Establishment and validation of an Alzheimer's disease diagnostic model on the basis of exhaled volatile organic compound characteristics
Breath tests combined with machine learning identified Alzheimer's with high accuracy by detecting metabolic changes, offering a noninvasive screening method needing human validation.
In 285 participants, exhaled volatile organic compound profiling combined with machine learning achieved 93% accuracy for AD diagnosis in internal validation and 75% in an independent external cohort, with 90% accuracy for distinguishing AD from non-AD dementias. Metabolic pathway analysis identified butyrate, pyruvate, and glycolysis disruptions as AD-specific VOC signatures.
What the study was
- Study design
- Validation study (internal + external cohorts)
- Population
- AD patients (dementia + MCI), non-AD dementia, and healthy controls
- Sample size
- 285
- Category
- Diagnostics
- Maturity
- Exploratory
- Journal
- Translational Psychiatry
Why it surfaced
Exhaled breath non-invasive AD diagnostics with external validation; external AUC 0.75 is modest; promising non-invasive approach for population screening.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.