By Amy Easton, Senior Director of Scientific Programs, Target ALS
Amyotrophic lateral sclerosis (ALS) has long grappled with a significant challenge in clinical research: the limited and often inadequate tools available to measure disease progression. Despite advances in medical research, the gold standard for tracking ALS—the ALS Functional Rating Scale (ALSFRS)—has remained largely unchanged for decades. Though useful, the scale has limitations that hinder both clinical care and drug development. Digital biomarkers, such as those offered by Modality.AI, now present an opportunity to revolutionize how ALS progression is tracked. By providing more precise, continuous, and nuanced measurements, these tools could transform drug development and personalized care for ALS patients.
The ALSFRS evaluates patients across 14 functional domains, such as speech, swallowing, and limb movement, with each domain scored from 0 (no function) to 4 (normal ability). Despite its widespread adoption, the tool is often criticized for its inherent subjectivity and lack of granularity. Patient or clinician interpretations of questions like “How well can you cut your food?” can vary, introducing inconsistencies in the data. Additionally, the infrequency of assessments, typically administered every three to four months, results in sparse data points that are vulnerable to variability based on a patient’s “good days” or “bad days.” Perhaps most critically, the ALSFRS fails to capture subtle changes that may indicate disease progression, such as slight tremors or gradual facial muscle deterioration. In essence, it is akin to using a sledgehammer to fix a watch—a blunt instrument for a complex and nuanced disease.
One of the key advantages of digital health technologies is their ability to enable frequent and objective monitoring, a direct measurement of patients doing tasks that require respiratory, speech, cognitive, emotion, and motor function. Rather than relying on sporadic clinical visits and a questionnaire, this offers a more comprehensive method for capturing data, where patients can perform 15- to 20-minute tasks daily or weekly. This approach generates dense data streams that can be far more representative of the patient’s overall condition than standard assessments. The other advantage of DHTs is the ability to measure patient symptoms at-home, important for enabling and supporting decentralized clinical studies for patients where mobility is prohibitive.
Researchers have been evaluating the utility of at-home digital health technologies across neurodegenerative disease areas for many years. There are a number of good studies demonstrating promise for devices that detect features of motor, speech, sleep, or cognitive impairment, comparable to standard clinical assessments. However, questions remain whether these platforms provide measures that have sufficient sensitivity and specificity to be relied upon for early detection, phenoconversion in genetic mutation carriers, or as a functional biomarker readout for clinical trials. There are a number of emerging digital technologies that are being evaluated to support these use cases. The holy grail would be identification of one device that supports multiple digital platforms to measure the spectrum of clinical symptoms.
Modality has taken a novel approach by using an AI-based digital platform that integrates video and audio-based data to measure micro-changes in a patient’s condition. This platform has the capability to provide a comprehensive analysis across many symptom domains and may even detect new features of these clinical symptoms. In the future, the platform may also be used to help predict disease progression, offering a powerful prognostic tool for both clinical care and research.
Target ALS is proud to partner with Modality to evaluate patient symptoms over time in a large Global Natural History Study. Patients come into premier ALS clinics every 4 months for a period of 16-18 months where they perform traditional clinical assessments, neurological exams, and donate urine, blood, and CSF specimens for genetic and biomarker research. At-home, they are asked to use Modality platform every two weeks and are guided through a short series of tasks by Tina. The tasks are designed to measure facial muscle movements, such as asymmetrical twitching, changes in speech patterns indicating vocal cord weakness, and changes in limb strength and functioning during activities of daily living, new for this particular study. For the first time, the Modality audio-visual metrics will be compared to the Patient Self-report of Problems (PROP), an important comparison needed to support regulatory requirements in clinical trials. The study is designed to provide a robust evaluation of disease progression across thousands of ALS patients and healthy controls and will enable comparison of different digital platforms to fluid biomarkers of neurodegeneration to gold standard assessments.
Modality’s platform can be easily adapted across different languages and cultural contexts. This makes it particularly valuable for conducting trials in diverse geographical locations such as Puerto Rico and Colombia, sites where Target ALS is conducting its GNHS. Although the prevalence of ALS is similar between Spanish-speaking populations to people of European-descent, there is underrepresentation of spanish-speaking patients in clinical trials and lack of clinical trials run in Latin and South America. Our study aims to characterize ALS patients across ancestries, learning from the differences and similarities across countries, and lay the needed groundwork to enable future clinical trials outside the US.
While these new digital technologies offer hope, challenges remain. One concern is the potential burden on patients to use an app on their phone or computer on a regular basis; the need for active participation may be overwhelming for some individuals. Simplifying protocols or integrating passive sensors could help address this issue. Regulatory validation is another hurdle. The FDA requires rigorous validation of new clinical outcome measures, particularly if used to approve a new drug. Target ALS is working together with the Accelerating Medicines Partnership for ALS (AMP ALS) in a strategic initiative to conduct forward-thinking research that will facilitate regulatory acceptance of new technologies. Finally, global equity must be a priority. Ensuring that patients in low-resource settings have access to these innovations is essential for creating truly inclusive and effective solutions.
Despite these challenges, digital biomarkers represent a paradigm shift in ALS research and care. The ALSFRS is unlikely to disappear overnight – Target ALS too is currently utilizing it as the current gold standard. However, the granular, real-world data captured by tools like Modality AI offer a transformative path forward. With the potential to accelerate drug development, provide personalized patient insights, and support truly global clinical trials, digital biomarkers bring us closer to turning ALS from a fatal diagnosis into a manageable condition. As the ALS community continues to rally behind these groundbreaking tools, hope is on the horizon—one data point at a time.
Interested in learning more? I joined Modality.AI on 14 February to discuss direct-to-participant ALS research. Watch the recording here.
Resources:
- Neumann, Kothare and Ramanarayanan (2023). Multimodal speech biomarkers for remote monitoring of ALS disease progression is a recently published journal article that describes the responsiveness, sensitivity and clinical utility of specific, timing-related speech biomarkers (including speaking duration and canonical timing alignment, CTA) towards tracking longitudinal ALS disease progression in as little as 2 months.
- Neumann et al. (2024). Multimodal Digital Biomarkers for Longitudinal Tracking of Speech Impairment Severity in ALS: An Investigation of Clinically Important Differences explores how clinical meaningful these metrics are by diving in depth into the idea of MCID, and explores listener effort for clinical meaningfulness.
- Neumann et al. (2023). Combining Multiple Multimodal Speech Features into an Interpretable Index Score for Capturing Disease Progression in Amyotrophic Lateral Sclerosis describes how we can optimally combine multiple speech and facial features into a single, interpretable index score for ALS.
