Improving Amide Proton Transfer-Weighted MRI Reconstruction Using T2-Weighted Images

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Biomedical Engineering
• Medical Imaging Physics

Background:

• Current Amide Proton Transfer-weighted (APTw) imaging protocols rely on high-resolution T2-weighted (T2w) images for guidance.
• Existing MRI acceleration methods for APTw imaging do not leverage structural information from T2w images during reconstruction.
• This limits the efficiency and quality of accelerated APTw MRI acquisition.

Purpose of the Study:

• To develop a novel framework for accelerating APTw imaging by reconstructing images directly from undersampled k-space data.
• To integrate structural information from T2w images into the APTw reconstruction process.
• To improve the quality and efficiency of APTw MRI acquisition.

Main Methods:

• A novel sparse representation-based slice matching algorithm was developed to identify corresponding T2w slices from undersampled APTw data.
• A Recurrent Feature Sharing Reconstruction network (RFS-Rec) was designed, incorporating a Convolutional Recurrent Neural Network (CRNN).
• The RFS-Rec network utilizes intermediate features from matched T2w images to reconstruct high-quality APTw images from undersampled k-space data.

Main Results:

• The proposed RFS-Rec framework successfully reconstructs APTw images from highly undersampled k-space data.
• The method effectively incorporates structural information from matched T2w images, enhancing APTw image quality.
• Experiments on rat and human brain datasets demonstrated that RFS-Rec outperforms state-of-the-art reconstruction methods.

Conclusions:

• The novel RFS-Rec framework offers an effective approach to accelerate APTw MRI acquisition.
• Integrating structural information from T2w images significantly improves the quality of reconstructed APTw images.
• The proposed method shows promise for advancing clinical and research applications of APTw imaging.