By Clinical Research News Staff 

February 11, 2026 | Clinical research in amyotrophic lateral sclerosis (ALS) has long been hindered by subjective endpoints, slow enrollment, and limited tools for tracking disease progression. EverythingALS, a nonprofit initiative, is addressing those barriers with a speech-based digital biomarker designed to modernize how ALS trials are conducted and evaluated. 

After less than three years of data collection, the Peter Cohen Foundation—operating as EverythingALS—is awaiting FDA review of a machine learning algorithm intended for use as a secondary endpoint in ALS clinical trials. The tool is a “listener effort” prediction model, a clinical measure traditionally scored by speech pathologists, using just one minute of recorded speech. If cleared, it would represent the first objective, scalable digital biomarker deployed in neurological disease research, according to founder and CEO Indu Navar. 

The clinical rationale is straightforward. ALS trials have historically relied on patient-reported outcomes and clinician questionnaires, which are inherently variable and poorly suited to detecting subtle changes over time. That subjectivity has contributed to larger, longer trials that frequently fail to meet regulatory thresholds. An objective, quantitative biomarker could reduce variability, increase statistical power, and enable smaller, faster studies. 

The algorithm was trained and validated using more than 1,200 participants and multiple independent datasets, including samples from the HEALEY ALS Platform Trial led by Massachusetts General Hospital. Three unaffiliated speech pathologists initially rated 6,000 samples, and the model matched their assessments 96% of the time. Subsequent testing across thousands more recordings demonstrated consistent performance in tracking dysarthria progression as ALS advances. 

Importantly for trial operations, the model requires no specialized equipment or clinic visits. Audio recordings alone were sufficient, making the approach compatible with decentralized and hybrid trial designs. Beyond speech, EverythingALS is applying a similar framework to other clinically relevant domains, including respiratory function, gait, balance, and fall risk.  

The organization is also addressing another persistent challenge in ALS research: enrollment. Fewer than 10% of people with ALS currently participate in trials, despite high reported interest. Diagnostic delays, restrictive eligibility windows, and limited physician awareness all contribute to missed opportunities. EverythingALS is conducting a baselining study, scheduled to launch in early 2026, to characterize disease severity and progression early and facilitate rapid matching to trials as they open. 

To read the full story written by Deborah Borfitz, go to Diagnostics World News.