Rapid, Efficient Imaging

1348975 - Simultaneous Multislice Imaging via Linear Phase Modulated Extended Field of View (SMILE) for Cardiac Cine and Perfusion

Thursday, January 26, 2023

5:50 PM – 6:00 PM

Location: Silver Pearl 2

Presenting Author(s)

SZ

Shen Zhao, PhD

Postdoctoral Scholar
Stanford University
Sunnyvale, California, United States

Co-Author(s)

XW

Xitong Wang, MSc

Graduate Student
Stanford University, United States
Junyu Wang, PhD

Postdoctoral Fellow
Stanford University
Palo Alto, California, United States
Michael Salerno, MD, PhD, MSc

Professor
Stanford University, California, United States

Disclosure(s):

Michael Salerno, MD, PhD, MSc: No relevant disclosure to display

Background: Simultaneous multislice imaging (SMS) significantly reduces acquisition time with little signal-to-noise penalty, and has found utility in CMR for cine, parametric mapping, and perfusion. A significant issue is "slice-leakage”, which is interference from the other excited slices leading to strategies to block this interference such as CAIPIRINHA, Slice GRAPPA(SG), Split Slice GRAPPA (SPSG), and ROCK-SPIRiT (RS). However, slice-leakage remains the main drawback and limitation to further acceleration. A technique called POMP (phase-offset multiplanar volume imaging) was developed in 1991 to use a larger FOV to disentangle different fully-sampled slices. In this abstract, we extend POMP and develop a new efficient SMS acquisition and reconstruction technique, Simultaneous Multislice Imaging via Linear phase modulated Extended field of view (SMILE) for cardiac Cine and perfusion that can avoid interslice leakage theoretically. Simulated SMS cine and perfusion experiments validate its superior performance.

Methods: For each cardiac phase, SMILE samples in the superposition of multiple linear-phase-modulated k-space with an extended FOV to avoid abrupt image content change. For SMILE we adopt CAVA spatiotemporal sampling, with HICU as the reconstruction algorithm. We compared SMILE results to conventional SMS acquisition (summation of multislice k-space) and corresponding reconstruction methods: SG, SPSG, and RS with CAIPIRINHA modulation. For the cine experiment, five 3T fully sampled slices from one volunteer are used as ground truth, with no additional in-plane acceleration for a net acceleration of rate R=5. For the perfusion study, four 1.5T fully sampled slices are used as ground truth. Data were simulated with SMS and R=2 in-plane uniform down-sampling achieving a total acceleration of R=8. SG and SPSG were followed by in-plane GRAPPA. Reconstructions were evaluated using NMSE and SSIM.

Results:

SMILE+HICU outperformed state-of-the-art SMS techniques for both the Cine and Perfusion Experiments. Figure. 1 and Table. 1 show the data from the 5 slices cine experiment. SMILE +HICU outperformed the techniques by 2.6 dB in terms of NMSE and has much less perceptual error as validated by SSIM. Figure. 2 and Table. 1 show that for the challenging 4 slices perfusion experiment with a net acceleration rate R = 8, SMILE+HICU can still reconstruct images with good image quality. In contrast, current SMS techniques show strong slice-leakage.

Conclusion:

SMILE represents a general approach for SMS which should be free of slice-leakage artifacts with potentially broad applicability in cardiac MRI. SMILE+HICU demonstrates high image quality even at high total acceleration factors.