Spatiotemporal Flexible Sparse Reconstruction for Rapid Dynamic Contrast-Enhanced MRI.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Radiology

Background:

• Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is crucial for assessing tissue perfusion.
• Current free-breathing DCE-MRI methods combine compressed sensing (CS) and parallel imaging for motion robustness but face reconstruction challenges at high acceleration rates.
• Existing techniques often suffer from degraded image quality and prolonged reconstruction times.

Purpose of the Study:

• To develop a novel parallel CS reconstruction model for DCE-MRI to address limitations of existing methods.
• To enhance reconstruction quality and reduce overall reconstruction time, particularly at high acceleration factors.
• To improve the diagnostic performance of DCE-MRI through optimized spatial and temporal resolution.

Main Methods:

• Proposed a new parallel CS reconstruction model for DCE-MRI.
• Incorporated flexible weighted sparse constraints along spatial and temporal dimensions.
• Developed a fast-thresholding algorithm for efficient model solving.

Main Results:

• The proposed method achieved the lowest reconstruction error and highest image structural similarity at high acceleration factors for both brain tumor and liver DCE-MRI.
• Demonstrated faster reconstruction times for liver datasets.
• Yielded improved physiological measures for tumor images.

Conclusions:

• The novel parallel CS reconstruction model offers superior performance for DCE-MRI compared to existing methods.
• The approach effectively balances spatial and temporal sparsity for enhanced image reconstruction.
• This technique holds promise for improving diagnostic accuracy and efficiency in clinical DCE-MRI applications.