Purpose: To determine the imaging sequence parameters that optimize SNR for spectroscopic imaging of hyperpolarized (HP) [1-Â¹Â³C]pyruvate.
Methods: Simulation experiments were conducted in Matlab to compare the performance of spatial encoding with both symmetric and flyback EPI readouts, and spectral encoding with both IDEAL and spectral-spatial excitations. Spatial encoding simulations calculated the relative SNRs for Â¹Â³C-EPI readouts achievable on a small-animal MRI (maximum gradient of 660 mT/m and maximum slew rate of 4570 T/m/s) for various readout bandwidths and T2* relaxation times. Spectral encoding comparisons were made by generating raw k-space data at 7T from a numerical phantom containing the six chemical shifts typically present in HP pyruvate imaging studies and reconstructing image data for various flip angles, T2* values and levels of complex noise in k-space. IDEAL data were simulated for eight excitations at equally spaced TEs, and spectral-spatial data were normalized to provide the signal level resulting from eight RF excitations at a given IDEAL flip angle (i.e. equivalent effects on longitudinal magnetization for both methods). Precise TEs for each spectral encoding strategy corresponded to those realizable for that method within the gradient system limitations listed above.
Results: Symmetric EPI readouts consistently provide better SNR than flyback trajectories due to more efficient sampling of k-space. The readout bandwidth that maximizes SNR decreases with increasing T2* due to the reduction of relaxation effects in the resulting images. IDEAL provides image SNRs superior to spectral-spatial pulses for short T2*s and excitation angles <40 deg due primarily to the shorter TE achievable with broadband IDEAL excitations.
Conclusion: Selection of the spectroscopic imaging approach for HP MRI entails choices of methods and parameters that have significant impacts on image SNR, however these effects are not immediately clear. These results demonstrate the conditions under which SNR is optimized for imaging of HP [1-Â¹Â³C]pyruvate.
Funding Support, Disclosures, and Conflict of Interest: This work was supported in part by R01-CA211150, R01-DK105346, and P30-CA016672.