Rapid flow-artifact-free high-resolution T₂ mapping via multi-shot multiple overlapping-echo detachment imaging.

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Medical Physics
• Biomedical Engineering

Background:

• T2 mapping is crucial for MRI-based tissue characterization.
• Existing methods often face trade-offs between resolution, acquisition time, and scan volume.
• Pulsatile cerebrospinal fluid (CSF) can introduce phase variations, complicating segmented k-space acquisitions.

Purpose of the Study:

• To develop a T2 mapping method with submillimeter resolution and ~1-minute acquisition time.
• To enable large volume coverage without console or coil modifications.
• To implement a deep learning approach for correcting inter-shot phase variations caused by CSF pulsation.

Main Methods:

• Integration of a multi-shot (msh-) acquisition scheme into multiple overlapping-echo detachment imaging (MOLED).
• Application of deep learning for correcting discontinuous phase jumps by leveraging signal and artifact incoherence.
• Validation of the method using phantoms (3T and 7T) and human subjects (3T).

Main Results:

• Phantom and human T2 mapping results showed low mean absolute errors (1.11/0.89 ms) compared to spin-echo.
• msh-MOLED demonstrated good structural depiction compared to turbo spin-echo.
• Successful elimination of CSF-induced phase errors without autocalibration, navigators, gating, or lengthy postprocessing.

Conclusions:

• A T2 mapping method with submillimeter resolution and high clinical practicality has been developed and validated.
• The navigator-free, calibrationless approach effectively addresses inter-shot phase variations from pulsatile CSF.
• This technique maintains quantification accuracy and rapid acquisition, suggesting potential for removing trajectory-related ghost artifacts via spatial encoding and inpainting.