Volume 16 Supplement 1

Abstracts of the 17th Annual SCMR Scientific Sessions

Open Access

Motion-corrected compressed-sensing enables robust spiral first-pass perfusion imaging with whole heart coverage

  • Yang Yang1,
  • Xiao Chen1,
  • Frederick H Epstein1, 3,
  • Craig H Meyer1, 3,
  • Sujith Kuruvilla2,
  • Christopher M Kramer2, 3 and
  • Michael Salerno2, 3
Journal of Cardiovascular Magnetic Resonance201416(Suppl 1):O81


Published: 16 January 2014


First-pass perfusion imaging using CMR is an important tool for diagnosing coronary artery disease (CAD), but most clinical techniques are limited in their spatial coverage. While compressed-sensing (CS) holds promise for highly accelerated perfusion spiral imaging, CS techniques suffer from blurring artifacts in the setting of respiratory motion. Spiral pulse sequences have multiple advantages for myocardial perfusion imaging including high acquisition efficiency, high signal to noise (SNR) and robustness to motion and for CS including a relatively incoherent aliasing pattern. Thus, we develop a whole heart spiral first-pass perfusion sequence combined with robust motion-correction.


Eight subjects undergoing clinical scans were recruited for rest perfusion studies using an accelerated 4× spiral perfusion sequence with whole heart coverage. The sequence parameters included: 2 variable density interleaves, 6.1 ms readout per interleaf, TE 1.0 ms, TR 9 ms, TI 80 ms, FA 45°, FOV 340 mm2, in-plane resolution of ~2 mm, 6 ~ 10 slices to cover the whole heart. All perfusion images were acquired on a 1.5T Siemens Avanto scanner during infusion of 0.1 mmol/kg of Gd-DTPA. The images were reconstructed by direct reconstruction with zero padding (DC), L1-SPIRiT using finite time difference as the sparsity transform, and Block LOw-rank Sparsity with Motion guidance compressed sensing (BLOSM). The images were evaluated for quality and blurring by two experienced cardiologists.


Figure 1 shows a comparison of perfusion data reconstructed by DC, L1-SPIRiT and BLOSM with (top) and without (bottom) motion from the same subject at different time points. In the absence of motion, both the L1-SPIRiT and BLOSM images have high SNR and image quality without motion artifacts. However, in the presence of motion, the L1-SPIRiT reconstructed images suffer from temporal blurring artifacts (arrows in Figure 1.) as finite difference was imposed on the non-registered images as the sparsity transform. The BLOSM technique imposes sparsity of the tracked blocks which preserves edge information with minimal blurring. Figure 2 shows the spatial-temporal profiles across one subject from the three reconstruction methods. The DC profile shows sharp edges but with high artifacts and noise level. The L1-SPIRIT profile demonstrates high SNR but severe temporal blurring(highlighted by the yellow arrow). The BLOSM images show preserved temporal fidelity and high SNR even in the presence of significant respiratory motion. One cardiologist commented that for all of the data set, the BLOSM had better image quality and less blurring than L1-SPIRiT while the other cardiologist rated BLOSM better for 7 of 8 data sets.
Figure 1

Perfusion images reconstructed by direct recon with DC, L1-SPIRiT and BLOSM with (top) and without(bottom) motion.

Figure 2

Time profile curve from DC, L1-SPIRiT, and BLOSM reconstructed images.


We demonstrate the feasibility of whole heart spiral first pass perfusion imaging using BLOSM to perform robust CS even in the setting of significant respiratory motion. Further validation will be required in patients undergoing vasodilator stress CMR.


K23 HL112910, R01 HL079110.

Authors’ Affiliations

Biomedical Engineering, University of Virginia
Medicine, University of Virginia
Radiology, University of Virginia


© Yang et al.; licensee BioMed Central Ltd. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.