Cartesian MaxGIRF: Model-based EPI reconstruction incorporating gradient nonlinearity and concomitant field effects

Area of Science:

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

Background:

• Lower field strength MRI scanners often suffer from EPI distortions caused by gradient nonlinearity and concomitant fields.
• Existing correction methods, relying on image-domain interpolation, can lead to undesirable spatial blurring, compromising image quality.
• These distortions are particularly problematic in scanners with large bore sizes, complex geometries, or strong gradient systems.

Purpose of the Study:

• To introduce a novel model-based EPI reconstruction framework, Cartesian MaxGIRF, designed to simultaneously compensate for EPI distortions.
• To avoid the spatial blurring typically introduced by conventional correction techniques.
• To address distortions arising from concomitant fields, gradient nonlinearity, and off-resonance effects during image reconstruction.

Main Methods:

• The Cartesian MaxGIRF framework was developed for model-based EPI reconstruction.
• Performance was evaluated against standard correction methods using phantom and in-vivo human brain datasets at 0.55T.
• Specific EPI artifacts, including parabolic shift and slice-dependent Nyquist ghost, were investigated and addressed using different EPI sequences (3D GRE-EPI).

Main Results:

• The proposed framework successfully mitigated parabolic shifts and slice-dependent Nyquist ghosts, artifacts induced by concomitant fields.
• Cartesian MaxGIRF demonstrated superior retention of image details compared to standard methods when correcting geometric distortions.
• Theoretical analysis confirmed the mitigation of parabolic shifts across various imaging scenarios.

Conclusions:

• The Cartesian MaxGIRF framework effectively mitigates EPI distortions from concomitant fields, gradient nonlinearity, and static off-resonance simultaneously.
• This approach is particularly beneficial for correcting artifacts caused by second-order concomitant fields in both symmetric and asymmetric gradient systems.
• The method offers an improved solution for enhancing image quality in EPI MRI without introducing spatial blurring.