Identification of Novel Smartphone-Based Digital Biomarkers to Characterize Lower and Upper Limb Motor Functions in Patients with Duchenne Muscular Dystrophy
Shotaro Tachibana1, Allison Grange2, Dominique Vincent-Genod3, Adeline Allara3, Mathilde Hirel3, Claude Cances4, Nicolas Chrestian5, Loïc Carment6, Arthur Feyt6, Clarissa Gorin6, Cédric Gormong6, Juan Camilo Luna-Escalante6, Alexandre Petitmangin6, Cecile Halbert7, Vincent Laugel8
1Hospices Civilis de Lyon, CHU Lyon, Escale Paediatric PMR Department, Lyon, France, 2I-Motion, Institut de Myologie, Paris, France, 3Service de Rééducation Fonctionnelle Pédiatrique des Hospices Civils de Lyon, Hôpital Femme Mère Enfant, Bron, France, 4Neuropaediatric Department, AOC (Atlantic-Oceania-Caribbean) Reference Centre for Neuromuscular Disorders, Toulouse University Hospital, Toulouse, France, 5Centre Mère Enfant Soleil CHU de Québec, Laval university, Quebec, Canada, 6Ad Scientiam, Paris, France, 7Centre de Référence des Maladies Neuromusculaires de l'enfant PACARARE, Service de Neuropédiatrie, Hôpital de la Timone Enfants, Marseille, France, 8Department of Pediatric Neurology, Strasbourg University Hospital, Strasbourg, France
Objective:
To assess the feasibility of smartphone-based digital biomarkers (dBMKs) to objectively measure lower and upper limb motor functions in patients with Duchenne muscular dystrophy (DMD).
Background:
DMD is a severe muscular dystrophy characterized by progressive muscle atrophy and weakness. Current clinical measures have limitations (eg., inter-rater variability) and are infrequent, which could impede patient’s care. These challenges could be overcome with smartphone-based dBMKs that offer innovative approaches to track the progression of functional parameters, from the patient's home.
Design/Methods:
A literature review and interviews with 11 experts (neuropaediatricians, physiotherapists, occupational therapists) was performed to identify and prioritize meaningful health concepts (HCs) and candidate dBMKs. A Proof-of-Concept study was conducted at a myology institute with 2 physiotherapists simulating different degrees of disease severity (mild, moderate, severe). We then assessed the feasibility of the measurement process (sensors and algorithms) to collect and transform these dBMKs in a Timed-up-and-Go (TUG) test and upper limb motor (ULM) test.
Results:
Ambulation, balance and transition tasks were prioritized as main HCs to capture key motor functions. 11 candidate dBMKs derived from accelerometer and gyroscope were selected for TUG and first results confirmed feasibility with a mean absolute error (MAE) of 1.6±1s when automatically detecting transition phases. Error was severity-dependent with a MAE of 0.8s in mild, versus 2.1s in severe phenotypes. The ULM test (ability to bring the hand to mouth, head and above), was captured using camera recording and an automated keypoint detection model. Feasibility was confirmed and a larger list of dBMKs was identified to measure arm function and compensation strategies.
Conclusions:
These insights pave the way for a promising approach to develop a smartphone-based measurement device to collect dBMKs in people living with DMD, and may provide clinicians with objective data collected from the patient's home to support medical decisions.