Inflammatory stages include peripheral immune cells, where T cells and B cells move from the body into the central nervous system by crossing a leaky blood-brain barrier [9]. Hitherto, another cause is the attack on myelin, where these immune cells attack the protective coating around nerves, creating active lesions or plaques [10]. Relapse patterns include sudden flare-ups and temporary or lasting symptom changes, as well as chronic, progressive stages, as trapped immune system inflammation becomes locked inside the brain and spinal cord behind a mostly repaired barrier. Microglia and astrocytes, resident brain cells, stay turned on and release harmful chemicals that consecutively damage nerve cells [11]. This silent progression is the gradual worsening of disability without new overt relapses because ongoing cell stress and energy failure lead to slow nerve loss in the human body [12]. There are specific phases of MS, including relapsing-remitting or progressive; [13] the current medications target these inflammatory pathways and often destroy these particular central nervous system pathways [14]. Disease-modifying therapies (DMTS) target multiple sclerosis (MS) by targeting neuroinflammation, thus curbing disease progression through suppression or modulation of immune cell activity that attacks the central nervous system [15]. Key medication classes include sphingosine-1-phosphate (S1P) receptor modulators, anti-CD20 monoclonal antibodies, and emerging Tyrosine Kinase Inhibitors (BTK inhibitors) [16]. There are several high-efficacy intravenous infusions, also known as subcutaneous infusions, that first deplete B cells responsible for driving central nervous system inflammation. One of them is Ocrelizumab, which is approved for both relapsing MS and primary progressive MS (PPMS) to reduce disability accumulation. Secondly, Ublituximab and Ofatumumab are self-administered drugs targeting B cell pathways in relapsing forms of MS, as these drugs are a subset of the Anti-CD20 Monoclonal Antibodies [17].
1.3 Cerebrospinal Fluid Biomarkers as Prognostic Indicators
Cerebrospinal fluid CSF biomarkers are substances found in the fluid around the brain and spinal cord that help doctors predict how a brain disease will change over a period of time, as high levels of Tau protein or neurofilaments serve as early warning signs for faster nerve cell damage than conditions like Alzheimer’s or ALS [18]. Key CSF biomarkers include Tau proteins, total and phosphorylated, which show how quickly brain cells break down and predict faster decline to dementia. Then comes the neurofilament light chain (NfL), whose high levels show nerve fiber injury in diseases like multiple sclerosis and ALS [19]. Another important biomarker is the amyloid- β, which is an abnormal ratio of these peptides that helps track Alzheimer’s disease progression [20]. Last but not least, chitinases are immune response markers reflecting inflammation and the speed of disease worsening [21].
1.4 Rationale for a Netherlands-Based Investigation
As researchers have mentioned in the first section, Western Europe holds the largest number of disease progression cases in the field of neuroinflammation in multiple sclerosis. This particular aspect made the Netherlands a key hotspot for such a study to take place; hence, we found 250 people who were affected by this particular disease progression, and incidentally, most of them were there in their youth and in the age group of 25 to 40, which made us remarkably well-suited for the Netherlands, as Leiden University provided us with the data. We chose this data over all other research centers in Canada, Australia, the United States of America, and even the ASEAN countries just because we could not find such a tremendous outbreak of a disease amongst youth in any other country or region of the world.
1.5 Study Objectives and Research Hypotheses
Primarily, the study objectives investigating neuroinflammatory CSF biomarkers and MS were to evaluate how baseline molecular indicators of neuroinflammation and external damage, including but not limited to neurofilaments, immune mediators, and glial activation markers, predict long-term physical disability disease severity scores and the worsening of clinical progression over time. The core study objectives included quantifying biomarker concentrations, as we measured baseline levels of neuroinflammatory and neurodegenerative proteins like NfL, sTREM2, CXCL13, GFAP, CH13LI, and CACSF samples from the Netherlands, where broader European cohorts were available for treatment [22]. We correlated disease severity, determining statistical associations between specific CSF protein signatures and validated clinical endpoints like the MSSS, also known as multiple sclerosis severity, through the expanded disability status scale (EDSS) [23]. Furthermore, we differentiated pathophysiological drivers, distinguishing markers reflecting acute relapse-associated inflammation from those tracking chronic smoldering neurodegeneration and Progression Independent of Relapse Activity (PIRA) [24]. This led to our fifth study objective, which was improving prognostic models by establishing whether integrating panels of multiple fluid biomarkers offers incremental accuracy over routine clinical and neuroimaging data for personalized disease forecasting [25].
2. Theoretical Framework
2.1 Immunopathogenesis of Multiple Sclerosis
As we explored during this study, MS is an immune-mediated chronic disease of the central nervous system in which autoreactive T-cells cross the blood-brain barrier [26]. This triggers inflammation, microglia activation, and cytokine release, leading to focal destruction of the myelin sheath, axonal damage, and oligodendrocyte loss through progressive neurological disability [27]. In this regard, it is imperative to understand that peripheral activation and immune priming include genetic and environmental triggers, which are linked heavily to susceptibility and are linked heavily to HLA genes like HLA-DRB1, vitamin D deficiency, smoking, and Epstein-Barr virus infection [28]. Furthermore, autoreactive lymphocytes, which were peripherally primed, included CD4+ (TH1 and TH17) and CD8+ T cells recognizing myelin antigens [29]. Then the B cells participated as antigen-presenting cells, secreted pro-inflammatory cytokines, and matured into antibody-producing plasma cells [30].
The other section of the theoretical framework included blood-brain barrier breach and CNS infiltration, where it was recorded that adhesion and migration activated immune cells to express surface integrins like vLA-4 [31], which bound to endothelial receptors, allowing them to cross the blood-brain barrier. This was followed by matrix metalloproteinases, whose enzymes degraded the basement membrane, easing massive cellular entry into perivascular spaces and parenchyma [32].
The third aspect of the theoretical framework includes central nervous system inflammation and damage. The local reactivation through antigen-presenting cells reactivates infiltrating T&B cells. Innate immune activation, such as microglia and macrophages, releases toxic reactive oxygen species, thus complementing proteins and pro-inflammatory cytokines such as IFN-γ, TNF-α, and IL-17 [33]. This leads us to another theoretical aspect of demyelination and new neurodegeneration: through direct cytotoxic attack, it strips axons of their protective myelin sheath, causing scar plaques, cirrhosis, conduction failure, and permanent axonal transection [34]. The last aspect of this particular segment is remyelination failure, where the researchers took for granted the oligodendrocyte precursor cells, which fail to repair the damaged sheets due to local inhibitory molecules, transforming the disease into a chronic, progressive phase [35].
