Related Experiment Video
Updated: Jul 18, 2026
10:46
A Method of Trigonometric Modelling of Seasonal Variation Demonstrated with Multiple Sclerosis Relapse Data
Published on: December 9, 2015
11.0K
Integrating big data and artificial intelligence to predict progression in multiple sclerosis: challenges and the
Hamza Khan1,2,3, Sofie Aerts1,4, Ilse Vermeulen1,2
1Biomedical Research Institute (BIOMED), University MS Center, Hasselt University, Agoralaan Building C, 3590, Diepenbeek, Belgium.
Journal of Neuroengineering and Rehabilitation
|September 30, 2025
Summary
Artificial intelligence (AI) can improve multiple sclerosis (MS) care by integrating diverse data. Addressing regulatory and ethical challenges is key to implementing AI for better MS prognosis and personalized patient outcomes.
Area of Science:
- Neurology
- Medical Informatics
- Artificial Intelligence
Background:
- Multiple sclerosis (MS) is a complex neurological condition with progressive disability, necessitating early detection and accurate prognosis.
- Integrating multimodal real-world data (clinical records, MRI, digital biomarkers) with AI for personalized MS care faces limitations.
- A gap exists between AI technical innovation and clinical implementation due to methodological, regulatory, and ethical barriers.
Purpose of the Study:
- To identify and analyze the gap between AI innovation and clinical implementation in multiple sclerosis (MS) care.
- To explore challenges in integrating real-world data, regulatory hurdles, and ethical concerns (bias, privacy, equity).
- To highlight emerging solutions and advocate for strategies to bridge the implementation gap.
Main Methods:
- Perspective paper analyzing the challenges and opportunities of AI in MS care.
- Exploration of the underuse of integrated real-world data, regulatory/ethical barriers, and potential solutions.
- Review of strategies including federated learning, regulatory initiatives (DARWIN-EU, EHDS), patient-led frameworks (PROMS, CLAIMS), and foundation models.
Main Results:
- Significant limitations exist in integrating multimodal real-world data for AI-driven MS care.
- Methodological constraints, evolving regulations, and ethical concerns impede clinical AI implementation.
- Promising solutions include federated learning, data infrastructure harmonization, patient-centered design, explainable AI, and real-world validation.
Conclusions:
- Bridging the gap requires aligning technical, regulatory, and ethical domains for effective AI implementation in MS.
- Harmonized data infrastructures, patient-centered design, explainable AI, and real-world validation are crucial.
- Successful AI integration can enhance MS prognosis, personalize care, and improve patient outcomes.
