As plasma-based Alzheimer biomarker testing moves from research labs into clinical practice, the gap between controlled trial performance and real-world accuracy is emerging as a serious concern — one that could translate into misdiagnosis, unnecessary treatment, and profound psychological harm for millions of patients.
The study examined real-world performance of the first FDA-cleared plasma test designed to detect Alzheimer's disease pathology through blood-based biomarkers, likely measuring phosphorylated tau (p-tau217) or amyloid-beta ratios. In controlled research settings, such tests have shown strong concordance with PET imaging and CSF analysis. However, in diverse clinical populations — which include patients with comorbid conditions, kidney disease, and non-Alzheimer cognitive disorders — specificity appears to drop meaningfully, generating false-positive rates that the research-grade validation studies did not fully anticipate. The analysis, published in JAMA Neurology and discussed in JAMA's Medical News section, underscores a recurring pattern: diagnostic technologies that perform well in curated cohorts often stumble when deployed across heterogeneous patient populations.
This finding carries weight beyond technical accuracy metrics. Blood-based Alzheimer tests are being positioned as a scalable gateway to anti-amyloid therapies like lecanemab and donanemab — treatments that carry real risks including amyloid-related imaging abnormalities (ARIA). A false-positive result could expose cognitively normal individuals to unnecessary PET confirmation costs, invasive CSF testing, or even eligibility for high-risk treatments. The broader landscape of biomarker-driven medicine is replete with cautionary examples where population-level test performance diverges sharply from clinical trial performance, particularly when pre-test probability is lower — as it would be in primary care settings where these tests are increasingly ordered. This study should be viewed as a calibration signal rather than a condemnation of plasma diagnostics; sequential testing algorithms and PET confirmation protocols remain important safeguards until real-world specificity data matures.