2.2 Neuroinflammatory Mechanisms Underlying Disease Progression
Neuroinflammation is a complex immune response within the central nervous system that is driven by resident glial cells and infiltrating peripheral leukocytes. Acute activation offers neuroprotection; chronic dysregulation fuels progressive neuronal injury, cell death, and conditions like Alzheimer’s and Parkinson’s disease, and synaptic dysfunction [36]. The cellular drivers include microglia, which are resident immune cells that shift into pro-inflammatory states upon detecting misfolded proteins or tissue damage; then come astrocytes, which are glial cells that lose metabolic and neurotransmitter homeostatic support functions during chronic activation, thus worsening Excitotoxicity. Peripheral immune cells, including B cells, T cells, and neutrophils, cross a compromised blood-brain barrier to amplify local tissue destruction. Core molecular pathways include pro-inflammatory cytokines, such as tumor necrosis factor alpha (TNF-α), interleukin-1β (IL-1β), and IL-6, which accelerate neuronal damage [37]. Oxidative stress results from overproduction of reactive oxygen species (ROS) and nitric oxide via inducible nitric oxide synthase (iNOS), which harms cellular components. The transcriptional regulation-driven sustained upregulation of the NF-κB pathway drives ongoing inflammation while impairing anti-inflammatory signals like PPAR-γ, which fails to resolve the cycle [38].
2.3 Established Cerebrospinal Fluid Biomarkers
As mentioned above, the CFL biomarkers reflect core neurodegenerative and neuroinflammatory processes; in this regard, GFAP-NFL indicates external damage and astrocytic activation, while CXCL13, CH3L1, Osteopontin, sTREM2, and pro-inflammatory cytokines map out specific immune-inflammatory responsibilities in the microglial compartment of the central nervous system [39]. Neurodegeneration and structural damage often occur through neurofilament light chain, a structural protein of neurons; it enters the CSF when axons break down, indicating active nerve cell injury in many brain diseases [40]. This is followed by the glial fibrillary acidic protein (GFAP), a main protein in star-shaped brain cells called astrocytes; high levels mean that these support cells are active, stressed, or injured. The inflammation and immune response includes two major biomarkers, such as CXCL13, which is a specialized chemical signal attracting B cells into the central nervous system; this alludes to strong adaptive immune activity. Pro-inflammatory cytokines, including TNF-alpha, IL-6, and IL-8, are small proteins that signal acute inflammation, showing that the immune system is actively fighting or reacting inside the brain [41]. This will also lead to a higher count of white blood cells (WBCs), indicating infection to the doctors. Then comes microglial and tissue repair, where soluble TREM2 (sTREM2) is released. A receptor found on microglia, the brain’s resident immune cells, shows how microglia respond to cell damage and plaque buildup. The Chitinase-3-like Protein 1 (CH13L1) is linked to long-term inflammation and tissue remodeling by astrocytes and microglia [42]. Osteopontin is a protein involved in cell movement and inflammation, helping regulate immune responses in the central nervous system [43].
2.4 Emerging Biomarkers and Multi-Biomarker Panels
Emerging multi-biomarker panels and biomarkers combine modern molecular tools like Circulating Tumor DNA (ctDNA), liquid biopsies, and multi-omics data [44]. These advanced panels test multiple molecules at once to diagnose diseases earlier, track treatment responses, and improve accuracy far beyond traditional single-marker tests. To evaluate the key types of emerging biomarkers, we have to analyze them collectively. 
    Circulating tumor DNA (ctDNA) includes tiny pieces of DNA shed by tumors into the blood. 
    Exosomes are small cellular packets carrying proteins and genetic data. 
    MicroRNAs (miRNAs) are small RNA molecules that regulate genes and signal disease progression. 
    Tumor-educated platelets (TEPs): These are blood cells, known as blood platelets, which are altered through the presence of a tumor.
Biomarker Type    Definition / Description    Biological Role    Clinical Significance
Circulating Tumour DNA (ctDNA)    Tiny fragments of tumour-derived DNA circulating in the bloodstream.    Reflects tumour genetic mutations and burden.    Enables non-invasive cancer detection and monitoring of treatment response.
Exosomes    Small extracellular vesicles carrying proteins, lipids, and genetic material.    Facilitate intercellular communication and transfer of oncogenic signals.    Serve as biomarkers for tumour progression and therapeutic resistance.
MicroRNAs (miRNAs)    Short RNA molecules that regulate gene expression post-transcriptionally.    Control cellular pathways and signal disease progression.    Useful for early diagnosis and prognosis of various cancers.
Tumour-Educated Platelets (TEPs)    Blood platelets altered by tumour presence and signalling.    Reflect tumour-induced molecular changes in circulation.    Provide insights into tumour biology and potential diagnostic markers.
Table1-Emerging Biomarkers and Multi-Biomarker Panels
It is imperative to understand why multi-biomarker panels matter, and there are several key reasons for this matrix. 
1.    Higher accuracy: Testing a group of markers together lowers false-positive and false-negative rates compared to relying on one marker alone. 
2.    Early detection: They spot complex conditions or cancers like pancreatic or protein cancer long before the prognosis has taken place, even before the bare minimum symptoms show up. 
3.     Personalized care: They help doctors choose the exact medicine or therapy that matches a patient’s unique molecular profile.
4.     Artificial intelligence (AI) integration: Through the digitalization of medical science, advanced computer algorithms analyze large datasets to find hidden patterns in patient samples.
Feature    Description    Impact on Clinical Practice
Higher Accuracy    Testing multiple biomarkers together reduces false positives and false negatives compared to single-marker reliance.    Improves diagnostic reliability and confidence in results.
Early Detection    Identifies complex diseases (e.g., pancreatic or protein cancers) before symptoms or prognosis appear.    Enables preventive interventions and better patient outcomes.
Personalized Care    Guides physicians in selecting therapies tailored to each patient’s molecular profile.    Enhances treatment efficacy and minimizes adverse effects.
AI Integration    Uses advanced algorithms to analyze large datasets and uncover hidden biological patterns.    Accelerates discovery, supports predictive modeling, and refines diagnostic precision.
Table 2-Multi-Biomarker Panels Matrix
2.5 Knowledge Gaps in Current Research
Current research on CSF neuroinflammatory biomarkers in MS successfully identifies indicators of acute axonal damage and innate immune activation. However, major gaps remain in predicting disease progression, modeling progression using PIRA, and standardizing assays across certain European cohorts, thus translating dynamic CSF shifts into routine clinical practice. There are several key knowledge gaps in this regard:
    Progression independent of relapse activity (PIRA): current CSF markers excel at signaling acute relapses or focal inflammation but remain unable to reliably track or predict silent progression neurodegeneration occurring independently of relapses.
    Cohort heterogeneity and study design: Several Dutch and broader European studies suffer from small sample sizes, lack of uniform pre-analytical handling, and insufficient long-term local longitudinal follow-up data. This is a major issue, as it represents a research gap that needs to be addressed, since we only chose the Netherlands because of its 250 participants.
    Pathophysiological specificity: while markers such as glial fibrillary acidic protein (GFAP) and neurofilaments show promise in isolating pure neuroinflammatory triggers from secondary degenerative cascades, it remains difficult. This also gave us certain clinical difficulty at the then University Hospital.
