Compressed sensing MRI combined with SENSE in partial k-space

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Medical Imaging Physics
• Radiology

Background:

• Accelerated magnetic resonance (MR) acquisition is crucial for reducing scan times.
• Compressed sensing (CS), parallel imaging, and partial Fourier (PF) acquisition are key acceleration techniques.
• Combining these methods offers flexibility in balancing scan speed and image quality.

Purpose of the Study:

• To evaluate the combined use of compressed sensing with SENSE (CS-SENSE) and partial Fourier (PF) acquisition for accelerating anatomical MRIs of the human brain.
• To demonstrate the potential of this combined undersampling strategy for high-resolution T1-weighted brain imaging.

Main Methods:

• Simulations were performed on full k-space data.
• A 3D fast spoiled gradient recalled echo (FSPGR) sequence on a GE 3T scanner was modified for undersampling.
• Undersampling schemes included CS, CS-SENSE, and partial Fourier CS-SENSE.
• Partially sampled k-space data were acquired and reconstructed into brain images.

Main Results:

• CS-SENSE combined with PF sampling yielded superior reconstructed images compared to CS alone or CS-SENSE without PF at the same acceleration factor.
• The combination allows for further acceleration or improved image quality at a constant acceleration rate.

Conclusions:

• Combining partial Fourier sampling with CS-SENSE is an effective strategy for accelerating high-resolution brain MRI acquisition.
• This approach provides enhanced image quality and/or scan speed, offering greater flexibility in clinical MRI protocols.