Ecosystem Requirements for E-Health and Digital Monitoring in Neuromuscular Disorders

In November 2024, we convened in Munich, Germany, for the 4th Congress to review the latest advancements in e-health and digital monitoring of neuromuscular disorders.

The Pathfinder manuscript of this special issue related to the 4^th eNMD Congress on e-health and digital monitoring in neuromuscular disorders is the report of the 3rd eNMD Congress in 2021. During the COVID-19 pandemic, a small group of experts convened in Pisa, Italy, focusing on the role of remote monitoring and telemedicine in managing neuromuscular diseases, which require multidisciplinary care due to their complex and evolving nature. The event highlighted telemonitoring as a transformative tool in diagnosis, treatment, and research, especially post-COVID-19 when remote technologies became vital. Sessions explored innovations such as wearable biosensors, digital biomarkers, AI applications, and soft robotics for motor rehabilitation. These technologies face standardisation, validation, and legal compliance challenges despite their promise. The congress stressed the need for hybrid care models and the importance of integrating patient-reported data into clinical practice. Ethical and regulatory concerns around data use, privacy, and equitable access were also discussed. A key takeaway was the proposal for a Europe-wide research network to accelerate the implementation of telemedicine tools for NMDs. A collaborative, patient-centred, and tech-enabled approach is essential to improve care and bridge gaps in NMD healthcare systems. So, what happened thereafter?

A new contender in data-sharing is the federated approach. The initial article reviewed by Süwer and team investigates how Federated Learning (FL) – a method of machine learning that prioritises privacy – has the potential to transform research on rare diseases (RDS). With over 6% of the global population impacted by approximately 10,000 RDs, advancements are stymied due to limited patient populations, lack of data, and stringent privacy laws such as GDPR. FL permits various institutions to collaboratively train AI models without sharing sensitive patient information, ensuring confidentiality while broadening access to datasets. This method enhances model accuracy, aids in biomarker identification, improves diagnostics, and allows for personalised treatment approaches. Notable FL applications can be found in cancer detection, telemedicine, and health monitoring using smart devices. Yet, challenges persist, including variability in data, infrastructure needs, and fairness of models. New solutions, such as federated transfer learning and privacy-enhancing methods, are being developed. The article calls for increased international cooperation and investment in FL technologies to fully harness AI's capabilities in RD research, ultimately improving patient outcomes worldwide. An important element is understanding the essentials of machine learning (ML) and its growing role in diagnosis, prognosis, and treatment, as covered by Yeo and colleagues in their paper. It highlights the transformative potential of ML in neuromuscular and electrodiagnostic medicine through the analysis of EMG, NCS, MRI, and -omics data. ML enhances diagnostic precision, supports disease progression prediction, and enables personalised care. The primer also covers the basics of ML models, including training processes, key algorithms, and neural networks. It emphasises the importance of data quality and ethical concerns such as bias, privacy, and transparency. Despite promising research, ML adoption in clinical practice remains limited due to regulatory and interpretability challenges. Future trends include generative AI, cloud-based inferencing, and ambient AI for disease monitoring. Ultimately, clinicians must grasp ML fundamentals to critically assess its applications and safeguard patient outcomes, ensuring that technological advances yield genuine clinical benefits. Siciliano and colleagues examined the prospects and challenges of telemedicine and remote monitoring in NMDs, particularly as the COVID-19 pandemic spurred their uptake. The article emphasises how digital health solutions – such as wearable sensors, mobile applications, and video consultations – facilitate diagnosis, rehabilitation, and ongoing care for various NMDs, including Duchenne muscular dystrophy, ALS, SMA, and myasthenia gravis. Although patients and clinicians express high satisfaction levels, obstacles like a lack of standardisation, restricted technology access, digital literacy deficits, and concerns about data privacy still exist. The review underscores that telemedicine is a valuable complement to in-person care but cannot entirely replace it, given the intricacies of clinical assessments. It also highlights the necessity for validated outcome measures, interoperable systems, and expansive digital inclusion. A combination of telehealth and traditional care is proposed as the ideal approach. Policy endorsement, equitable access, and more vigorous clinical research are vital for effective integration to ensure that digital tools improve care quality and research for NMDs. In accordance with this, the following paper by Stein and colleagues presents the MyaLink study, a telemedicine platform designed for patients with myasthenia gravis (MG). Given the intricate nature of MG management and the scarcity of specialists, this platform was created to enable remote symptom monitoring and enhance communication between patients and physicians. It features a patient app and a physician portal, allowing data exchange through patient-reported outcomes, wearable devices, and digital spirometry. In a 12-week randomised controlled trial with 45 patients, those using MyaLink benefited from continuous remote monitoring and regular telemedical consultations. Initial results indicate a strong interest in enhancing patient care, particularly regarding symptom tracking, medication management, and access to specialists. The study demonstrates the platform's feasibility and usability, suggesting that MyaLink could meet unmet needs by providing personalised, timely care. Additionally, it supports earlier interventions and may improve treatment efficiency while alleviating the burden on healthcare systems. Future research will evaluate long-term clinical outcomes and the potential for broader implementation.

