Trajectory Auto-Corrected image reconstruction.

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Medical Imaging
• Image Reconstruction

Background:

• Non-Cartesian MRI sequences are efficient but sensitive to k-space trajectory errors.
• These errors cause image distortions like blurring and streaking, degrading diagnostic quality.
• Accurate k-space trajectory estimation typically requires gradient calibration or direct measurement.

Purpose of the Study:

• To develop a method for estimating k-space trajectory errors in non-Cartesian MRI.
• To reconstruct distortion-free images from corrected trajectories.
• To eliminate the need for explicit trajectory measurements or gradient calibrations.

Main Methods:

• Introduced Trajectory Auto-Corrected image Reconstruction (TACR).
• TACR jointly estimates k-space trajectory errors and images using SENSE and SPIRiT principles.
• Exploits data redundancy from parallel imaging and k-space center oversampling.
• Trajectory errors modeled as weighted sums of basis functions, optimized via gradient descent.

Main Results:

• TACR successfully reconstructed images from golden angle radial, center-out radial, and spiral acquisitions.
• TACR reconstructions showed significantly reduced blurring and streaking compared to nominal trajectories.
• Image quality was comparable to reconstructions using measured k-space trajectories for radial and spiral data.
• Demonstrated reductions in reconstruction cost function and improvements in image gradient metrics.

Conclusions:

• TACR enables accurate non-Cartesian MRI reconstruction without trajectory measurement or gradient calibration.
• The method effectively corrects for k-space trajectory errors, improving image fidelity.
• This approach simplifies the workflow for non-Cartesian MRI acquisition and reconstruction.