Joint Eigenvector Estimation from Mutually Anisotropic Tensors Improves Susceptibility Tensor Imaging of the Brain, Kidney, and Heart

    loading  Checking for direct PDF access through Ovid

Abstract

Purpose:

To develop a susceptibility-based MRI technique for probing microstructure and fiber architecture of magnetically anisotropic tissues—such as central nervous system white matter, renal tubules, and myocardial fibers—in three dimensions using susceptibility tensor imaging (STI) tools.

Theory and Methods:

STI can probe tissue microstructure, but is limited by reconstruction artifacts because of absent phase information outside the tissue and noise. STI accuracy may be improved by estimating a joint eigenvector from mutually anisotropic susceptibility and relaxation tensors. Gradient-recalled echo image data were simulated using a numerical phantom and acquired from the ex vivo mouse brain, kidney, and heart. Susceptibility tensor data were reconstructed using STI, regularized STI, and the proposed algorithm of mutually anisotropic and joint eigenvector STI (MAJESTI). Fiber map and tractography results from each technique were compared with diffusion tensor data.

Results:

MAJESTI reduced the estimated susceptibility tensor orientation error by 30% in the phantom, 36% in brain white matter, 40% in the inner medulla of the kidney, and 45% in myocardium. This improved the continuity and consistency of susceptibility-based fiber tractography in each tissue.

Conclusion:

MAJESTI estimation of the susceptibility tensors yields lower orientation errors for susceptibility-based fiber mapping and tractography in the intact brain, kidney, and heart.

Related Topics

    loading  Loading Related Articles