A brain signature highly predictive of future progression to Alzheimer's dementia
share
Early prognosis of Alzheimer's dementia is hard. Mild cognitive impairment (MCI) typically precedes Alzheimer's dementia, yet only a fraction (30 MCI individuals will progress to dementia. Even when a prognosis of dementia is established using machine learning models and biomarkers, the fraction of MCI progressors remain limited (50 large clinical cohorts known for their heterogeneity, we propose here to identify only a subset of individuals who share a common brain signature highly predictive of oncoming dementia. This signature was discovered using a machine learning model in a reference public sample (ADNI), where the model was trained to identify patterns of brain atrophy and functional dysconnectivity commonly seen in patients suffering from dementia (N = 24), and not seen in cognitively normal individuals (N = 49). The model then recognized the same brain signature in 10 MCI individuals, out of N = 56, 90 within three years. This result is a marked improvement on the state-of-the-art in prognostic precision, while the brain signature still identified 47 MCI progressors (N = 19). We thus discovered a sizable MCI subpopulation with homogeneous brain abnormalities and highly predictable clinical trajectories, which may represent an excellent recruitment target for clinical trials at the prodromal stage of Alzheimer's disease.
READ FULL TEXT