Accelerated phosphorus magnetic resonance spectroscopic imaging using compressed sensing method

The purpose of this simulation study is to assess the performance of accelerated phosphorus magnetic resonance spectroscopic imaging (31P MRSI) using latest fast MR imaging technique, called compressed sensing. A volunteer was scanned at a 3T MR scanner (Philips Medical Systems, Best, Netherlands) using a 31P surface coil. Acquired 31P spectra dataset was quantified using AMARES tool in jMRUI. Healthy and tumor datasets were created in MATLAB (The Mathworks Inc., Natick, MA, USA) based on metabolite peak intensities and ratios of 31P spectra dataset. A 2D 32x32 sized phosphorus spectra dataset which contains healthy and tumor (10x10 voxels on the left top side) regions were simulated. A randomly undersampled pattern which reduces k-space data along x by a reduction factor of 2.23 was implemented to k-space of 2D phosphorus spectra dataset. This undersampled k-space was inverse Fourier transformed along ky and then kx-f data was reconstructed for each y point using Sparse MRI software package that uses compressed sensing (CS) method. Results of our simulation study showed that compressed sensing applied datasets had higher signal to noise (SNR) ratio than the original datasets due to denoising effects of compressed sensing reconstruction technique. However, peak ratios were similar between both compressed sensing and original datasets. This study indicated that compressed sensing reconstruction could be used for faster phosphorus magnetic resonance spectroscopic imaging.