Prediction of Disease Severity in Multiple Sclerosis

Machine learning algorithms that integrate clinical, imaging, and omics data can effectively predict disease activity and treatment responses in multiple sclerosis (MS) patients. Analysis of data from 322 MS patients and 98 healthy controls showed that these algorithms achieved intermediate to high accuracy in predicting outcomes like disability accumulation and therapy escalation. The algorithms performed well with clinical and imaging data alone, with slight improvements when omics data was included, demonstrating their potential to aid in identifying patients at risk for worsening disability.