Can machine-learning models differentiate between radiologically- and clinically-isolated syndromes?

Takeaway

  • Machine-learning models, when applied to multimodel magnetic resonance imaging (MRI), can differentiate between the earliest clinical expressions of multiple sclerosis (MS) with an accuracy of 78%.

Why this matters

  • Differentiation between radiologically isolated syndrome (RIS) and clinically isolated syndrome (CIS) will ensure disease-modifying therapies are administered to appropriate patients, but this cannot currently be done on a single-patient level.