Machine learning based prediction of clinical progression in multiple sclerosis

Machine learning may aid in individualized prediction of disease progression in multiple sclerosis (MS). In this study, we used data from 354 patients with MS to evaluate the capability of different machine learning approaches at predicting future disease worsening. Multiple clinical end points of disease worsening were tested and different combinations of clinical and structural MRI measures were used as inputs. Machine learning models were capable of discriminating between patients with low and high disability but did not perform well in predicting future disease course.