Predicting Amyloid-beta Positivity in Cognitively Intact Adults Using Multi-contrast Three-dimensional MRI and Deep Learning (P10-13.002)

Key Points

• The aim is to develop and test a model predicting CNS amyloid-β positivity using MRI data.
• Developed a pretrained gated fusion model
• Tested the model on two distinctive datasets
• Utilized multi-contrast three-dimensional MRI scans
• The model accurately predicts amyloid-beta positivity in cognitively normal individuals
• Demonstrated robustness across two different datasets
• Showed potential for enhancing early detection of cognitive decline