Estimating individual treatment effect on disability progression in multiple sclerosis using deep learning

Disability progression in multiple sclerosis (MS) is challenging to treat due to the lack of suitable biomarkers for phase 2 clinical trials. This study proposes a deep-learning predictive enrichment strategy to enhance statistical power in short proof-of-concept trials. A multi-headed multilayer perceptron (MLP) estimates the conditional average treatment effect (CATE) using baseline clinical and imaging features. Patients predicted to be most responsive are preferentially randomized. Using data from six randomized clinical trials (n=3,830), the model showed larger average treatment effects for the most responsive patients compared to the entire group for both anti-CD20 antibodies and laquinimod. This approach significantly increases the power of short trials, accelerating therapeutic advancements.

Jean-Pierre R. Falet, Joshua Durso-Finley, Brennan Nichyporuk, Julien Schroeter, Francesca Bovis, Maria-Pia Sormani, Doina Precup, Tal Arbel, Douglas Lorne Arnold