Abstract #0905

PET by MRI: Glucose Imaging by 13C MRS by Noise Suppression through Tensor Decomposition Rank Reduction

Jeffrey R Brender¹, Shun Kishimoto¹, Hellmut Merkle², Galen Reed³, Ralph E Hurd⁴, Albert P Chen⁵, Jan Henrik Ardenkjaer-Larsen⁶, Jeeva Munasinghe², Keita Saito¹, Tomohiro Seki¹, Nobu Oshima¹, Kazu Yamamoto¹, James Mitchell¹, and Murali C. Krishna¹

¹Radiation Biology, NIH/NCI, Washington DC, DC, United States, ²NIH/NINDS, Bethesda, MD, United States, ³GE HealthCare, Dallas, TX, United States, ⁴Applied Science Laboratory, GE HealthCare, Menlo Park, CA, United States, ⁵GE HealthCare, Toronto, ON, Canada, ⁶Technical University of Denmark, Lyngby, Denmark

We show here a new method of imaging glucose metabolism in vivo by ¹³C MRS without DNP that relies on a simple but efficient postprocessing procedure using the higher dimensional analogue of SVD, tensor decomposition. Using this procedure, without sacrificing accuracy, we achieve an order of magnitude increase in SNR in both DNP and non-hyperpolarized single voxel experiments. In CSI imaging experiments, we see an approximately 30 fold increase in SNR, enough that the glucose to lactate conversions can be imaged with a time resolution of 12 seconds and an overall spatial resolution that compares favorably to 18F-FDG PET.

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

Join Here