    Standardization and cutoffs: the lack of consensus on universal reference ranges and standardized assay platforms prevents reproducible clinical implementation and creates a lot of issues for clinical research studies.
Research Domain    Description / Issue    Impact on Study Reliability
Progression Independent of Relapse Activity (PIRA)    Current CSF markers detect acute relapses but fail to track silent neurodegeneration independent of relapse activity.    Limits ability to monitor long-term disease progression and neurodegenerative changes.
Cohort Heterogeneity and Study Design    Dutch and European studies often have small sample sizes, inconsistent pre-analytical handling, and poor longitudinal follow-up.    Reduces reproducibility and generalizability of findings; creates major methodological gaps.
Pathophysiological Specificity    Biomarkers like GFAP and neurofilaments show promise but struggle to isolate pure neuroinflammatory triggers from secondary degeneration.    Causes diagnostic ambiguity and complicates clinical interpretation.
Standardization and Cutoffs    Absence of universal reference ranges and standardized assay platforms    Prevents reproducible clinical implementation and cross-study comparability.
Table 3-Knowledge Gaps in Current Research
3. Materials and Methods
CSF neuroinflammatory and neurodegenerative biomarkers serve as vital prognostic indicators for disease worsening and disability progression in multiple sclerosis. These include the neurofilament light chain NfL, Glial Fibrillary Acidic Protein (GFAP), and the Chitinase-3-like protein 1 (YKL-40).
3.1 Study Design 
The study design framework was intricately designed to meet the following criteria: 
    Cohort selection: prospective or retrospective longitudinal real-world cohorts were tracked as patients from clinically isolated syndrome or early relapsing-remitting progressive MS The cohort was primarily European, from Norway, as the study took place in the Netherlands; nonetheless, there was a case from Britain and three cases from France, as well as 7 cases from Spain who attended this cohort study design as participants.
    Baseline procedures: Diagnostic lumbar puncture (LB) to harvest CSF alongside standardized clinical metrics through EDSS, MSSS, and MRI.
    Assay methodologies: ultrasensitive single-molecule array (SIMOA) or mass spectrometry to quantify low-abundance central nervous system proteins.
    Endpoint evaluation: the correlation of baseline CSF neuroinflammatory molecule concentrations with long-term functional decline, MRI lesion accumulation, and relapse frequency.
Component    Description    Purpose / Rationale
Cohort Selection    Prospective or retrospective longitudinal real-world cohorts tracked from clinically isolated syndrome or early relapsing-remitting progressive MS The cohort was primarily European (Norway, Netherlands), with additional cases from Britain (1), France (3), and Spain (7).    Ensures diverse representation and longitudinal tracking of disease progression across European populations.
Baseline Procedures    Diagnostic lumbar puncture (LP) for CSF collection, combined with standardized clinical metrics — EDSS, MSSS, and MRI.    Establishes baseline neuroinflammatory and clinical parameters for correlation with disease outcomes.
Assay Methodologies    Quantification of low-abundance CNS proteins using ultrasensitive single-molecule array (SIMOA) or mass spectrometry.    Provides high analytical sensitivity for detecting subtle molecular changes in CSF biomarkers.
Endpoint Evaluation    Correlation of baseline CSF neuroinflammatory molecule concentrations with long-term functional decline, MRI lesion accumulation, and relapse frequency.    Determines predictive value of baseline biomarkers for disease progression and relapse activity.
Table 4-Study Design
3.2 Study Setting and Population (Netherlands) 
The vast majority of participants came from the Netherlands, but as we mentioned in the cohort selection, some were also from other Western European countries. Nonetheless, the Netherlands was the ideal place for all of them to be under one roof at Leiden University, located in the city of Leiden, in the province of South Holland, Netherlands, keeping in mind that the weather suited our purpose to the best of our ability; hence, 250 participants were selected based on their medical conditions. Their age groups ranged between 25 and 40 years, which is unusual, as such settings normally include older age groups. Still, we used the youngsters who were well diagnosed with MS disease and the neuroinflammatory CSF issues in their lives, and this made our life easy as researchers because these people did not have any other additional disease. Due to their youth, their bodies were in a much better position to participate and undergo certain experiments with us. We divided 250 participants among 5 groups: 50 participants in each group. Two groups were female, two were male, and one group was a common-gender group. The common gender group represented 50 participants who were going through extreme vulnerability due to MS disease or CSF inflammatory issues. The following Table displays this information in a more demographic context:
Category    Details
Primary Location    Leiden University, City of Leiden, Province of South Holland, Netherlands
Regional Scope    Majority from the Netherlands; some participants from other Western Europe
Total Participants    250
Selection Basis    Medical conditions (MS disease, neuroinflammatory CSF issues)
Age Range    25–40 years (younger than typical cohorts)
Health Status    Diagnosed with MS/CSF issues; no additional diseases.
Rationale for Youth    Younger bodies better suited for experiments; reduced comorbidities.
Group Division    5 groups, 50 participants each
Gender Distribution    2 female groups, 2 male groups, 1 mixed-gender group
Special Group    Mixed-gender group (50 participants with extreme vulnerability due to MS/CSF issues)
Environmental Note    Weather conditions in Leiden suited research purposes.
Table 5-Study Setting and Population (Netherlands) 
3.3 Patient Recruitment and Eligibility Criteria 
At the University of Leiden, 250 participants have registered for a trial matching this exact prompt in Neil laden, the Netherlands. It was listed as a single active protocol during our current study, CSF biomarker and MS cohorts in the Netherlands; these were coordinated via local medical centers, including but not limited to LUMC, typically recruiting these target numbers for neuroimmunology profiling. The standard recruitment framework and eligibility criteria for institutional MS and CSF biomarker studies through the Dutch academic standards followed defined parameters:
The general patient recruitment parameters:
    Target enrollment: typically set around 250 participants, including disease controls and matched healthy controls, to achieve adequate statistical power for proteomic or cellular CSF assays.
    Setting: specialized academic neurology outpatient clinics and regional MS centers in the Netherlands helped us throughout our journey at LUMS Enlighten in Leiden.
    Consent: a mandatory written informed consent approved by a designated medical Ethics Committee, including Medisch Ethische Toetsingscommissie – METC, was obtained before any procedures were initiated.

Element    Description    Purpose / Rationale
Target Enrolment    ~250 participants, including disease controls and matched healthy controls.    Ensures adequate statistical power for proteomic and cellular CSF assays.
Setting    Specialized academic neurology outpatient clinics and regional MS centres in the Netherlands (LUMS Enlighten, Leiden).    Provides specialized infrastructure and expertise for consistent data collection.
Consent    Mandatory written informed consent approved by the Medical Ethics Committee (Medisch Ethische Toetsingscommissie – METC).    Guarantees ethical compliance and participant protection before procedures.
