AAN 2022 | Predicting treatment effect in MS patients: enrichment of certain baseline features in responders

Jean-Pierre Falet, MD, CM, McGill University, Montreal, Canada, discusses findings from a model used to estimate treatment response for patients with multiple sclerosis (MS). Neural networks are notorious for being difficult to interpret; having complex, non-linear functions, there is difficulty obtaining coefficients that can explain the likely response to medication. One method of attempting to interpret the prediction of the neural network is to compare the baseline characteristics of the predicted responder group to the predicted non-responder group. The study noted that Certain features were notably enriched across the variety of thresholds for what is considered a responder and non-responder. Responders tended to be younger, mostly men, have more MRI activity at baseline, and exhibit a shorter time from symptom onset. They also had greater disability scores across multiple disability metrics. This interview took place at the American Academy of Neurology 2022 Congress in Seattle, WA.