We are aiming to greatly reduce the radioactive radiotracer dose administered to subjects during PET scanning. In this work we propose to leverage the perfect spatiotemporal correlation of hybrid PET/MRI scanning to synthesize diagnostic PET images from multiple MR images and a noisy PET image reconstructed from acquisitions with actual ultra-low-dose (as low as ~1% of the original) amyloid radiotracer injections, using trained deep neural networks. This technique can potentially increase the utility of hybrid amyloid PET/MR imaging and remove the limiting factors to large-scale clinical longitudinal PET/MRI studies.
This abstract and the presentation materials are available to members only; a login is required.