OUTCOME/OBJECTIVES

A spatial localization technique recently developed by our laboratory is described here. It is based on the sensitivity heterogeneity of multichannel coils and high-resolution anatomical imaging. Since sensitivity profiles of the coils contain spatial information of the detected MRS signal the new technique accelerates multi-region MRS data acquisition to obtain spectral information from regions with irregular boundaries.

PURPOSE

The SLIM (Spectral Localization by Imaging) technique pioneered by Hu et al (1) achieves rapid spatial localization of irregularly shaped regions using high resolution anatomical images and multiple k-space points to extract spectral information from compartments with relatively homogeneous signal distribution. Recently, we introduced the SPLASH (SPectral Localization Achieved by Sensitivity Heterogeneity) technique (2, 3) for performing MRS using a multi-element receiver coil. Similar to SLIM, SPLASH uses anatomical information extracted from high-resolution imaging to define irregularly shaped compartments. Different from SLIM, SPLASH uses inherent spatial heterogeneity of multiple receiver coil elements to resolve signals from different spatial regions. In SPLASH, both the amplitude and phase heterogeneity of the receive coils are used for extracting spectral and spatial information. With phased array coils, a large number of spectra are obtained in a single shot. This feature allows SPLASH to accelerate multi-region MRS data acquisition. The first goal of SPLASH is to achieve short scan time and high SNR efficiency by constructing compartmental spectra from data acquired with no or very few k-space points. The second goal of SPLASH is to achieve spatial resolutions comparable to conventional Fourier CSI but with much reduced partial volume effect and less cross contamination.

METHODS

To account for potential heterogeneous signal distribution within compartments defined by high-resolution imaging, each compartment was subdivided into multiple subcompartments and their spectra were reconstructed by Tikhonov regularization to impose smoothness of MRS signal distribution within each anatomical compartment. The spectrum of a given compartment was generated by combining the spectra of the subcompartments. The SPLASH method was validated using Monte Carlo simulations and phantom experiments. It has been applied to reconstructing in vivo spectra from irregularly shaped ischemic stroke and normal tissue compartments.

RESULTS

The theory of coil-sensitivity based MRS localization using SPLASH was developed and validated using phantom experiments, Monte Carlo simulations and in vivo MRS of stroke patients. In particular, we showed that spectra of irregularly shaped ischemic stroke and normal tissue compartments can be acquired in a single shot at 3 Tesla using an eight-element phased array coil. We also showed that Tikhonov regularization significantly reduces localization and metabolite quantification errors in cases where signal distribution within compartments is significantly heterogeneous.

DISCUSSION and CONCLUSION

SPLASH accelerates MRS data acquisition from multiple regions with irregular boundaries. Tikhonov Regularization addresses the intracompartment heterogeneity problem by subdividing the compartments into many smaller subcompartments and enforces intracompartment smoothness. A major advantage of this approach is that the inverse problem of reconstructing MRS spectra remains linear. The compartmental spectra can be reconstructed by the conventional generalized least square algorithm. In summary, SPLASH took advantage of the intrinsic spatial heterogeneity of multi-element phased array coils to achieve spatial localization. Like SLIM, structural information extracted from high-resolution anatomical imaging was utilized to define MRS compartments. In SPLASH, inherent spatial heterogeneity of multiple receiver coil elements was used along with optional k-space encoding to resolve signals from different compartments. The reconstructed compartmental spectra showed that SPLASH effectively reduced the partial volume effect and cross contamination compared to conventional Fourier CSI performed using the same multi-element receiver coil. Compared to SLIM, SPLASH further shortens scan time, increases sensitivity, and improves localization reliability.

Acknowledgements

The authors acknowledge support from NIMH Intramural Research Program, NIH.