Embedded Quantitative MRI T_1ρ Mapping Using Non-Linear Primal-Dual Proximal Splitting.

Area of Science:

• Medical Imaging
• Biophysics
• Computational Imaging

Background:

• Quantitative MRI (qMRI) offers objective, numerical tissue property assessments, reducing clinical subjectivity.
• Current qMRI methods, like T1ρ mapping, are time-consuming due to multiple measurements.
• Compressed sensing (CS) reconstruction techniques can accelerate qMRI by using undersampled data.

Purpose of the Study:

• To develop and evaluate a novel, direct, embedded reconstruction method for T1ρ mapping.
• To improve the efficiency and accuracy of quantitative MRI by reducing scan time.
• To enable more robust parameter map estimation, especially at higher acceleration factors.

Main Methods:

• A direct, embedded reconstruction approach utilizing a known signal model within a non-linear optimization framework.
• Reconstruction directly estimates the T1ρ parameter map, reducing the number of unknowns.
• Validation using simulated radially sampled data and experimental ex vivo mouse kidney data.

Main Results:

• The embedded reconstruction method demonstrated a 37-76% reduction in T1ρ Root Mean Square Error (RMSE) compared to standard CS reconstructions on simulated data.
• Performance improvements increased with higher undersampling (acceleration) factors.
• The proposed method outperformed reference techniques in experimental ex vivo measurements, showing enhanced robustness.

Conclusions:

• The embedded reconstruction method offers a significant advancement for accelerated quantitative T1ρ MRI.
• This approach enhances accuracy and robustness, particularly under accelerated acquisition conditions.
• The method holds promise for reducing MRI scan times while maintaining or improving diagnostic quality.