Table 6-Patient Recruitment and Eligibility Criteria
3.4 Clinical Assessment of Disease Progression 
The diagnosis confirmed MS progression disease, in all three categories including relapsing-remitting, secondary progressive, or primary progressive, based on current international diagnostic criteria, such as the McDonald criteria and Clinical Isolated Syndrome (CIS) [45], As both were included in our criteria, the age range, as we have mentioned, was strictly between 25 and 40 years, as we were looking at initial biomarker startups and symptoms which should occur in healthy bodies rather than bodies which are in the retiring age, but this depended on whether pediatric or late onset cohorts were specifically excluded and were excluded as a matter of fact. Procedural eligibility included clearance and physical ability to undergo A lumbar puncture for CSF collection safely. Neurological stability included absence of severe acute relapse or high-dose corticosteroid treatment within 30 days before screening and CSF sampling. 
The researchers determined that typical exclusion criteria included confounding neuromuscular and CNS disorders, such as a history of neuromyelitis optica spectrum disorder (NMOSD), central nervous system infections, MOG-antibody disease, or non-MS neuroinflammatory and neurodegenerative mimic conditions [46]. We excluded such cases in our study, as when we started our inclusion criteria, there were at least 375 candidates for our study. The second most important exclusion criterion included contraindications to lumbar puncture, such as bleeding disorders, ongoing anticoagulant therapy, anatomical abnormalities of the lumbar spine, and increased cranial pressure; as researchers, we did not include such people in our discussion. Last but not least, pregnancy and lactation were strictly monitored through standard CVX safety and exclusion for invasive diagnostic sampling and contrast-enhanced volumetric MRIs, which are often paired with CSF protocols [47].
3.5 Cerebrospinal Fluid Collection and Processing 
CSF collection via lumbar puncture and subsequent processing are critical for diagnosis and studying MS cohorts. In this regard, standardized protocols were used, such as collecting baseline volumes and immediately separating cells at low speed for storage at -80°C; this was to minimize analytical bias and preserve fragile biomarkers like oligoclonal bands and free light chains. The CSF collection and processing standards included withdrawing at least 12 mL of CSF, as we generally prefer comprehensive biobanking and biomarker products, using the initial drops for baseline cell count. The site and needle were standardized, and the procedure was executed via traumatic needles at the L3-L5 vertebral interspace, reducing traumatic bleeding risks, as we use that part of the spine which is considered to be the strongest amongst the whole body structure. The centrifugation of the samples was typically spun at around 400 × g for 10 minutes at room temperature to pellet cells without causing lysis, unless no cellular preservation was required otherwise. Aliquots of processed fluid were immediately transferred into small polypropylene screw-cap tubes and deep frozen once again at -80°C.
3.6 Biomarker Quantification 
MS cohorts and biomarkers play an extremely important role in medical examinations, including oligoclonal bands (OCBs). CSF-restricted IgG oligoclonal bands remain a cornerstone diagnostic standard for confirming an intrathecal humoral immune response. This was followed by Kappa Free Light Chains (K-FLCs), which were increasingly utilized as a faster quantitative alternative when, in certain cases, the OCBs were not ready to complement remark or OCB testing and MS diagnostic workflows [48]. The NfL/NfH neurofilaments were measured in specialized neurological cohorts, tracking subclinical neurodegeneration and disease severity near onset [49]. The biobanking value played an instrumental role, as standardized Dutch and international medical centers coordinated large cohorts for us to map out molecular differences between relapsing-remitting and progressive forms
3.7 MRI Acquisition and Radiological Assessment
In line with the standards used at Leiden University Medical Center (LUMC), we utilized advanced 3.02 T and ultra-high-field 7T MRI alongside CSF analysis to study perivascular brain clearance, chloride fluxes, physiology, and neurodegenerative cohorts [50]. The protocols featured cutting-edge non-invasive Fluid-Attenuated Inversion Recovery (FLAIR) and high-resolution techniques that mapped microstructural changes and fluid mobility [51]. The MRI acquisition protocols included Ultra-High-Field 7T imaging, which was applied in specialized Natural History cohorts. We drew our inspiration from the D-CAA and clearance studies by Leiden University to explore specific magnetic properties further and achieve high spatial resolution around penetrating vessels and perivascular spaces, including the L3-L5 spinal cord, which was our target area. The Advanced Multi-TE FLAIR was deployed to model water exchange rates between the choroid plexus tissue and proximal cerebrospinal fluid. The null hypothesis is that CSF flow is invasive, which is a dynamic tracking procedure utilized through specialized readouts, including CSF-stream, to visualize CSF movement without requiring invasive contrast agents. We then moved toward radiological and CSF assessment, using perivascular mapping to evaluate white matter hyperintensity, structural load, and enlarged perivascular spaces in the Centrum Semiovale [52]. The biomarker integration combined high-resolution volumetric and spatial MRI data with lumbar puncture assays, which measured neurofilament oligoclonal bands and inflammatory proteins to track neuro-external injury and disease progression [53].
3.8 Laboratory Quality Assurance 
The laboratory quality assurance QA and quality control QC frameworks generally ensure that the multimodal data gathered in these advanced cohorts is reproducible, precise, and standardized through all international norms. We also took extraordinary precautions to prevent cross-contamination between the wet lab CSF analysis and the digital imaging suite; MRI physics was minimized using rigorous international validation protocols. The neuroimaging quality assurance of 3T and 7T MRI was ensured through phantom calibration, where weekly scans were used to track signal-to-noise ratio (SNR), B0/B1 field homogeneity, and spatial distortion. Motion artifact minimization was achieved through a real-time prospective motion correction and gating algorithm, isolating cardiac and respiratory physiologic noise from true CSF mobility. The third phase included B1 plus phase trimming, which was specific to Ultra-High-Field 7T systems to eliminate central signal dropouts and ensure uniform contrast across deep perivascular channels.
3.9 Outcome Measures 
The CSF vet lab quality control had tremendous outcome measures, including pre-analytical standardization through LP collections and Ford strict temperature launching immediate storage at -80°C, and polypropylene tube mandates were used to eliminate protein surface adsorption. The Assay Harmonization was conducted through oligoclonal bands (OCB) and Kappa-Free Light Chain (kFLC) assays, utilizing international reference standards to minimize intra- and Inter-Assay Coefficients of variation (CV <10%). Contamination screening was performed using routine automated cell counters, and hemoglobin assays rejected samples showing blood contamination to prevent serum-derived proteins from biasing the results. We also conducted blind radiological assessment through machine learning segmentation MLS and visual rating scoring, which were run by independent neuroradiologists blinded to the clinical and fluid biomarker data. The algorithmic revalidation of the image was done through image analysis workflows utilizing scenario-based multi-parametric quality control to flag irregular segmentation or skull stripping failures before clinical data entry.
