PrecivityAD2 APS2 Shows 91% Accuracy vs Amyloid PET

Key Takeaways

• In an independent ADNI cohort of cognitively impaired participants, APS2 showed high concordance with amyloid PET; the authors said these findings support use in symptomatic patients evaluated in clinical care.
• More than half of participants had a positive APS2 result, showing that amyloid-positive classifications were common in this validation cohort.
• Using the previously validated APS2 cut point, overall accuracy was high, and sensitivity and specificity were at or above 90%.
• The authors reported that the prespecified algorithm and cut point were repeatable, reliable, robust, accurate, and generalizable in this intended clinical context.

Evaluating cognitive impairment increasingly requires deciding when brain amyloid status is relevant to diagnosis and downstream management discussions. Blood-based biomarker tools are entering that pathway as a less invasive option than conventional confirmatory testing, but their intended use depends on performance in symptomatic patients rather than broad screening settings. PrecivityAD2 is one such blood-based test, and investigators assessed its Amyloid Probability Score 2 result in cognitively impaired patients in an independent clinical cohort.

Amyloid Probability Score 2 (APS2), the algorithmic result of the PrecivityAD2 blood test, had previously been trained and validated in two cohorts of cognitively impaired individuals and was already in clinical use. Investigators further tested the prespecified APS2 algorithm and cut point in an independent cohort of cognitively impaired individuals, using amyloid positron emission tomography (PET) as the reference standard for brain beta-amyloid pathology. The analysis re-evaluated the existing threshold rather than deriving a new one.

In the npj Dementia validation study of PrecivityAD2, concordance with amyloid PET was high, with an AUC-ROC of 0.95 (95% CI 0.93-0.98), and 54% of participants had a positive APS2 result. The blood-based score tracked closely with PET status in this cohort.

Using the previously validated APS2 cut point, overall accuracy was 91% (95% CI 86-94%), sensitivity was 90% (95% CI 83-94%), and specificity was 92% (95% CI 84-96%). The prespecified threshold retained strong classification performance in the independent cohort.

These findings support APS2 performance in cognitively impaired individuals similar to those studied in the ADNI validation cohort, rather than definitively establishing performance across all routine clinical-care settings, and they should not be extended to asymptomatic population screening. Because amyloid PET served as the reference standard, the results establish PET-referenced diagnostic performance for APS2 in this population rather than universal replacement of PET across care settings. This validation adds intended-use context without broadening the reported use case.

The authors concluded that the prespecified APS2 algorithm and cut point were repeatable, reliable, robust, accurate, and generalizable for identifying brain beta-amyloid pathology in cognitively impaired patients undergoing evaluation in clinical care.

Clinician Questions

Which patients do these PrecivityAD2 APS2 validation findings apply to?

The reported population was cognitively impaired individuals described as patients being evaluated in a clinical care setting, so the validation is bounded to symptomatic assessment rather than screening of asymptomatic people.

Did this validation study test a new APS2 threshold or re-evaluate the existing cut point?

Investigators re-tested the prespecified, previously validated APS2 algorithm and cut point in an independent cohort instead of deriving a new threshold, which is what makes this an independent validation study.

What does this PrecivityAD2 analysis establish about amyloid PET and what does it leave unresolved?

Because amyloid PET was the reference standard for brain beta-amyloid pathology, the study establishes PET-referenced performance of PrecivityAD2 APS2 in cognitively impaired patients but does not by itself show that blood testing can replace PET in all settings.