O2‐05‐01: A DATA‐DRIVEN MODEL OF BIOMARKER CHANGES IN SPORADIC ALZHEIMER'S DISEASE

Abstract

Determining the sequence in which Alzheimer's disease (AD) biomarkers become abnormal would provide important insights into disease biology and a mechanism for disease staging. Here we implement a probabilistic event-based model (EBM) (Fonteijn et al. 2012) to determine the sequence of biomarker abnormality in sporadic AD, characterise the uncertainty in the ordering, and provide a natural patient staging system. Unlike previous attempts to construct such a model, our method does not rely on a-priori clinical diagnostic information, or explicit biomarker cut-points, and it allows for a proportion of cases and controls to be misdiagnosed. We included 285 ADNI subjects (92 cognitively normal (CN), 129 mild cognitive impairment (MCI), 64 AD) with measurements of 14 AD-related biomarkers including CSF A β 1-42, p-tau, t-tau, whole brain and regional MRI brain volumes and rates of atrophy over one year, and three cognitive test scores. We used the EBM to determine the sequence of biomarker abnormality and its uncertainty in various population subgroups. We used patient stages assigned by the EBM to discriminate CN from AD subjects, and predict conversion from MCI to AD and CN to MCI. The model predicts (Figure 1) that CSF measures become abnormal first, followed by rates of atrophy, then cognitive test scores, and finally regional brain volumes. In amyloid positive (CSF A β 1-42<192 pg/ml) or APOE positive subjects, the model predicts with high confidence that the CSF biomarkers b…