Coil profile estimation strategies for parallel imaging with hyperpolarized ¹³ C MRI

Area of Science:

• Magnetic Resonance Imaging
• Metabolic Imaging
• Biophysics

Background:

• Hyperpolarized 13C MRI enables real-time metabolic imaging.
• Parallel imaging accelerates data acquisition but requires accurate coil profile estimation.
• Current coil calibration methods can be suboptimal for hyperpolarized 13C MRI.

Purpose of the Study:

• To compare auto-calibration, phantom calibration, and theoretic coil profile estimation for hyperpolarized 13C MRI.
• To evaluate the impact of different coil profile estimation methods on parallel imaging reconstruction.
• To assess the feasibility of these methods for in vivo metabolic imaging.

Main Methods:

• Investigated three coil profile estimation approaches: auto-calibration, phantom calibration, and theoretic calibration.
• Utilized 3D stack-of-spirals sequence with Cartesian undersampling and spectral-spatial excitation.
• Performed reconstructions using conjugate gradient SENSE and 3D gridding with inhomogeneity correction.
• Compared methods via simulations, phantom experiments, and an in vivo pig kidney study at 3T.

Main Results:

• Theoretic and phantom calibrated approaches yielded superior structural similarity in simulations and higher image quality in phantom experiments compared to auto-calibration.
• Accelerated acquisitions with improved temporal resolution enhanced in vivo mapping of pyruvate uptake and lactate conversion.
• Theoretic coil profiles provided the best practical results and image quality.

Conclusions:

• The theoretic coil profile estimation method is robust, showing minimal load effects on sensitivity profiles at 3T.
• Theoretic or phantom calibrated parallel imaging enables accelerated 3D volumetric reconstruction with adequate resolution for abdominal organ metabolism mapping.
• This approach facilitates clinical translation of parallel imaging in hyperpolarized 13C MR.