Machine learning models based on conventional MRI predicts clinical worsening in multiple sclerosis

Early identification of patients with multiple sclerosis (MS) at high risk of clinical and cognitive deterioration is possible using conventional MRI imaging and machine learning, according to researchers in Italy.

The researchers collected T2- and T1-weighted brain MRI images from over 370 patients with MS, using the Italian Neuroimaging Network Initiative. Changes in clinical functioning (Expanded Disability Status Scale [EDSS]) and cognitive functioning (Symbol Digit Modalities Test [SDMT]) over a 2-year period were also collected.