Shifting the focus to neuromuscular rehabilitation and assistive devices, Mijic and colleagues systematically reviewed and evaluated the effectiveness of assistive gait devices (AGDs) in improving mobility and daily activities for individuals with NMDs. Forty studies involving 596 participants assessed various devices, including gait-assisting exoskeletons (GAEs), ankle-foot orthoses (AFOs), orthopaedic footwear (OF), and neuroprostheses. GAEs demonstrated modest improvements in walking endurance (2-min walk test) but no significant gains in speed or functional independence. AFOs and OFs showed no noticeable benefits in walking metrics, while neuroprostheses lacked sufficient evidence. Methodological limitations hindered definitive conclusions, including a high risk of bias, small sample sizes, and diverse study designs. Most studies employed pre-post designs with few randomised controlled trials. Despite the interest of patients and clinicians in AGDs in enhancing independence, the current evidence is of very low certainty. The review emphasises the need for higher-quality studies and standardised outcomes to guide the selection and prescription of AGDS for NMD care. In their systematic review, Calderone and colleagues investigate how virtual reality (VR) and gamification can improve rehabilitation outcomes for individuals with NMDs. Traditional rehabilitation methods often become monotonous and unmotivating, especially for young patients. The review underscores that VR and gamified therapies enhance patient motivation, engagement, and treatment adherence by providing immersive environments, real-time feedback, and interactive activities. The studies reviewed indicate advancements in motor function, balance, cognition, and psychosocial health through VR tools. Specifically, children diagnosed with Duchenne and Becker muscular dystrophies experienced better strength, less pain, and improved mobility after participating in VR-based interventions. Incorporating gamification elements, such as storylines and reward systems, turned exercises into enjoyable activities, promoting regular participation. Furthermore, the review emphasises the necessity for improved study designs and additional research on long-term effects. In conclusion, VR and gamified rehabilitation present promising, patient-centric strategies for addressing NMDs and supporting functional and emotional recovery. In a pilot proof of concept study, Fossomo and colleagues investigated rehabilitation technology's effectiveness in enhancing arm and hand function for adults with myotonic dystrophy type 1 (DM1). They utilised a single-subject experimental design where six participants completed a three-week inpatient rehabilitation program incorporating technology-aided exercises with tools such as AMADEO (for hand dexterity) and ArmeoSenso (for arm movements). Weekly video assessments were performed using standardised tests like the Nine Hole Peg Test (NHPT), grip strength measurements, and patient-reported outcome measures. Five participants demonstrated improved dexterity in their dominant hand, three exhibited gains in range of motion, and two reported increased hand strength. All participants tolerated the intervention well, with fatigue, sleep, and myotonia levels either stable or improved. The study confirmed the feasibility and potential advantages of digital rehabilitation in DM1, particularly for individuals with significant impairments. Although there were limitations, including a small sample size and variability in baseline function, the findings support further exploration into technology-based interventions and telerehabilitation models for neuromuscular disorders. This approach also underscores the usefulness of remote monitoring in enhancing access and continuity of care. Peruzzo and colleagues present a study applying machine learning techniques to investigate brain changes in children diagnosed with Duchenne Muscular Dystrophy (DMD), utilising various MRI data types. DMD predominantly affects muscle tissue but also has repercussions on the central nervous system, frequently causing cognitive and behavioural