Iterative 3D projection reconstruction of (23) Na data with an (1) H MRI constraint

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Image Reconstruction
• Medical Physics

Background:

• Non-proton MRI often suffers from low signal-to-noise ratio (SNR) and artifacts.
• Iterative reconstruction methods can improve image quality but may require prior information.
• Incorporating anatomical knowledge can further enhance reconstruction accuracy.

Purpose of the Study:

• To enhance signal-to-noise ratio (SNR) and reduce artifacts in non-proton MRI.
• To integrate a priori information from (1)H MR data into iterative reconstruction.
• To develop and evaluate a novel Binary Mask and Total Variation (BM&TV) algorithm.

Main Methods:

• Developed an iterative reconstruction algorithm for 3D projection reconstruction (3DPR).
• Combined prior anatomical knowledge using a binary mask (BM) with image sparsity under a total variation (TV) constraint.
• Evaluated the BM&TV method using simulations and MR measurements in human volunteers.

Main Results:

• BM&TV reduced artifact level by 20% in simulated data while preserving structures.
• Achieved up to 100% SNR gain in simulated data and 29±7% in human brain (23)Na MRI.
• Demonstrated superior anatomical structure preservation compared to standard iterative reconstruction (14% vs 66% contrast loss).

Conclusions:

• The BM&TV algorithm effectively increases SNR and reduces artifacts in non-proton MRI.
• This method outperforms traditional gridding reconstruction and unspecific TV-regularized iterative methods.
• Incorporating anatomical constraints significantly improves the performance of iterative MRI reconstruction.