What You Should Know:
– Two newly published peer-reviewed studies validate that Linus Health’s AI-driven, 3-minute digital assessment can concurrently detect cognitive impairment and predict underlying Alzheimer’s pathology.
– This “multiplexing” capability allows primary care providers to identify at-risk patients years before symptoms appear, potentially streamlining specialist referrals and improving eligibility for new disease-modifying treatments.
Linus Health Study Confirms AI Clock Drawing Test Predicts Alzheimer’s Biomarkers
For years, the digital health sector has promised “early detection” of Alzheimer’s that rarely survived the transition from a controlled lab to a chaotic primary care workflow. However, the data published in Alzheimer’s Research & Therapy and Annals of Clinical and Translational Neurology signals a shift from marketing “fluff” to clinical utility.
By leveraging “process metrics”—analyzing how a patient draws (pauses, pen pressure, stroke velocity) rather than just the final image—Linus Health is moving toward a behavioral surrogate biomarker. For a healthcare system facing 18-month wait times for neurologists, a 3-minute “multiplexed” test that flags both cognitive status and pathology likelihood is a significant operational win.
Key Evidence: The “Multiplexing” Breakthrough
The core of the announcement is the Digital Clock and Recall (DCR) assessment. Unlike traditional “pass/fail” paper tests, the AI analyzes a single stream of behavioral data to power separate models.
1. Concurrent Detection of Pathology and Impairment
In a clinical analysis of 930 participants from the Bio-Hermes-001 study, the DCR demonstrated:
- Pathology Prediction: An AUC of 0.81 for predicting Amyloid-beta (Aβ) PET status, outperforming traditional cognitive tests like the MMSE (AUC 0.70) and RAVLT (AUC 0.73).
- Impairment Detection: An AUC of 0.83 for classifying cognitive impairment, proving superior to the FDA-listed Cognivue Clarity (AUC 0.75).
- Synergy with Biomarkers: Combining the DCR with p-tau217 blood tests increased predictive accuracy to AUC 0.91, suggesting that digital tools can significantly “boost” the value of expensive blood-based biomarkers (BBMs).
2. Predicting Longitudinal Decline
An independent study by Massachusetts General Hospital involving 204 cognitively normal adults found that those with higher amyloid or tau burden on PET scans showed faster decline on the DCTclock over time.
- Specific Insight: The decline was most prominent in the “Information Processing” domain, driven by subtle pen-stroke latencies—details invisible to the human eye but captured at 120+ Hz by a digital stylus.
Impact of AI-Enabled Digital Cognitive Assessments (DCAs)
Patient Experience: Immediate results at the point of care; reduces the anxiety of the “9- to 18-month wait” for specialist answers.
Health Outcomes: Enables earlier lifestyle interventions and identifies candidates for DMTs (Lecanemab/Donanemab) when they are most effective.
Reducing Costs: Streamlines clinical trial recruitment (where 50-70% of costs go to screening) and prevents unnecessary, expensive imaging for low-risk patients.
Clinician Experience: A 3-minute test administered by medical assistants reduces the cognitive load and “time-pressure” on PCPs.
“This is the first demonstration of a true behavioral surrogate biomarker for Alzheimer’s disease,” said David Bates, PhD, CEO of Linus Health. “AI is detecting the earliest disruptions in brain function, likely several years before clinical symptoms.”
What’s Next?
As blood-based biomarkers become more accessible, the DCR’s role as a “pre-screener” will be critical for payers to control costs.
