To determine the feasibility of video artificial intelligence (vAI)-based assessments and reliability of extracted digital biomarkers of motor function in people with relapsing multiple sclerosis (RMS) and non-relapsing secondary progressive multiple sclerosis (nrSPMS).

Current clinical tools for monitoring multiple sclerosis (MS) progression often miss early signs of disability accumulation due to poor generalizability, limited sensitivity, and fragmented data capture across outcomes. AI-powered video approaches help derive integrative and patient-centric digital measures to address these issues. The FREXADIGITAL AI substudy pioneered vAI-based assessment and analysis for disease monitoring and treatment response in MS clinical trials.

FREXADIGITAL AI is an optional exploratory substudy of the phase 3 FREXALT (RMS) and FREVIVA (nrSPMS) studies comparing the efficacy and safety of frexalimab to teriflunomide (FREXALT, NCT06141473) and/or placebo (FREVIVA, NCT06141486) over 24 months. Smartphone/tablet video recordings assess gait, balance and upper extremity function in at-hospital (Timed-25-Foot-Walk-Test [T25FWT], 9-Hole-Peg-Test [9HPT]) and at-home settings (Timed-Up-and-Go [TUG], balance and [un]buttoning shirt tests). Video-based AI body pose estimation algorithms are utilised to estimate key body landmarks and to assess objective motor function measures such as symmetry and balance indices, movement duration, speed, error count, sit-to-stand duration, and angles when using upper/lower limbs. Clinical validation includes reliability, anchor analysis against Expanded Disability Status Scale (EDSS), 9HPT and Multiple Sclerosis Walking Scale (MSWS-12), time trend and treatment response estimation for 85 participants per study. A Human Factors (HF) study assessed smartphone application feasibility in 15 representative MS participants (ages 18-60 years; nine with mild, six with moderate disease severity).

HF results showed high overall usability scores (8.3/10). Results from baseline vAI-based assessments of related metrics will be shared. The main results are expected at study completion.

This innovative approach offers a non-invasive, patient-centric, objective, standardized, and accessible method to estimate disease progression and treatment efficacy in people with MS.