Phase-sensitive deep reconstruction method for rapid multiparametric MR fingerprinting and quantitative...

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Medical Physics
• Quantitative Imaging

Background:

• Magnetic Resonance Fingerprinting (MRF) is an advanced MRI technique.
• Simultaneous quantification of multiple tissue properties (T1, T2, Proton Density, B1+, phase, QSM) is challenging.
• Novel deep learning approaches are needed to improve MRF efficiency and accuracy.

Purpose of the Study:

• To evaluate the efficacy of a novel Phase-Sensitivity Deep Reconstruction Network (PS-DRONE) for MRF.
• To enable simultaneous quantification of T1, T2, Proton Density, B1+, phase, and quantitative susceptibility mapping (QSM).
• To assess the performance of PS-DRONE in both phantom and in vivo studies.

Main Methods:

• Acquisition of data at 3 Tesla using an optimized EPI-based MRF sequence.
• Phantom experiments utilizing standardized phantoms for T1/T2 and custom phantoms for B1/QSM.
• In vivo experiments on healthy volunteers and a patient with brain metastasis, comparing PS-DRONE maps to standard sequences.

Main Results:

• Achieved a scan time of 2 minutes for 32 slices with high resolution.
• Reconstruction of all quantitative maps (T1, T2, PD, B1+, phase) completed within 1 second.
• PS-DRONE demonstrated strong correlations (r=0.99 for T1, T2, QSM) with reference data in phantoms and in vivo, with minor bias noted for B1 and QSM.

Conclusions:

• The PS-DRONE network successfully enables concurrent acquisition of multiple quantitative MRI maps.
• Achieved rapid acquisition (2 min) and reconstruction (1 s) times.
• PS-DRONE shows significant potential for efficient and accurate quantitative MRI.