Artificial Intelligence in Multiple Sclerosis: Possibilities in Radiological Diagnostics and Progression Assessment

Abstract

Background

Multiple sclerosis (MS) is characterized by clinical and radiological heterogeneity. Recent refinements in the McDonald criteria and the integration of advanced MRI biomarkers, such as the central vein sign and paramagnetic rim lesions, can enhance diagnostic precision but may also push the manual radiological workload to an unsustainable level. This growing diagnostic complexity makes artificial intelligence (AI) a critical area of development in MS radiology to address the dual challenges of feasibility and personalized care.

Materials and methods

Relevant studies were identified through a literature search in PubMed of articles published between January 1, 2020, and August 31, 2025. Additional studies were included through manual searching of references.

Results

This review examines the current landscape of AI applications in this field, with a particular focus on deep learning. It details how AI can automate lesion quantification and aid differential diagnosis, as well as how it is being developed to make the evaluation of complex biomarkers clinically practical. The review also analyzes the emerging evidence for AI in prognostic modelling and treatment optimization. We argue that robust development of AI in MS depends on the integration of multimodal data. Although commercial volumetric tools exist, the integration into clinical practice presents recognized challenges, including the need for large-scale validation datasets and ethical frameworks.

Conclusion

AI is thus positioned as an essential technological response to the evolving demands of modern, personalized MS care.

Key points

• Recent MS criteria create an increasing radiological workload

• Artificial intelligence offers a possible solution

• Robust AI development needs multimodal data integration

• Commercial tools already automate lesion segmentation and volumetric analysis

• Clinical adoption requires large-scale validation and ethical frameworks

Citation Format

• Müller D, Bellenberg B, Lukas C. Artificial Intelligence in Multiple Sclerosis: Possibilities in Radiological Diagnostics and Progression Assessment. Rofo 2024; DOI 10.1055/a-2808-0083

Zusammenfassung

Hintergrund

Multiple Sklerose (MS) ist durch klinische und radiologische Heterogenität gekennzeichnet. Durch die kürzlich vorgenommenen Verfeinerungen der McDonald-Kriterien und die Integration erweiterter MRT-Biomarker, wie dem zentralen Venenzeichen und paramagnetischen Randläsionen, kann die Diagnosegenauigkeit erhöht werden, was jedoch auch die manuelle radiologische Arbeitsbelastung auf ein untragbares Maß erhöht. Diese zunehmende diagnostische Komplexität macht künstliche Intelligenz (KI) zu einem wichtigen Entwicklungsbereich in der MS-Radiologie, um die doppelte Herausforderung der Umsetzbarkeit und personalisierten Patientenversorgung zu bewältigen.

Methode

Relevante Studien wurden durch eine Literaturrecherche in PubMed nach Artikeln identifiziert, die zwischen dem 1. Januar 2020 und dem 31. August 2025 veröffentlicht wurden. Weitere Studien wurden durch manuelle Suche in Referenzen hinzugefügt.

Ergebnisse

Diese Übersicht untersucht die aktuelle Landschaft der KI-Anwendungen, insbesondere des Deep Learning, in diesem Bereich. Wir stellen dar, wie KI die Quantifizierung von Läsionen automatisieren und die Differenzialdiagnose unterstützen kann, sowie wie KI weiterentwickelt wird, um die Nutzung komplexer Biomarker klinisch praktikabel zu machen. Darüber hinaus analysieren wir die neuen Erkenntnisse zur KI in der Prognosemodellierung und Behandlungsoptimierung. Wir vertreten die Auffassung, dass eine robuste KI-Entwicklung im Bereich MS von der Integration multimodaler Daten abhängt. Zwar gibt es kommerzielle volumetrische Tools, doch ihre Integration in die klinische Praxis steht vor bekannten Herausforderungen, darunter die Notwendigkeit groß angelegter Validierungsdatensätze und der Klärung der ethischen Rahmenbedingungen.

Schlussfolgerung

KI wird daher als eine wichtige technologische Antwort auf die sich wandelnden Anforderungen einer modernen, personalisierten MS-Versorgung angesehen.

Kernaussagen

• Aktuelle MS-Kriterien führen zu einer zunehmenden radiologischen Arbeitsverdichtung.

• Künstliche Intelligenz bietet eine mögliche Lösung.

• Eine robuste KI-Entwicklung erfordert die Integration multimodaler Daten.

• Kommerzielle Tools automatisieren bereits die Segmentierung von Läsionen und die volumetrische Analyse.

