Nonlinear parameter estimation with physics-constrained spectral-spatial priors for highly accelerated chemical...

Area of Science:

• Magnetic Resonance Imaging
• Biophysics
• Medical Imaging

Background:

• Chemical Exchange Saturation Transfer (CEST) MRI is crucial for detecting subtle biological changes.
• Current CEST MRI methods often require long acquisition times or suffer from artifacts due to incomplete data.
• Robust spectral analysis is essential for accurate CEST MRI.

Purpose of the Study:

• To develop a nonlinear, model-based parameter estimation method for robust spectral analysis in highly accelerated CEST MRI.
• To enable direct parameter estimation from incomplete k-w space measurements.
• To improve the delineation of tumor-specific CEST maps.

Main Methods:

• Incorporated a CEST-specific, separable nonlinear model derived from the Bloch-McConnel equation into a CEST MRI measurement model.
• Formulated signal drop using a nonlinear spectral prior.
• Estimated parameters directly from incomplete k-w space data using constrained optimization with physics-constrained and sparsity priors.

Main Results:

• The proposed method produced clearer delineation of tumor-specific CEST maps compared to conventional methods.
• The method effectively reduced artifacts and noise in the resulting CEST maps.
• Demonstrated feasibility for CEST MRI with highly incomplete measurements.

Conclusions:

• The developed nonlinear, model-based method enables robust spectral analysis in highly accelerated CEST MRI.
• This approach allows for high-resolution whole-brain CEST MRI in clinically reasonable imaging times.
• The method overcomes limitations of conventional techniques by utilizing incomplete k-w space data effectively.