Radiomics and Artificial Intelligence in Multiple Sclerosis MRI: A Comprehensive Review.
Authors
Affiliations (2)
Affiliations (2)
- From the From the Department of Applied Medical Physics, Medical School (K.P., A.P., E.P.E.), 2nd Department of Radiology (G.V.), University General Hospital Attikon, National and Kapodistrian University of Athens, Rimini 1, Str, Haidari, 12462, Athens, Greece; Medical School (I.S.), National and Kapodistrian University of Athens, 75 Mikras Assias str., 11527, Athens, Greece; Medical Physics Laboratory (E.K.), Democritus University of Thrace, 69100 Alexandroupolis, Greece.
- From the From the Department of Applied Medical Physics, Medical School (K.P., A.P., E.P.E.), 2nd Department of Radiology (G.V.), University General Hospital Attikon, National and Kapodistrian University of Athens, Rimini 1, Str, Haidari, 12462, Athens, Greece; Medical School (I.S.), National and Kapodistrian University of Athens, 75 Mikras Assias str., 11527, Athens, Greece; Medical Physics Laboratory (E.K.), Democritus University of Thrace, 69100 Alexandroupolis, Greece. [email protected].
Abstract
Radiomics is a process of extracting quantitative features from medical images, such as MRI. This process, combined with artificial intelligence, has already been investigated in several studies on the management of multiple sclerosis. The aim of this review article was to provide an overview of the various applications of MRI radiomics in the diagnosis and prognosis of multiple sclerosis. The literature search was conducted in PubMed and Scopus for articles published between 2015 and 2025. A total of 26 articles met the specified criteria. Studies found that radiomics features from brain MRI images, combined with Artificial Intelligence models, were able to distinguish between healthy tissues and multiple sclerosis lesions, predict disability, detect disease activity, and differentiate between conditions with similar symptoms. The extraction of radiomic features and their utilization with Artificial Intelligence models could enhance the effectiveness of multiple sclerosis management. However, several limitations, such as an unbalanced dataset and a lack of external validation, must be addressed before they can be integrated into clinical practice.