• Die klinische Umsetzung erfordert eine umfassende Validierung und ethische Rahmenbedingungen.

Keywords

multiple sclerosis - artificial intelligence - deep learning - MR-imaging - diagnosis - prognosis

Publication History

Received: 16 December 2025

Accepted after revision: 03 February 2026

Article published online:
26 February 2026

• References

• 1 Thompson AJ. et al. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria. Lancet Neurol 2018; 17: 162-173
• 2 Solomon AJ. et al. Differential diagnosis of suspected multiple sclerosis: An updated consensus approach. Lancet Neurol 2023; 22: 750-768
• 3 Montalban X. et al. Diagnosis of multiple sclerosis: 2024 revisions of the McDonald criteria. Lancet Neurol 2025; 24: 850-865
• 4 Gaetani L. et al. 2017 revisions of McDonald criteria shorten the time to diagnosis of multiple sclerosis in clinically isolated syndromes. J Neurol 2018; 265: 2684-2687
• 5 Sastre-Garriga J. et al. MAGNIMS consensus recommendations on the use of brain and spinal cord atrophy measures in clinical practice. Nat Rev Neurol 2020; 16: 171-182
• 6 Wattjes MP. et al. 2021 MAGNIMS–CMSC–NAIMS consensus recommendations on the use of MRI in patients with multiple sclerosis. Lancet Neurol 2021; 20: 653-670
• 7 Lublin FD. et al. Defining the clinical course of multiple sclerosis: The 2013 revisions. Neurology 2014; 83: 278-286
• 8 Patzig M. et al. Comparison of 3D cube FLAIR with 2D FLAIR for multiple sclerosis imaging at 3 Tesla. Rofo 2014; 186: 484-8
• 9 Sati P. et al. FLAIR*: A combined MRI contrast technique for visualizing white matter lesions and parenchymal veins. Radiology 2012; 265: 926-932
• 10 Cortese R. et al. Value of the central vein sign at 3T to differentiate MS from seropositive NMOSD. Neurology 2018; 90: e1183-e1190
• 11 Maggi P. et al. Central vein sign differentiates multiple sclerosis from central nervous system inflammatory vasculopathies. Ann Neurol 2018; 83: 283-294
• 12 Bagnato F. et al. Tracking iron in multiple sclerosis: A combined imaging and histopathological study at 7 Tesla. Brain 2011; 134: 3602-3615
• 13 Dal-Bianco A. et al. Slow expansion of multiple sclerosis iron rim lesions: Pathology and 7 T magnetic resonance imaging. Acta Neuropathol 2017; 133: 25-42
• 14 Absinta M. et al. Association of chronic active multiple sclerosis lesions with disability in vivo. JAMA Neurol 2019; 76: 1474-1483
• 15 Solomon AJ. et al. The contemporary spectrum of multiple sclerosis misdiagnosis: a multicenter study. Neurology 2016; 87: 1393-1399
• 16 Lui YW. et al. Artificial intelligence in neuroradiology: Current status and future directions. Am J Neuroradiol 2020; 41: E52-E59
• 17 Goodfellow I, Bengio Y, Courville A. Deep Learning, MIT Press. 2016
• 18 Shoeibi A. et al. Applications of deep learning techniques for automated multiple sclerosis detection using magnetic resonance imaging: A review. Comp Biolog Med 2021; 136: 104697
• 19 Aslam N. et al. Multiple sclerosis diagnosis using machine learning and deep learning: Challenges and opportunities. Sensors 2022; 22: 7856
• 20 Eitel F. et al. Uncovering convolutional neural network decisions for diagnosing multiple sclerosis on conventional MRI using layer-wise relevance propagation. Neuroimage Clin 2019; 24: 102003
• 21 Lopatina A. et al. Investigation of deep-learning-driven identification of multiple sclerosis patients based on susceptibility-weighted images using relevance analysis. Front Neurosci 2020; 14: 609468
• 22 Zhang Y. et al. Grad-CAM helps interpret the deep learning models trained to classify multiple sclerosis types using clinical brain magnetic resonance imaging. J Neurosci Methods 2021; 353: 109098
• 23 Kim H. et al. Deep learning-based method to differentiate neuromyelitis optica spectrum disorder from multiple sclerosis. Front Neurol 2020; 11: 599042
• 24 Rocca MA. et al. Deep learning on conventional magnetic resonance imaging improves the diagnosis of multiple sclerosis mimics. Invest Radiol 2021; 56: 252-260
