Velocity reconstruction with nonconvex optimization for low-velocity-encoding phase-contrast MRI

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging (MRI)
• Biomedical Engineering

Background:

• Phase-contrast MRI is crucial for quantifying blood flow velocity.
• Low-velocity-encoding (V_enc) acquisitions are prone to errors like intravoxel dephasing and velocity aliasing.
• Conventional reconstruction methods struggle with these challenges, limiting accuracy.

Purpose of the Study:

• To introduce and demonstrate a novel nonconvex optimization method for reconstructing velocity data.
• To address limitations in low-V_enc phase-contrast MRI data reconstruction.
• To improve the accuracy and robustness of velocity measurements.

Main Methods:

• Velocity reconstruction formulated as a nonconvex optimization problem.
• Incorporated weighting for intravoxel dephasing and Laplacian-based regularization for velocity aliasing.
• Tested with dual-V_enc and multidirectional high-moment encoding schemes in simulations, flow phantoms, and healthy volunteers.

Main Results:

• The nonconvex optimization method significantly reduced velocity reconstruction error compared to conventional methods in simulations and phantoms.
• Qualitative assessments showed decreased apparent noise and fewer phase wraps in velocity fields across all experiments.
• No additional velocity bias was observed in healthy volunteers using the proposed method.

Conclusions:

• The proposed nonconvex optimization method effectively reconstructs velocity data from low-V_enc acquisitions.
• It reduces phase aliasing and velocity errors from intravoxel dephasing, enabling more robust MRI acquisitions.
• Allows for V_enc to be lowered, increasing the velocity-to-noise ratio and improving diagnostic utility.