Exploring a Cost-Efficient Model for Predicting Cerebral Aβ Burden Using MRI and Neuropsychological Markers in the ADNI-2 Cohort.

Many studies have focused on the early detection of Alzheimer's disease (AD). Cerebral amyloid beta (Aβ) is a hallmark of AD and can be observed in vivo via positron emission tomography imaging using an amyloid tracer or cerebrospinal fluid assessment. However, these methods are expensive. The current study aimed to identify and compare the ability of magnetic resonance imaging (MRI) markers and neuropsychological markers to predict cerebral Aβ status in an AD cohort using machine learning (ML) approaches. The prediction ability of candidate markers for cerebral Aβ status was examined by analyzing 724 participants from the ADNI-2 cohort. Demographic variables, structural MRI markers, and neuropsychological test scores were used as input in several ML algorithms to predict cerebral Aβ positivity. Out of five combinations of candidate markers, neuropsychological markers with demographics showed the most cost-efficient result. The selected model could distinguish abnormal levels of Aβ with a prediction ability of 0.85, which is the same as that for MRI-based models. In this study, we identified the prediction ability of MRI markers using ML approaches and showed that the neuropsychological model with demographics can predict Aβ positivity, suggesting a more cost-efficient method for detecting cerebral Aβ status compared to MRI markers.