Quick Fire Session
SCMR 22nd Annual Scientific Sessions
Yang Yang, PhD
Assistant Professor
Icahn School of Medicine at Mount Sinai
Austin Robinson, MD
Fellow
University of Virginia
Roshin Mathew, MD
Cardiovascular Advanced Imaging Fellow
University of Virginia Health System
Christopher Kramer, MD
Ruth C. Heede Professor of Cardiovascular Medicine
University of Virginia Health System
Michael Salerno, MD
Associate Professor of Medicine (Cardiology), Radiology, and Biomedical Engineering
University of Virginia
Background: First-pass contrast-enhanced CMR myocardial perfusion is a valuable noninvasive tool to assess patients with coronary artery disease1, but current clinical available techniques are still limited by spatial-temporal resolution (~2mm) and ventricular coverage (3 slices)2, which reduces the sensitivity to detect perfusion difference between the endocardium and epicardium. We have previously demonstrated the feasibility of performing perfusion imaging ultra-high spatial resolution of 1.25 mm with whole heart coverage using dual density spiral trajectories at 3T5.In this study, we aim to further compare the imaging performance with combination of different techniques including varying spiral readout duration, outer-volume suppression (OVS)3, and simultaneous multi-slice (SMS) imaging4 to optimize the protocols for high-resolution clinical myocardial perfusion.
Methods: We evaluated the high-resolution pulse sequences by varying the readout length of spiral trajectory (4ms or 5ms per spiral interleave), imaging with or without the OVS module, or using interleaved or SMS acquisition as detailed in Table 1. Resting first-pass perfusion was performed with a 0.075 mmol/kg Gd-DTPA bolus in 64 scans from 43 normal subjects on a 3T Prisma Siemens scanner. The images were reconstructed by L1-SPIRiT6 or SMS-L1-SPIRiT7 using finite temporal difference as the sparsity transform. Images were graded by a cardiologist on overall quality of a 5-point scale (5-excellent, 1-poor) as well as signal dropout, dark rim artifact and blurriness.
Results: Figure 1 shows an example of eight slices perfusion images with 1.25 mm resolution from 4ms interleaved sequence without OVS in the top row and with OVS in the bottom row. Both of the pulse sequences with and without OVS produced high quality images. Figure 2 presented an example of high spatial resolution perfusion images using 5ms interleaved readout with OVS. There was some signal dropout in the inferior wall due to the off-resonance effect from longer spiral readout. Figure 3 showed an example of 4ms SMS readout without OVS. The average image scores of the interleaved and SMS sequences were 3.75±0.51 and 2.69±0.40 respectively (p<0.05) demonstrating better performance of the interleaved approach. Scores of 4ms and 5ms readout were 3.36±0.72 and 2.94±0.55 (p<0.05) largely due to less dropout and image blurring as expected. Image quality was better without OVS (3.33 versus 3.02) which is due to some dropout likely from B1 inhomogeneity at 3T.
Conclusion:
This study quantitatively evaluated ultra-high resolution sequences using different spiral readouts and acquisition strategies. The highest image quality was using 4ms readouts and an interleaved approach without OVS. High-resolution whole-heart perfusion with optimal protocols will provide a tool to quantify regional difference in perfusion of the subendo and subepi myocardium. Further validation will be required in patients undergoing adenosine stress CMR.