Psychiatry
Partner With Us
Advertisement
From the Journals

New Tool Estimates Amyloid Risk in Early Alzheimer’s

Cognitive performance–based score stratifies older adults by risk of cerebral amyloid deposition.

Conexiant

A brief computerized test combined with a newly developed scoring system may help stratify older adults by risk of early Alzheimer's disease pathology and potentially reduce reliance on positron emission tomography imaging. The Cognivue Amyloid Risk Measure, recently evaluated in a study published in Neurology and Therapy, estimates the likelihood of cerebral amyloid deposition using age and performance on three cognitive subtests: adaptive motor control, visual salience, and shape discrimination.

The Cognivue Amyloid Risk Measure (CARM) is integrated into the Cognivue Clarity system, a Food and Drug Administration–cleared computerized cognitive test, and generates a 4-point risk score. Its use in primary care settings may assist in identifying patients with Alzheimer's disease (AD)-related cognitive impairment, particularly in settings without access to advanced diagnostic tools.

"CARM is an added-value component to performing the Cognivue Clarity,” said lead study author James E. Galvin, MD, MPH, of the University of Miami Miller School of Medicine, in an interview with Conexiant News. “While the Clarity is used to screen for the likelihood of cognitive impairment, CARM provides the likelihood that impairment is due to [AD]. This could greatly facilitate referrals to neurology for treatment with amyloid-lowering therapies and the conduct of more expensive biomarkers such as positron emission tomography (PET) scans or lumbar punctures,” he added.

Development and Validation

The tool was evaluated in 887 individuals aged 60 to 85 years from the Bio-Hermes Study. Participants underwent Cognivue Clarity testing, amyloid PET imaging with 18F-florbetapir, and blood-based biomarker assessments. The participants were classified into six groups: cognitively normal controls (true controls), preclinical AD (pAD), mild cognitive impairment due to AD (MCI-AD), AD, MCI due to non-AD causes, and dementia due to non-AD causes.

CARM scores were derived using gradient boosting machine (GBM) regression, which outperformed support vector machines and ensemble models. Risk was categorized into four levels: CARM1 and CARM2 (low risk), and CARM3 and CARM4 (high risk).

“An individual presenting with cognitive impairment and CARM levels 3 or 4 would benefit from having a confirmatory biomarker toward the treatment or clinical trial path,” Dr. Galvin said. “Those with CARM levels 1 or 2 are less likely to have amyloid, but clinical judgment should guide whether further testing is warranted.”

Clinical Utility and Implications for Research

CARM scores were significantly associated with both PET-determined amyloid status and plasma biomarkers, including amyloid-beta 42/40, pTau181, and pTau217. Rates of amyloid positivity increased across score levels—19.9% in CARM1, 11.8% in CARM2, 25.7% in CARM3, and 42.6% in CARM4. Cognitive impairment prevalence also increased from 40% in CARM1 to 75% in CARM4.

When scores were grouped into low and high risk, CARM correctly identified 68.3% of amyloid-positive individuals (odds ratio = 3.67, 95% confidence interval = 2.76–4.89). Most cognitively normal, amyloid-negative participants were classified as low risk, while individuals with pAD, MCI-AD, or AD were more frequently classified as high risk.

“Another added value of the CARM is that for older adults with normal performance and a CARM suggesting the presence of amyloid, this could mean the presence of [pAD],” Dr. Galvin explained. “While no therapies are yet approved for the treatment of [pAD], there are a number of clinical trials that are testing agents.”

Integration Into Primary Care

Although not a diagnostic tool, CARM may be used as a prescreening method to inform decisions around confirmatory testing, treatment referrals, or potential trial enrollment. Dr. Galvin noted that it can help reduce unnecessary referrals and prioritize patients who may be more likely to benefit from amyloid-lowering therapy.

The researchers proposed a 2 × 2 matrix integrating CARM and overall Clarity scores to categorize patients into one of four groups:

  • True Controls (normal cognition, low amyloid)
  • pAD (normal cognition, high amyloid)
  • Non-AD MCI/dementia (impaired cognition, low amyloid)
  • MCI-AD/AD (impaired cognition, high amyloid).

This classification may inform decisions such as plasma biomarker testing (eg, pTau217), discussions about treatment eligibility, or continued monitoring.

“CARM should inform—but not replace—clinical judgment, especially when scores fall into intermediate ranges,” Dr. Galvin emphasized.

Limitations and Future Directions

Study limitations included its cross-sectional design, a limited neuropsychological testing battery, and the absence of structural imaging such as magnetic resonance imaging. The raw CARM score isn't directly interpretable on its own and should be considered within the clinical context. Although the study cohort was large and diverse, certain racial and ethnic populations were underrepresented.

Despite these limitations, the integration of CARM with a brief cognitive screen and its correlation with multiple biomarkers suggests potential utility in supporting early-stage AD assessment workflows.

Disclosures: Available in the original publication.

Source: Neurology and Therapy, Cognivue Press