• 25 Cortese R. et al. Deep Learning Modeling to Differentiate Multiple Sclerosis From MOG Antibody–Associated Disease. Neurology 2025; 105: e214075
• 26 Tan IL. et al. MR venography of multiple sclerosis. Am J Neuroradiol 2000; 21: 1039-1042
• 27 Tallantyre EC. et al. Ultra-high-field imaging distinguishes MS lesions from asymptomatic white matter lesions. Neurology 2011; 76: 534-539
• 28 Maggi P. et al. CVSnet: A machine learning approach for automated central vein sign assessment in multiple sclerosis. NMR in Biomed 2020; 33: e4283
• 29 Bozsik B. et al. Reproducibility of lesion count in various subregions on MRI scans in multiple sclerosis. Front Neurol 2022; 13: 843377
• 30 Afzal HMR. et al. The emerging role of artificial intelligence in multiple sclerosis imaging. Mult Scler J 2022; 28: 849-858
• 31 Smeets D. et al. Reliable measurements of brain atrophy in individual patients with multiple sclerosis. Brain and behavior 2016; 6: e00518
• 32 Haider L. et al. Linear brain atrophy measures in multiple sclerosis and clinically isolated syndromes: A 30-year follow-up. J Neurol Neurosurg Psych 2021; 92: 839-846
• 33 Goebl P. et al. Enabling new insights from old scans by repurposing clinical MRI archives for multiple sclerosis research. Nat Commun 2025; 16: 3149
• 34 Durso-Finley J, Douglas LA, Tal A. “Saliency based deep neural network for automatic detection of gadolinium-enhancing multiple sclerosis lesions in brain MRI.”. In: Crimi A, Bakas S. , ed. Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries. Cham: Springer; 2019: 108-118
• 35 Preetha CJ. et al. Deep-learning-based synthesis of post-contrast T1-weighted MRI for tumour response assessment in neuro-oncology: A multicentre, retrospective cohort study. Lancet Digit Health 2021; 3: e784-e794
• 36 Gong E. et al. Deep learning enables reduced gadolinium dose for contrast‐enhanced brain MRI. J Magn Reson Imaging 2018; 48: 330-340
• 37 Martire MS. et al. What is the potential of paramagnetic rim lesions as diagnostic indicators in multiple sclerosis?. Expert Rev Neurother 2022; 22: 829-837
• 38 Clarke MA. et al. Value of 3T susceptibility-weighted imaging in the diagnosis of multiple sclerosis. Am J Neuroradiol 2020; 41: 1001-1008
• 39 Barquero G. et al. RimNet: A deep 3D multimodal MRI architecture for paramagnetic rim lesion assessment in multiple sclerosis. Neuroimage Clin 2020; 28: 102412
• 40 Zhang H. et al. QSMRim-Net: Imbalance-aware learning for identification of chronic active multiple sclerosis lesions on quantitative susceptibility maps. Neuroimage Clin 2022; 34: 102979
• 41 Tousignant A. et al. Prediction of disease progression in multiple sclerosis patients using deep learning analysis of MRI data. Proceedings of Machine Learning Research. In: . 2019: 483-492
• 42 Storelli L. et al. A deep learning approach to predicting disease progression in multiple sclerosis using magnetic resonance imaging. Invest Radiol 2022; 57: 423-432
• 43 Coll L. et al. Deep learning to predict progression independent of relapse activity at a first demyelinating event. Brain Commun 2025; 7.4: fcaf243
• 44 Falet J-PR. et al. Estimating individual treatment effect on disability progression in multiple sclerosis using deep learning. Nat Commun 2022; 13: 5645
• 45 Praet J. et al. A future of AI-driven personalized care for people with multiple sclerosis. Front Immunol 2024; 15: 1446748
• 46 Praet J. et al. RECLAIM – A retrospective, multicenter observational study aimed at enabling the development of artificial intelligence-driven prognostic models for disease progression in multiple sclerosis. Front Neurol 2025; 16: 1557947
• 47 Mendelsohn Z. et al. Commercial volumetric MRI reporting tools in multiple sclerosis: A systematic review of the evidence. Neuroradiology 2023; 65: 5-24
• 48 Lekadir K. et al. FUTURE-AI: International consensus guideline for trustworthy and deployable artificial intelligence in healthcare. BMJ 2025; 388: e081554
• 49 Barnett M. et al. A real-world clinical validation for AI-based MRI monitoring in multiple sclerosis. NPJ Digit Med 2023; 6: 196