Anatomical structure-guided joint spatiotemporal graph embedding framework for magnetic resonance fingerprint...

Area of Science:

• Medical Imaging
• Computational Imaging
• Biophysics

Background:

• Undersampled Magnetic Resonance Fingerprinting (MRF) acquisition schemes often result in aliasing artifacts, compromising quantitative imaging accuracy.
• Current graph-based reconstruction methods struggle with high computational costs and capturing complex spatiotemporal dynamics.

Purpose of the Study:

• To develop an efficient and accurate MRF reconstruction framework.
• To address limitations of existing methods in handling aliasing artifacts and computational complexity.

Main Methods:

• Proposed an anatomical structure-guided joint spatiotemporal graph embedding framework.
• Integrated anatomical segmentation and homogeneity clustering to partition MRF data.
• Constructed subgraphs for each cluster to capture spatial correlations and temporal dynamics.

Main Results:

• Achieved a ~2 dB higher signal-to-noise ratio (SNR) in reconstructed data compared to state-of-the-art methods.
• Demonstrated a ~70% reduction in reconstruction time.
• Outperformed existing methods on both simulated and in vivo MRF datasets.

Conclusions:

• The proposed framework effectively reduces aliasing artifacts in MRF reconstruction.
• The method offers significant improvements in reconstruction accuracy and computational efficiency.
• Publicly available source code facilitates further research and application.