A dynamic approach for MR T2-weighted pelvic imaging.

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging (MRI)
• Deep Learning

Background:

• Standard T2-weighted pelvic MRI protocols face challenges with motion artifacts and blurring due to peristalsis.
• Current methods require patient preparation with antiperistaltic agents, causing discomfort.
• Peristalsis-induced motion significantly impacts diagnostic accuracy in pelvic MRI.

Purpose of the Study:

• To introduce a novel dynamic MRI approach for T2-weighted pelvic imaging.
• To address peristalsis-induced motion artifacts without requiring patient preparation.
• To develop a motion-reconstruction strategy for improved pelvic MRI quality.

Main Methods:

• A rapid dynamic data acquisition strategy with a complementary sampling trajectory was employed.
• Highly undersampled, motion-resistant data sampling was achieved.
• An unrolling method based on a deep equilibrium model was used for image reconstruction from dynamic k-space data.
• The fix-point convergence of the equilibrium model ensured reconstruction stability.

Main Results:

• The dynamic approach demonstrated superior performance in reducing motion artifacts compared to standard static imaging.
• Accurate depiction of structural details was achieved in both retrospective and prospective data.
• The method effectively reduced blurring caused by involuntary patient motion.

Conclusions:

• The proposed dynamic approach effectively captures motion states through dynamic acquisition and deep learning reconstruction.
• This method addresses motion-related challenges in pelvic MRI, offering a more comfortable and accurate diagnostic tool.
• The technique transforms pelvic MRI from motion prevention to motion reconstruction, enhancing diagnostic capabilities.