issues. The researchers performed MRI scans on 18 boys with DMD alongside 18 age-matched control subjects, implementing T1-weighted and Diffusion Tensor Imaging (DTI) data. Employing a multivariate approach alongside Support Vector Machine classification, the study found notable differences in white matter integrity, particularly within the cerebellar peduncles, posterior thalamic radiation, fornix, and medial lemniscus. Additionally, DMD patients displayed decreased cortical thickness, especially in the motor and cingulate cortices, hippocampus, and insula. The classifier attained an accuracy of 97.2%, indicating a strong association between the observed changes and DMD, regardless of comorbid conditions such as autism or intellectual disabilities. This research marks the first instance of combining structural and microstructural MRI data using machine learning to pinpoint a distinctive brain alteration pattern related to DMD. It highlights the value of sophisticated neuroimaging and computational methods in revealing the cerebral implications of DMD and implies a connection between dystrophin deficiency and brain structure abnormalities. These results could pave the way for future studies focusing on targeted therapies and neurodevelopmental assessments in children with DMD. Finally, Barral and Servais share their insights on creating and validating digital outcome measures (DOMs) for interventional clinical trials in DMD. DOMs, such as stride velocity, cadence, and activity counts measured by accelerometers, provide a more continuous, objective, and sensitive method for tracking disease progression and treatment efficacy. Notably, the stride velocity 95th centile (SV95C) has become the first DOM to gain regulatory approval as a primary endpoint for ambulatory DMD patients from the European Medicines Agency. Generally, patients tolerate and accept DOMS well, but challenges persist in areas like compliance, standardisation, and data collection in uncontrolled settings. The integration of artificial intelligence further enhances their capabilities by automating data analysis and pinpointing predictive biomarkers, such as Kinedmd. These advancements improve reliability, potentially reducing the sample sizes and duration required for clinical trials. Although there are limitations in current evidence, primarily due to small sample sizes and differences in study design, the field is progressing, and broader adoption of DOMs is anticipated to boost clinical trial efficiency and therapeutic assessment in DMD. More research is necessary to back regulatory acceptance and fine-tune their application across diverse patient populations and treatment scenarios.

In conclusion, the 4th eNMD Congress advanced digital technologies in diagnosing, monitoring, and treating NMDs, building on the 3rd Congress. It showcased an evolved ecosystem of telemedicine, AI, and digital outcome measures (DOMs), promoting a paradigm shift in NMD care. A key advancement is Federated Learning (FL), enabling privacy-preserving, decentralised AI model training across institutions, essential for rare disease research, where small populations hinder data sharing. FL enhances diagnostic accuracy, biomarker identification, and treatment personalisation. ML and AI analyse EMG, NCS, MRI, and -omics data, improving prognostic models and precision medicine. DOMs like the Stride Velocity 95th Centile (SV95C) show high sensitivity and have been validated as a regulatory endpoint by the European Medicines Agency, a breakthrough in clinical trial methodology. Innovations in remote monitoring, such as wearable biosensors and digital spirometry, facilitate continuous data collection and early intervention. Virtual reality (VR), gamification, and robotic rehabilitation boost therapy engagement. However, challenges like standardisation, long-term validation, regulatory alignment, and equitable access persist. This special issue highlights the shift toward hybrid care models, emphasising the need for international collaboration, regulatory engagement, and investment in scalable infrastructure to enhance digital health outcomes for NMD patients.