Accelerating the estimation of 3D spatially resolved T₂ distributions

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Biomedical Engineering
• Image Reconstruction

Background:

• Quantitative 3D spatially-resolved T2 maps are crucial for medical and porous media analysis.
• Long MRI acquisition times necessitate accelerated experimental methods.
• Undersampling schemes and advanced reconstruction techniques are key to faster data acquisition.

Purpose of the Study:

• To investigate methods for accelerating the acquisition of 3D T2 maps.
• To compare different k-space sampling patterns and reconstruction algorithms.
• To achieve quantitative T2 mapping with significantly reduced acquisition times.

Main Methods:

• Utilized a multi-echo spin echo pulse sequence for data acquisition.
• Implemented two k-space sampling patterns: coherent and incoherent.
• Compared reconstruction techniques including Total Variation, Nuclear Norm, and Nuclear Total Generalized Variation (NTGV) regularization.

Main Results:

• Incoherent sampling patterns yielded superior reconstruction results when processed with NTGV regularization.
• Quantitative T2 maps were successfully obtained using as little as 3.1% of k-space data.
• This approach demonstrated a 32-fold reduction in acquisition time compared to fully sampled datasets.

Conclusions:

• The combination of incoherent k-space sampling and NTGV regularization is highly effective for accelerating 3D T2 map acquisition.
• This method enables significant time savings without compromising quantitative accuracy.
• The findings have broad implications for medical imaging and porous media research.