Simultaneous multislice cardiac magnetic resonance fingerprinting using low rank reconstruction

Abstract

This study introduces a technique for simultaneous multislice (SMS) cardiac magnetic resonance fingerprinting (cMRF), which improves the slice coverage when quantifying myocardial T1, T2, and M0. The single‐slice cMRF pulse sequence was modified to use multiband (MB) RF pulses for SMS imaging. Different RF phase schedules were used to excite each slice, similar to POMP or CAIPIRINHA, which imparts tissues with a distinguishable and slice‐specific magnetization evolution over time. Because of the high net acceleration factor (R = 48 in plane combined with the slice acceleration), images were first reconstructed with a low rank technique before matching data to a dictionary of signal timecourses generated by a Bloch equation simulation. The proposed method was tested in simulations with a numerical relaxation phantom. Phantom and in vivo cardiac scans of 10 healthy volunteers were also performed at 3 T. With single‐slice acquisitions, the mean relaxation times obtained using the low rank cMRF reconstruction agree with reference values. The low rank method improves the precision in T1 and T2 for both single‐slice and SMS cMRF, and it enables the acquisition of maps with fewer artifacts when using SMS cMRF at higher MB factors. With this technique, in vivo cardiac maps were acquired from three slices simultaneously during a breathhold lasting 16 heartbeats. SMS cMRF improves the efficiency and slice coverage of myocardial T1 and T2 mapping compared with both single‐slice cMRF and conventional cardiac mapping sequences. Thus, this technique is a first step toward whole‐heart simultaneous T1 and T2 quantification with cMRF.

3D free‐breathing cardiac magnetic resonance fingerprinting

“…56 2D cMRF studies comparing against T2prep-bSSFP and T2prep-FLASH (fast low angle shot) have also reported an underestimation bias. [27][28][29] T2prep methods and GraSE based methods have shown a negative and positive bias relative to multi-echo spin echo, 57 respectively. Underestimation was observed for high T 2 values (not relevant for myocardial characterization), likely due to the absence of long T2prep modules (to encode high T 2 values) in the proposed sequence.…”

“…Model corrections such as B 1 and slice profile for 2D cMRF have also been evaluated, 28 leading to reduced bias in the estimated parameters. Simultaneous multi-slice imaging has been combined with 2D cMRF 29 to increase the coverage of the heart; however, this is still limited to three slices and is sensitive to through-plane motion.…”

T1ρ magnetic resonance fingerprinting

“…For all acquisitions using the proposed sequence, reconstruction of the k ‐space data was performed using a low rank reconstruction similar to that performed by Hamilton et al, Zhao et al, and Tamir et al ESPIRiT coil calibration was performed and then low rank images were reconstructed using publicly available code for low rank NUFFT reconstruction from Assländer et al The low rank images were generated by restricting the image data to the subspace of the dictionary generated in the previous section. The subspace was calculated by taking the singular value decomposition (SVD) of the dictionary along the temporal dimension and multiplying the reconstructed images by the left‐hand operand of the SVD.…”

From Compressed-Sensing to Artificial Intelligence-Based Cardiac MRI Reconstruction

“…Globally low-rank (GLR) reconstructions, exploiting lowrankness on the entire image series, have been exploited in many cardiac applications such as dynamic cine MRI (39-41), real-time CMR (42), cardiac perfusion (43), or simultaneous multislice CMR fingerprinting (44). GLR reconstruction techniques are particularly suited for image series that exhibit strong correlation over time.…”