Establishing Walking-related Digital Biomarkers in Rare Childhood Onset Progressive Neuromuscular Disorders

PROJECT SUMMARY Individuals with childhood onset, genetically determined neuromuscular disorders (NMD) experience progressive neuromuscular degeneration resulting in weakness that affects ambulatory function. Genetic testing and consensus-derived diagnosis guidelines have led to an increasing prevalence of neuromuscular diseases approaching other more common neurological conditions. Recent therapeutic approaches in Duchenne muscular dystrophy (DMD) and spinal muscular atrophy (SMA) have resulted in disease modifying therapies commercially available or in clinical trials with the best outcomes with early treatment. While results from novel drug therapies are promising, symptoms of weakness and impaired function persists. Patients report changes in motor function in real-world settings, but these changes often elude standard in-clinic examinations. There is a compelling need to develop more sensitive, quantitative assessments that focus on ambulatory function in real- life settings. Digital mobility outcomes (DMOs) use wearable technology for continuous remote monitoring in a person’s own environment. Consumer-grade activity trackers monitor healthy and clinical cohorts, but with limited accuracy. Research grade trackers and foot-worn gait monitoring devices measure mobility-related volume (step count, distance, duration of walking bouts, etc.), but can only capture a minimal set of stride-by-stride real-life gait parameters (e.g., stride time, length, velocity) and they cannot measure kinetic parameters. We argue that the modest accuracy and granularity of current wearable devices hamper the objective characterization of subtle but clinically meaningful changes. Our recent findings suggest that novel machine-learning (ML)-based abstraction models may map noisy signals from foot-worn sensors (namely, instrumented insoles developed by the project team) into accurate and clinically relevant spatiotemporal and kinetic gait parameters. These gait parameters derived from instrumented insoles, which we refer to as digital mobility outcomes (DMOs), may serve as functional biomarkers to detect changes in real world function. The purpose of this research is twofold: (1) to identify disease-specific walking-related digital biomarkers of disease severity, and (2) to determine the insole-derived DMOs that are most sensitive to longitudinal changes in ambulatory function and best able to predict 12-month changes in ambulatory function in DMD and SMA.