Cartesian vs radial MR-STAT: An efficiency and robustness study.

Area of Science:

• Magnetic Resonance Imaging
• Quantitative MRI
• Medical Physics

Background:

• MR Spin Tomography in Time-domain (MR-STAT) is a quantitative MRI technique.
• It estimates multiple quantitative parameters from a single scan using non-linear optimization.
• This study extends the MR-STAT framework to non-Cartesian trajectories.

Purpose of the Study:

• To compare Cartesian and radial MR-STAT in terms of time-efficiency and robustness.
• To evaluate performance in simulations, phantom, and in vivo measurements.
• To determine the optimal MR-STAT trajectory for clinical applications.

Main Methods:

• MR-STAT framework extended to non-Cartesian (radial) gradient trajectories.
• Comparison of Cartesian and radial MR-STAT using simulations.
• Validation with gel phantom and in vivo MRI measurements.

Main Results:

• Both Cartesian and radial MR-STAT demonstrated robustness against undersampling in simulations.
• Radial MR-STAT showed higher dynamic encoding power but lower spatial encoding power.
• Cartesian MR-STAT exhibited robustness against overfitting, unlike radial MR-STAT which showed artifacts in phantom and in vivo scans.
• Cartesian MR-STAT had higher overall time-efficiency, though radial MR-STAT was faster for T1 in white matter.

Conclusions:

• Cartesian MR-STAT is more robust against overfitting and artifacts, suggesting its preference for clinical settings.
• Radial MR-STAT may require further optimization to mitigate hardware-related artifacts.
• The choice between Cartesian and radial MR-STAT may depend on specific clinical needs and tissue types.