Enhanced monitoring of Alzheimer's disease brain atrophy using composite value ratios of volumes

Brain atrophy is a natural consequence of ageing but can be accelerated by neurodegenerative diseases such as Alzheimer's disease. We apply an existing algorithm to identify new biomarkers that better track volumetric changes over time, i.e. atrophy. These new biomarkers are the volumetric ratio of two composite regions of interest, identified by an algorithm optimized to enhance the monitoring of Alzheimer's disease progression. The algorithm prioritizes biomarkers with less noisy trajectories (quantified by lower sample size estimates for clinical trials) and that capture disease signal by showing a greater rate of change in amyloid-positive versus amyloid-negative individuals, ensuring the biomarker effectively reflects Alzheimer's pathology. Data from 1381 individuals from the Alzheimer's Disease Neuroimaging Initiative database having multiple MRI scans were analysed. The new biomarkers outperformed traditional volumetric measures (whole brain, hippocampus and ventricles) across all metrics. This improvement was particularly pronounced in cognitively impaired individuals, where atrophy is more severe. Among the traditional measures, ventricular volume had the best performance. Results suggest that ratios of regional brain volumes could enhance disease progression tracking, as has been shown in other modalities like fluid biomarker ratios and standardized uptake value ratios from positron emission tomography.