High-Resolution Self-Gated Dynamic Abdominal MRI Using Manifold Alignment.

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging (MRI)
• Computational Imaging

Background:

• Abdominal MRI often suffers from motion artifacts due to respiration.
• Current self-gating methods may not fully capture respiratory motion, limiting image quality.
• High spatial and temporal resolution imaging is crucial for accurate abdominal diagnostics.

Purpose of the Study:

• To develop a novel retrospective self-gating method for free-breathing abdominal MRI.
• To improve motion representation and image reconstruction accuracy using manifold alignment (MA).
• To achieve high spatial and temporal resolution in abdominal MRI sequences.

Main Methods:

• A retrospective self-gating method based on manifold alignment (MA) was developed.
• Radial golden-angle acquisition trajectory was used to extract multidimensional self-gating signals from k-space data.
• MA aligned low-dimensional manifolds of k-space profiles to represent respiratory positions accurately for image reconstruction.

Main Results:

• The MA-based method demonstrated high correlation with ground truth on synthetic data.
• Evaluated on in vivo data, the method significantly improved quantitative measurements compared to center of k-space gating.
• Improved organ sharpness and image gradient entropy were observed in reconstructed abdominal MRI sequences.

Conclusions:

• The novel MA-based self-gating method enables accurate motion representation and reconstruction of free-breathing abdominal MRI.
• This approach significantly enhances image quality and quantitative measurements.
• The method offers a promising advancement for high-resolution abdominal MRI acquisition.