Predicting Amyloid Burden Using a Masked Multimodal‐Multitask Deep Learning Framework with Latent Diffusion‐based Synthetic PET

By reducing reliance on cost-intensive PET scans and handling missing data more effectively, our framework could substantially broaden access to early AD screening in diverse clinical settings. Future work will involve testing in larger, more varied cohorts, applying different tracers, and further evaluate its real-world impact on AD management.