King's College London

Research portal

Combined simultaneous multislice bSSFP and compressed sensing for first-pass myocardial perfusion at 1.5 T with high spatial resolution and coverage

Research output: Contribution to journalArticle

Sarah McElroy, Giulio Ferrazzi, Muhummad Sohaib Nazir, Karl P Kunze, Radhouene Neji, Peter Speier, Daniel Stäb, Christoph Forman, Reza Razavi, Amedeo Chiribiri, Sébastien Roujol

Original languageEnglish
JournalMagnetic Resonance in Medicine
Publication statusE-pub ahead of print - 12 Jun 2020

Bibliographical note

© 2020 The Authors. Magnetic Resonance in Medicine published by Wiley Periodicals LLC on behalf of International Society for Magnetic Resonance in Medicine.

King's Authors


PURPOSE: To implement and evaluate a pseudorandom undersampling scheme for combined simultaneous multislice (SMS) balanced SSFP (bSSFP) and compressed-sensing (CS) reconstruction to enable myocardial perfusion imaging with high spatial resolution and coverage at 1.5 T.

METHODS: A prospective pseudorandom undersampling scheme that is compatible with SMS-bSSFP phase-cycling requirements and CS was developed. The SMS-bSSFP CS with pseudorandom and linear undersampling schemes were compared in a phantom. A high-resolution (1.4 × 1.4 mm2 ) six-slice SMS-bSSFP CS perfusion sequence was compared with a conventional (1.9 × 1.9 mm2 ) three-slice sequence in 10 patients. Qualitative assessment of image quality, perceived SNR, and number of diagnostic segments and quantitative measurements of sharpness, upslope index, and contrast ratio were performed.

RESULTS: In phantom experiments, pseudorandom undersampling resulted in residual artifact (RMS error) reduction by a factor of 7 compared with linear undersampling. In vivo, the proposed sequence demonstrated higher perceived SNR (2.9 ± 0.3 vs. 2.2 ± 0.6, P = .04), improved sharpness (0.35 ± 0.03 vs. 0.32 ± 0.05, P = .01), and a higher number of diagnostic segments (100% vs. 94%, P = .03) compared with the conventional sequence. There were no significant differences between the sequences in terms of image quality (2.5 ± 0.4 vs. 2.8 ± 0.2, P = .08), upslope index (0.11 ± 0.02 vs. 0.10 ± 0.01, P = .3), or contrast ratio (3.28 ± 0.35 vs. 3.36 ± 0.43, P = .7).

CONCLUSION: A pseudorandom k-space undersampling compatible with SMS-bSSFP and CS reconstruction has been developed and enables cardiac MR perfusion imaging with increased spatial resolution and myocardial coverage, increased number of diagnostic segments and perceived SNR, and no difference in image quality, upslope index, and contrast ratio.

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454