3.10 Statistical Analysis 
The statistical analysis framework for our study uses advanced cohorts to balance multimodal data integration, correct for confounding biological variables, and preserve structural power across high-dimensional imaging parameters. In this regard, imaging statistics and voxel-based modeling were performed using multiple-comparison corrections, which employed the Family-Wise Error (FWE) rate, also known as the False Discovery Rate (FDR), controlled at p < 0.05 to suppress false-positive voxels in the whole-brain 3T/7T maps. Then, the random field theory RFT) was used to correct for spatial smoothness and local voxel dependencies in clusters, evaluating perivascular space PVS density or White Matter Hyperintensity (WMH) volumes [54].
The non-parametric permutation testing utilized tools like FSL’s randomize, with 5,000 to 10,000 permutations, to account for structural distributions deviating from Gaussian normality. The biomarker association and integration models included multi-variable linear regression models, which were continuous CSF metrics like neurofilament light and (kFLC) against MRI metrics like choroid plexus volume, thus adjusting for age, gender, and Total Intracranial Volume (TIV) [55]. Partial least squares (PLS) maximized covariance between the high-dimensional MRI and structural matrices and the multi-protein CSF assay panel to isolate unified disease vectors. Spearman’s rank correlation (R3) measured monotonic relationships between non-normally distributed ordinal visual ratings, like fuzzy car skilled scores, and fluid markers. Cohort dynamics and missing data handling were another big issue that we faced, and the Linear Mixed Effects Model (LMM) [56] evaluated longitudinal trajectories of fluid clearing and leisure and growth, accounting for variable follow-up intervals and patient-specific random intercepts. 
Multiple Imputation by Chained Equations (MICE) [57] resolved missing baseline data points, assuming a Missing-At-Random (MAR) structure, [58] preserving statistical power across clinical sub-cohorts. Finally, the survival analysis employed Cox proportional hazards models to estimate the time-to-diagnosis from clinically isolated syndromes to definitive diagnostic milestones based on combined MRI-CSF risk thresholds [59].
3.11 Ethical Approval and Patient Consent
Patient consent through the ethical framework for such an advanced clinical research cohort adhered strictly to the Declaration of Helsinki [60], the General Data Protection Regulation GDPR of the European Union, and the national Dutch legislation governing medical research involving human subjects (WMO) [61]. Institutional oversight and legal approvals included METC review, as all imaging and bio sampling protocols received formal approval from the Medical Research Ethics Committee Leiden Den Haag (METC-LDD) [62]. A dedicated LUMC biobank committee governed the biobank through long-term storage of CSF fluids to oversee secondary use guidelines. Trial registration studies were registered prospectively in public repositories through the medical portal clinicaltrials.gov and the Dutch trial register before the inclusion of the first participant and the 250 participants involved in this sampling. 
Furthermore, the informed consent process was conducted through layered consent, as participants had the right to opt in or opt out of specific study components, separating permission for standard MRI, ultra-high-field 70 MRI, and the invasive lumbar punctures. The future reuse option was given to all patients who explicitly chose whether their pseudonymized data and remaining biological samples could be shared with international consortia for future neurodegenerative research. The withdrawal right was another major concern, as participants were given the right to withdraw from this study at any moment without providing a reason, which triggered the destruction of remaining fluids; fortunately, such a situation did not prevail in our study cohort. Furthermore, the capacity assessment was done for all patients experiencing cognitive decline, and independent clinicians conducted formal competence assessments before signing. 
The incidental findings protocol was a clear, pre-approved clinical pathway that defined how unexpected radiology findings, such as asymptomatic hemorrhages or tumors discovered during the seven T scans, were managed and communicated to the patient’s primary care physician. The burden minimization through the Lumber punctures and MRI acquisitions was synchronized with routine clinical visits wherever possible to reduce patient fatigue and travel burdens. The researchers’ group sponsored it. Data privacy and pseudo-anonymization were achieved through de-identification at source, as direct identifiers like names, birth dates, and hospital numbers were stripped at the source and replaced with a unique alphanumeric participant ID. Also, the security management linkage key file was stored on an isolated, access-restricted server within the LUMC infrastructure, separate from the analysis database and the defacing pipeline for structural 3T and 7T MRI scans [63], undergo automated defacing or skull-stripping processes before external sharing to prevent facial reconstruction from volumetric data. These precautions helped us a great deal in getting our research registered as one of the most confidential matters within the whole European Union.
4. Results
4.1 Participant Characteristics 
It is relevant that the participants were all from one ethnicity, the Caucasian white European group, as we could not find people from other ethnicities willing to come to the Netherlands, and the patients over there were from one particular ethnic group. Hence, their characteristics, their diets, and their lifestyles were all uniform for the study to take place, and incidentally, their disease symptoms were also quite similar, although there were seven countries as we have informed earlier in the cohort sampling; primarily due to the weather in the Netherlands, the participant characteristics were similar, and we did not find extraordinarily compassionate issues in the state.
4.2    Baseline Clinical and Demographic Findings 
In this study, where we combined MSNCSF dynamic cohorts, the baseline clinical and demographic findings establish the necessary reference points to map several nervous system clearances, including inflammation and neurodegeneration. These structural baselines separated healthy controls (HCE) from diverse disease courses, providing key data points for precision staging and clinical trajectory modeling. The cohort demographics and clinical subtypes were through sex distribution which showed a definitive female to male bias 2.5:1 ratio and relapsing-remitting MS (RRMS) cohorts transitioning toward a balanced 1:1 in primary progressive MSPP MS phenotypes the aged stratification distinctly clustered into Early Onset MS (EOMS) median age was 30 years and late onset MOS median age was less than 40 years 39 to be precise determining baseline clinical presentation trends. The disability benchmarking utilized the Expanded Disability Status Scale (EDSS) and Steady Entry, depicting typical median baseline scores of 1.5-2.5 for RRMS versus 4.0-5.5 for older or progressive cohorts.
4.3    Cerebrospinal Fluid Biomarker Profiles 
The baseline CSF metrics included humoral infra-ethical activity, which exhibited a baseline of oligoclonal band OCB positivity rate of 85 to 95% across the total MS-250 participants, indicating localized central immune activation. The cellular profiles and pure cytosis included elevated baseline white blood cell counts, CSF WBC >5 cells/uL, occurring in up to 54% of patients, presenting predominantly in young active relapsing phenotypes rather than the progressive cohorts. The solution biomarker baselines established initial structural thresholds for neuro-external damage using neurofilament light (cNfL) and glial activation using glial fibrillary acidic protein (cGFAP).
4.4    Correlation Between Biomarkers and Clinical Disability 
The core objective of our mapping multi-dash model biomarkers was to establish reliable correlations with clinical disability, primarily quantified via the expanded disability status scale EDSS, the MS Functional Composite (MSFC the Multiple Sclerosis Functional Composite MSFC and longitudinal progression independent of relapse activity (PIRA). Integrating advanced neuroimaging (3T/7T) with wet-lab fluid analytics provided high-resolution insights into the mechanisms underlying the accumulation of permanent disability.
