A predictor-corrector phase unwrapping algorithm for temporally undersampled gradient-echo MRI.

Area of Science:

• Magnetic Resonance Imaging
• Image Processing
• Medical Physics

Background:

• Gradient Recalled Echo (GRE) phase imaging is crucial in MRI.
• Temporal undersampling and nonlinearities complicate GRE phase unwrapping.
• Accurate phase unwrapping is essential for quantitative MRI.

Purpose of the Study:

• To develop a robust method for unwrapping temporally undersampled and nonlinear GRE phase data.
• To improve the accuracy and spatial continuity of phase unwrapping in MRI.
• To evaluate the proposed method across various data types and field strengths.

Main Methods:

• A spatio-temporal extension of the predictor-corrector unwrapping (PCU) algorithm was developed.
• The method performs sequential one-step prediction and correction of echo phase.
• Evaluation involved numerical, physical phantoms, and in vivo brain data at 3T and 9.4T.

Main Results:

• The PCU algorithm demonstrated significant reduction in unwrapping errors compared to state-of-the-art methods, especially at higher echoes.
• Spatially smooth phase images were generated for in vivo data, eliminating the need for additional spatial unwrapping steps.
• PCU showed superior performance over the iVENyS algorithm at higher echoes.

Conclusions:

• The PCU algorithm is a robust solution for phase unwrapping in temporally undersampled and nonlinear GRE MRI.
• It is particularly effective in high-field MRI environments.
• The method enhances the reliability of quantitative MRI analyses.