Meeting Banner
Abstract #3871

Diagnostic Ultra-low-dose 18F-PI-2620 Tau PET/MRI with Generative Adversarial Network-based Enhancement

Kevin T Chen1,2, Robel Tesfay3, Mary Ellen I Koran2, Jiahong Ouyang2, Sara Shams2, Tie Liang2, Mehdi Khalighi2, Elizabeth Mormino2, and Greg Zaharchuk2
1National Taiwan University, Taipei, Taiwan, 2Stanford University, Stanford, CA, United States, 3Meharry Medical College, Nashville, TN, United States


With the focal and lower uptake in tau PET imaging compared to other tracers such as amyloid, we aim to use multimodal simultaneous PET/MR imaging combined with training a generative adversarial network (GAN) to enhance ultra-low-dose 18F-PI-2620 tau PET images. We showed that the deep learning-enhanced images greatly reduced image noise as compared to the ultra-low-dose images, outperformed the ultra-low-dose images metrics-wise, and were able to be read clinically for regional uptake patterns of tau accumulation similarly as the full-dose images.

This abstract and the presentation materials are available to members only; a login is required.

Join Here