Undersampling in Magnetic Resonance Metabolic Imaging Using Prior Anatomical Knowledge

Studying the metabolic pathways is important as insufficient energy supply to the cardiac muscle has a major role in heart failure. Dysfunctional substrate metabolism eventually leads to an impaired contractile function typical for heart failure. For assessing cardiac metabolism health, hyperpolarized magnetic resonance spectroscopy is used herein, for imaging pyruvate, lactate and bicarbonate metabolites. Due to the inherently short lifetime of the hyperpolarized signal, the use techniques to acquire dynamic images in a short amount of time are necessary. To this end, we use Cartesian undersampling techniques in conjunction with a spectral localization algorithm (SLIM) based on prior anatomical knowledge. Successful image reconstructions were obtained using undersampling factors of up to three, with minor degradation in image quality in simulations (11.8%) and preliminary in-vivo case (24.6%).