Fast magnetic resonance spectroscopic imaging using echo-time optimization

In recent decades, magnetic resonance spectroscopic imaging (MRSI) has become an influential technique to provide information about metabolite distributions. However, long acquisition time is required in standard MRSI to achieve satisfactory resolution in both spatial and spectral dimensions. This limitation is a key challenge for MRSI applications. To overcome this problem, we propose to implement an imaging protocol that uses echo-time optimization in the acquisition process. An efficient sequential backward selection (SBS) is used to optimize the echo-time selection in the context of echo planer imaging (EPI). To investigate this approach with a simpler acquisition method, we applied the proposed selection method to standard chemical shift imaging (CSI), using experimental phantom data. This acquisition method simplifies the reconstruction process and avoids Nyquist ghost artifacts as a means of testing the potential of optimized echo-time EPI. Comparisons are made in acquisition data volume and reconstructed spatial and spectral distribution from the proposed technique, standard MRSI and equally spaced echo-time selection. Results show that our method has the potential to greatly reduce the acquisition time without sacrificing spatial and spectral accuracy.