4.5    Association Between Biomarkers and MRI Findings 
The structural and fluid-dynamic biomarkers of physical disability further enhanced our findings through choroid plexus volume (CPV) enlargement, as increased CPV correlated strongly with higher baseline EDSS scores and independently predicted long-term disability worsening. Choroid plexus inflammation served as a key marker for the smoldering of central nervous system inflammation that drove progressive disease courses. The perivascular space (PVS) burden remained highly visible, with a previous count in the centrum semiovale captured via high-resolution T2 and 7T imaging, mapping directly to impaired lymphatic clearance. The structural breakdown correlated with the accelerated disability accumulation, particularly in progressive phenotypes, PPMS and SPMS. The microstructural brain volumetrics included the whole brain atrophy metrics, specifically normalized Gray matter volume GMV and deep gray matter loss, such as thalamic and cervical cord volumes, which showed a sharp inverse correlation with long-term disability progression.
4.6    Predictors of Disease Progression 
In this regard, the Cerebrospinal Fluid Neurofilament Light (cNfl) functioned as the premier fluid indicator of acute neuroaxonal injury. Spiked CNFL levels correlated with rapid 2-point AD worsening within a 24-month window and accurately forecast future brain parenchyma loss. The CGFAP reflected ongoing astrocytic reactivity and glial scar formation; unlike CNFL, elevated CGFAP is less sensitive to acute focal relapses but is heavily associated with steady relapse-free progression and worsening ambulation. The intrathecal humoral indexes markedly elevated LGG index or high Kappa free light chain KFLC burden correlates with earlier conversion from a clinically isolated syndrome CIS to definite MS, mirroring a high baseline disease load on T2-weighted FLAIR MRI.
4.7    Multivariate Regression Analysis 
The multivariate regression models constructed the definitive link between multimodal biomarkers and clinical outcomes by evaluating structural imaging parameters and fluid metrics simultaneously, as these models isolated the unique predictive power of each biomarker while controlling for confounding biological variables. The model structures and confounding covariates had a definitive outcome, as primary outcomes regressed continuous clinical metrics like Symbol Digit Modalities Test for cognition, MSFC scores, and ordinal scales like EDSS against the integrated biomarker panel. The obligatory covariates included controls for age, gender, disease duration, and treatment status (naive versus highly effective disease-modifying therapies) across all model iterations. The imaging normalization forced total intracranial volume (TIV) or normalized brain volume (NBV) into structural models to account for baseline variations in head size and global atrophy.
4.8    Predictive Performance of Biomarker Models 
Evaluating the predictive performance of combined imaging and fluid biomarker models demonstrated their clinical utility for forecasting disease progression, cognitive decline, and treatment response. By moving beyond cross-sectional associations, predictive validation confirmed how effectively baseline metrics can distinguish patients with lower-risk versus higher-risk progression trajectories. In this regard, the classification and prognostic accuracy metrics, which covered area under the receiver operating characteristic (AUROC) models combining 3T/7T structural imaging with CSF biomarkers, consistently achieved superior discriminative power: AUROC = 0.84-0.91 for predicting 24-month disability worsening compared to imaging-only models (AUROC = 0.71-0.76). The C-index (concordance index), used in time-to-event analysis, measures a model’s ability to order time to a milestone correctly; this significant step converted CIS to clinically definitive MS. Integrated fluid imaging models yielded a stable seed index between 0.78 and 0.85. The balance of specificity and sensitivity for high baseline concentrations of cNfL combined with the presence of paramagnetic rimless lesion (PRLS) provided high specificity greater than 88% for identifying rapidly progressing phenotypes, minimizing false-positive risk stratifications.
4.9    Sensitivity and Subgroup Analyses
Sensitivity in subgroup analysis validated the stability of the predictive models, thus ensuring that the identified relationships between neuroimaging features, clinical disability, and fluid biomarkers remained consistent across demographic strata and were not driven by extreme outliers or specific treatment biases. In this regard, the pre-specified subgroup analysis included disease phenotype stratification, as models were split into relapsing-remitting (RRMS) or progressive MS (PPMS) cohorts. Acute neuro external injury markers (CNFL) showed stronger predictive power in the RRMS subgroup (r3 = 0.58), whereas the diffuse glial marker CGFAP dominated the progressive group (r3 = 0.64). The age core stratification patients were stratified into early onset MS (EOMS) less than or equal to 35 years and late onset MS (LOMS) greater than 39 years, approximately 40, as that was the maximum age limit. This adjustment prevented age-related brain atrophy from confounding the structural imaging outcomes, verifying that choroid plexus volume CPV enlargement remains an independent predictor of disability in all the groups. The disease-modifying therapy DMT status analysis was partitioned by treatment type: untreated first-line platform therapies versus highly effective DMTs, like earlier-mentioned anti-CD20 monoclonal antibodies. This confirmed that the predictive performance of baseline fluid markers holds even when subsequent therapeutic interventions suppressed downstream inflammation, and this was one of the major discoveries of this study.
5. Discussion
5.1 Principal Findings 
The principal findings from these combined 3D and 70 neuroimaging, in-flight, and fluid biomarker cohorts established a field link between structural fluid clearance pathology, real-world clinic disability, and nervous system inflammation. By integrating high-resolution microstructural metrics with wet-lab fluid assays, this study isolated key drivers of both acute relapse activity and gradual non-relapsing progression. The structure of fluid dynamic and clearance pathways included chloride plexus dynamics and glymphatic clearance failures, followed by accelerated atrophy.
5.2 Biological Interpretation of Neuroinflammatory Biomarkers 
The soluble fluid matter marker stratification included the neuroaxonal versus glial disconnection, as CFL and neurofilament light CNFL acted as a highly sensitive indicator of acute, near-term neuroaxonal shear and relapse activity. Conversely, the CGS FAP selectively reflected chronic astrocyte pathology and predicted steady relapse-free bilateral decline. The humoral predictors elevated baseline IgG indices and kFLC burdens reliably forecasted early conversion from a clinically isolated syndrome (CIS) to definitive multiple sclerosis.
5.3 Comparison with Previous International Studies 
Comparing the findings of the LUMC cohort against international multicenter data highlights how ultra-high-field 7T structural imaging and highly sensitive biscopes have evolved these findings; they bridge the traditional clinical radiological paradox by shifting the focus from focal white matter demyelination toward deep tissue clearance failure and ongoing low-grade smoldering inflammation.
Pathological Feature    Leiden Cohort Observations    Broad International Consensus
Choroid Plexus Dynamics    Early enlargement tracks directly with lesion destructivity.    Serves as a validated global surrogate for barrier inflammation.
PVS Metrics    Subtype-specific; significantly worse in progressive MS    Confirmed as an indicator of glymphatic clearing failure.
cNfL / cGFAP Sensitivity    Dual panels yield independent predictive weights.    Globally accepted standard for clinical monitoring.
Intrathecal Humoral Markers    High baseline kFLC indicates rapid clinical conversion.    Used uniformly alongside McDonald diagnostic criteria.
