GPND-AI NULISA: A 15-Protein AI classifier for diagnosis and co-pathology profiling across neurodegenerative diseases
A blood test using artificial intelligence can now distinguish between four major forms of dementia with 92% accuracy, offering hope for faster, more precise diagnosis.
This study developed and validated a plasma-based AI classifier (GPND-AI) using a 15-protein NULISA panel that achieves 92.3% accuracy and AUC 0.955 for differential diagnosis of Alzheimer's, Parkinson's, frontotemporal dementia, and Lewy body dementia while also predicting co-existing pathologies. External validation against neuropathology-confirmed diagnoses at Banner Sun Health Research Institute supports translational readiness of this approach for improving clinical diagnostic accuracy in neurodegenerative disease.
What the study was
- Study design
- Validation study with external cohort
- Population
- Patients with Alzheimer's disease, Parkinson's disease, frontotemporal dementia, dementia with Lewy bodies, and healthy controls
- Category
- Diagnostics
- Maturity
- Validated
- Journal
- Alzheimer's & Dementia
Why it surfaced
Externally validated plasma proteomic AI classifier for differential diagnosis of four major neurodegenerative diseases simultaneously; addresses major unmet need (accurate neurodegenerative disease differentiation is currently challenging in clinical practice); high AUC 0.955 and validated against neuropathology gold-standard; plasma-based assay is near-term implementable.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.