Title:A brain signature highly predictive of future progression to Alzheimer's dementia

Abstract:Early prognosis of Alzheimer's dementia is hard. Mild cognitive impairment (MCI) typically precedes Alzheimer's dementia, yet only a fraction (30%-50%) of 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%-75%). Instead of aiming at precise diagnosis in 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% of which progressed to dementia 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% of all 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.