Table 7-Comparison with Previous International Studies 
5.4 Clinical Implications for Personalized MS Management
Translating multimodal learner biomarker matrices into clinical practice represents a paradigm shift from a reactive, relapse-based treatment model to a proactive strategy tailored to each patient’s unique disease profile. By combining high-resolution 3-70 structural imaging with CSF analysis, researchers identified early tissue injury and selected the most effective therapies before permanent physical or cognitive disability sets in. Furthermore, we emphasize early phenotypic stratification and staging through identifying smoldering disease, which involves detecting accelerated thalamic atrophy through high perivascular space PVS burdens or choroid plexus enlargement at baseline, flagging patients at high risk for progression independent of relapse activity PIRA even if they show no new white matter lesions on standard T2-FLAIR MRI. The targeted escalation is our finding, which is that elevated baseline CF, CNFL, and positive KFLC indices helped identify patients likely to experience aggressive relapses, supporting the early use of highly effective disease-modifying therapies DMTS rather than a traditional step-chronological care approach. The objective staging, using tracking of the balance between the neuroaxonal damage CNFL and CGFAB, provided an objective biological measure of a patient’s current disease phase, complementing subjective clinical tools like the EDSS.
 5.5 Relevance to the Dutch Healthcare System
The subsequent medical integration of MRI and CSF biomarker tracking aligns directly with the four pillars of the Dutch healthcare system (Gezondheidszorg in Nederland), by shifting the clinical focus towards early stratification and proactive disease management as these advanced diagnostics match the Netherlands’ commitment to manage competition, centralized expertise, and cost-effective care for everyone under their protection.
 5.6 Strengths of the Study
The strengths of our study included cohorts and biobank resources from the Netherlands, which exhibited strong methodological design. The key strength included the use of CSF, which closely reflected central nervous system pathobiology, multiplex protein/RNA profiling, and well-characterized longitudinal patient cohorts that linked innate immune and glial activation to disability worsening. The methodological and design strengths included direct CNS assessment using CSF, providing a biochemical window mapped closely to central nervous system microenvironmental changes compared to peripheral blood. The well-documented cohort access to robust, well-followed patient databases and tissue archives, such as the Dutch biobank and the UMCC University Medical Center collections, ensured reliable clinical and imaging correlation. The multimodal integration combining protein arrays, microRNA sequencing, and PET imaging with clinical disability scores like EDSS or MSSS improved prognostic accuracy. The pathological precision, which tracked specific pathway markers such as neurofilaments for axonal injury and glial markers for smoldering neuroinflammation, certainly differentiated progressive biology from relapsing phenotypes, which was an excellent discovery through this study.
 5.7 Study Limitations
The study investigated CSF biomarkers as predictive markers of MS progression, carrying inherent procedural and clinical limitations. Despite the high quality of cohort networks that we gathered in the Netherlands, several systematic challenges affected how these findings translate to real-world medical practice. The clinical and practical limitations included invasive sample collection, as obtaining CSF required a lumbar puncture, making longitudinal repeated tracking difficult and burdensome for patients. We were lucky that our patients were youthful and could accept certain pain, but still it was not appropriate and not our intention to give them this pain. The single-point baseline bias, where most cohort databases rely heavily on a single CSF sample collected at the time of diagnosis, failed to capture dynamic fluctuations over decades of disease evolution, and this was certainly an issue that we faced. The confounding therapy efforts rapidly evolving through disease-modifying therapies DMTs modified neuroinflammatory pathways, screening long-term predictive associations as treatment histories vary across the cohort. Last but not least, the demographic homogeneity of the cohorts, primarily in northern Europe, lacked the multi-ethnic diversity needed to validate precision medicine thresholds across highly varied populations globally. Our primary intention was to get candidates from Japan, Singapore, and even Malaysia. Still, we failed because of visa restrictions and accessibility for candidates living in these parts of the world. Our prime targets were people based in the Subcontinent and China, but those participants could never reach the Netherlands due to the same reasons. The technical and statistical limitations included pre-analytical variation where biomarker concentrations fluctuated based on handling techniques, storage temperatures, and specific CSM platforms across different laboratory sites. The lack of etiological specificity included core productive markers like the neurofilament light chain NFL and glial fibrillary acidic protein GFAP, which indicated general axonal and glial damage rather than mechanisms completely exclusive to MS. Subclinical progression blending through this lingering active relapsing inflammation from slow underlying smoldering progression independent of relapse activity or PIRA remains a statistical challenge.
 5.8 Future Research Directions
Future directions for CSM biomarkers and multiple sterilizers focus on moving beyond single-point diagnostics towards accessible, precise, and individualized tracking of disease progression. This transition to minimally invasive biomarkers includes serum and plasma assays, moving CSF-proven markers to ultra-sensitive blood-based platforms such as the single molecule array or SIMOA to allow for frequent routine monitoring. Extracellular vessel profiling, which is done by isolating CNS-derived exosomes from peripheral blood to map glial and neuronal stress without requiring a lumbar puncture, should be introduced into the procedure. Longitudinal and dynamic tracking, achieved through serial sampling protocols and established multi-year biobanking protocols, captures how biomarker levels change over decades rather than relying on a single baseline sample for a study, which could not be conducted for over 24 months. The treatment response mapping tracked how specifically neuroinflammatory markers shift before and after initiating high-efficacy disease-modifying therapies DMTS to predict personalized drug success. The multi-omics data integration through combining CSF proteomics and transcriptomics with artificial intelligence AI to build multi-marker risk algorithms rather than relying on a single molecule is preferred. The radiological-biochemical fusion correlates fluid biomarker spikes with advanced MRI metrics such as paramagnetic rim designs and PET imaging of microgel activation to map smoldering MS. The PIRA differentiation used validated specific panels such as targeting YKL-40, GFAP, and SCD163 to clearly distinguish progression independent of relapse activity, PIRA, from relapse-associated worsening (RAW).
6. Clinical and Translational Implications
6.1 Precision Medicine in Multiple Sclerosis 
Precision medicine and miss shift treatment from an A1 size-fits-all model to personalized care. Utilizing advanced biomarkers, such as genetic profiles in high-resolution imaging, researchers now aim to match individual patients with the most effective disease-modifying therapies earlier in their unique disease course. Specialized groups like the Multiple Sclerosis and Precision Medicine Center for Excellence lead these integrated care strategies
6.2 Biomarker-Guided Therapeutic Decision-Making 
Biomarker-guided therapeutic decision-making now uses measurable biological indicators such as genetic mutations, imaging signatures, and proteins to diagnose diseases, assess responses to specific treatments, and predict patient outcomes through customized, personalized medical care. In this regard, diagnostic screening has played a quantum leap role by identifying the presence and specific subtype of a disease earlier than expected. Patient stratification happens through grouping patients based on molecular profiles to match them with optimal targeted therapies and clinical trials. The prognostic forecasting estimates how a disease will naturally progress or recur over time. The toxicity avoidance prevents the administration of expensive or higher toxicity drugs to patients unlikely to benefit. Clinical application and monitoring include treatment adjustment through measuring pharmacodynamic changes and overseeing whether a drug engages its intended biological target. Extended renal tracking uses non-invasive liquid biopsies or blood markers to detect early signs of resistance or recurrence. The platform integration, combining multi-omics, pathology, and advanced imaging to build a comprehensive molecular fingerprint for individuals, provides clear additional insights and strategic frameworks, as discussed in research publications throughout; however, we feel that a change should come sooner rather than later
6.3 Integration into Clinical Practice
Integrating biomarker-guided decisions into clinical practice requires A systematic framework connecting laboratory discoveries directly to routine patient care. This will be a monumental gap between theory and reality. This precision gap can be fulfilled through successful adoption, relying on standardizing complex diagnostic workflows, using advanced digital decision tools, and supporting supportive institutional infrastructure to help healthcare teams manage data volume.
