Least Squares Reconstruction for Sparse Spectra: Application to Hyperpolarized 13C Imaging
Levin Y, Yen Y, Spielman D, Hurd R, Mayer D
Stanford University School of Medicine, Stanford University School of Engineering
Hyperpolarization of 13C allows for imaging of an injected compound and its downstream products. A single-interleaf spiral CSI pulse sequence that allows for aliasing of peaks from the broad but sparse 13C spectrum is explored here. A k-space based least-squares algorithm using a priori knowledge of peak characteristics for reconstruction of metabolic images is proposed. Using simulation and phantom measurements, the technique compares favorably to conventional 3-D spectroscopic reconstruction and allows optimization of acquisition parameters such as spectral width and number of echoes. The work implies that very rapid (<100 ms) spectroscopic imaging of hyperpolarized 13C can be performed.