Machine Learning Models Classify Tau Pathology in Amyloid-Positive Cohorts
Personalized briefing
Discovery of the day · Neurology
Classification of tau status with machine learning models in amyloid‐positive cohorts
Dear Damien Boorman, this is your personalized scientific intelligence briefing — curated for your work in Neurology.
Key finding
Medicine · Neurology
Discovery of the day
Machine learning models trained on MRI and amyloid PET features successfully classify tau positivity in amyloid-positive cohorts, achieving an AUC of 0.92 in both internal and external validation. Logistic regression demonstrated the best performance, with a combined external validation yielding 85% accuracy, 83% sensitivity, and 85% specificity across the ADNI, OASIS-3, and SCAN datasets. Crucially, subjects with mild cognitive impairment and predicted tau positivity progressed to Alzheimer’s disease at a significantly faster pace (p < 10⁻⁶), suggesting this non-invasive approach could serve as a practical surrogate biomarker for tau pathology in clinical settings, addressing the cost and accessibility limitations of tau PET imaging.
Novelty
91%
Rigor
85%
Significance
90%
Validity
84%
Clarity
87%
Advertisement
GroupRFQ — join group deals on lab & medical equipment. Pool orders, unlock volume pricing. See open deals →
Your briefing is personalized based on your selected fields, keywords, and research interests.