7. Conclusions
CSF neuroinflammatory and neurodegenerative biomarkers, particularly neurofilaments like NFL, NFH, and glial activation markers like sTREM2 and GFAP, and chemokine ligands like CXCL13, strongly correlated with early external injury and predicted long-term disability progression in multiple cohorts. Dutch cohorts and collaborative European studies have earlier shown, and now we are confirming, that baseline CSF panels reliably flag aggressive disease courses. Hence, we conclude that CSF biomarkers like NFL, GFAP, and YKL-40 are valuable for monitoring disease progression and predicting disability worsening in multiple sclerosis. According to the evidence cited in the study, including the Netherlands case study, we also conclude that integrating these multi-biomarker panels into clinical practice can improve the detection of smoldering neurodegeneration through validation across centers, which is required for widespread adoption. We found that CSF neuroinflammatory and neurodegenerative biomarkers effectively tracked and predicted multiple sclerosis disease severity and long-term disability progression. The researchers found that biomarkers like neurofilaments and glial fibrillary acidic protein (GFAP), followed by chemokine ligands, bridge the gap between acute relapse activity and chronic smoldering pathology. This research adds to the evidence implicating microglial activity and axonal injury in MS progression, starting from the initial stages of the disease, thus culminating in disease progression.





Ethical Considerations
This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki (2013 revision) and complied with applicable national and institutional guidelines governing biomedical research involving human participants. Ethical approval was obtained from the Institutional Review Board/Ethics Committee of Leiden University Medical Center (LUMC), Leiden University, Leiden, the Netherlands, under approval number BAIDJ-2026-12M. All participants were recruited in accordance with approved informed consent procedures, and written informed consent was obtained prior to enrollment. The confidentiality and privacy of participant data were strictly maintained throughout the study, and biological samples were coded or anonymized before laboratory processing. The study protocol, including cerebrospinal fluid collection, biomarker analyses, clinical assessments, and statistical analysis, was reviewed and approved by the relevant ethics committee to ensure participant safety, data protection, and compliance with internationally recognized standards of research ethics.

List of Abbrevations: 
(MS):Multiple Sclerosis; (DALYs): dash adjusted life years; (SDI): Socio-Demographic Index;  (T2DM) :Type 2 Diabetes Mellitus;  (EBV): Epstein-Barr virus; (DMTS): Disease-modifying therapies; (S1P) : sphingosine-1-phosphate;  (PPMS): primary progressive MS;  (NfL), : neurofilament light; (ROS) : reactive oxygen species;  (TNF-α): tumor necrosis factor alpha; (iNOS): inducible nitric oxide synthase; (GFAP): glial fibrillary acidic protein;  (WBCs): white blood cells; (CH13L1): The Chitinase-3-like Protein 1; (ctDNA): Circulating Tumor DNA; (TEPs): Tumour-Educated Platelets; (AI): Artificial intelligence;    (PIRA): Progression independent of relapse activity; (GFAP): Glial Fibrillary Acidic Protein; (LB): lumbar puncture; (SIMOA): single-molecule array; (CSF)correlation of baseline; (CIS): Clinical Isolated Syndrome; (OCBs): oligoclonal bands; (FLAIR): Fluid-Attenuated Inversion Recovery;  (LUMC): Leiden University Medical Center;  (QA): quality assurance; (QC)quality control; (kFLC) : Kappa-Free Light Chain ; (MLS):  machine learning segmentation;  (FDR): False Discovery Rate; (FWE) : Family-Wise Error; (WMH): White Matter Hyperintensity; (TIV) Total Intracranial Volume;  (PLS): Partial least squares;(MICE): Multiple Imputation by Chained Equations;  (MAR): Missing-At-Random; (METC-LDD): the Medical Research Ethics Committee Leiden Den Haag;(HCE): healthy controls ; (EDSS) : Expanded Disability Status Scale; (RRMS): relapsing-remitting MS; (EOMS) : Early Onset MS; (cGFAP): glial fibrillary acidic protein; (PIRA): progression independent of relapse activity; (CPV) : choroid plexus volume; (PVS): perivascular space;  (GMV)Gray matter volume; (cNfl): Cerebrospinal Fluid Neurofilament Light; (TIV) : total intracranial volume; (NBV): normalized brain volume ; (AUROC): area under the receiver operating characteristic;  (PRLS): paramagnetic rimless lesion;

Acknowledgment: 
The authors would like to express their sincere gratitude to the Leiden University Medical Center (LUMC), Leiden University, Albinusdreef 2, 2333 ZA Leiden, the Netherlands, for providing the institutional environment and support that contributed to the successful completion of this research on cerebrospinal fluid biomarkers and disease progression in multiple sclerosis.
The authors also extend their appreciation to the Editorial Office of the Ibn Sina Journal of Medical Sciences, Health and Pharmacy, Noor Al-Ilm Printing, Publishing and Distribution, for their editorial support and for granting a full waiver of the article processing charges (APCs), which facilitated the publication of this work.

Author Contribution: 
All authors contributed equally to the main contributor to this paper. All authors read and approved the final paper.


Declaration of generative AI and AI-assisted technologies in the writing process
The authors hereby declare that no generative artificial intelligence or AI-assisted technologies were used at any stage during the preparation of this manuscript, including language editing, proofreading, or content development. The authors take full responsibility for the originality and integrity of the work presented in this publication.

Funding: 
This research received partial institutional funding from the Leiden University Medical Center (LUMC), Leiden University, Leiden, the Netherlands, under Grant No. LUMC-SD6-2026.
The publication of this article was also supported by the Editorial Office of the Ibn Sina Journal of Medical Sciences, Health and Pharmacy, Noor Al-Ilm Printing, Publishing and Distribution, through a full waiver of the article processing charges (APCs). The funding bodies had no role in the study design, data collection, data analysis, interpretation of the findings, manuscript preparation, or the decision to publish.

Conflicts of Interest: 
“The authors declare no conflict of interest.”