Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI

Elena Martín-González; Ebraham Alskaf; Amedeo Chiribiri; Pablo Casaseca-de-la-Higuera; Carlos Alberola-López; Rita G. Nunes; Teresa Correia

doi:10.1007/978-3-030-88552-6_9

Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI

Elena Martín-González^*, Ebraham Alskaf, Amedeo Chiribiri, Pablo Casaseca-de-la-Higuera, Carlos Alberola-López, Rita G. Nunes, Teresa Correia

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

2 Citations (Scopus)

Abstract

First-pass perfusion cardiac magnetic resonance (FPP-CMR) is becoming an essential non-invasive imaging method for detecting deficits of myocardial blood flow, allowing the assessment of coronary heart disease. Nevertheless, acquisitions suffer from relatively low spatial resolution and limited heart coverage. Compressed sensing (CS) methods have been proposed to accelerate FPP-CMR and achieve higher spatial resolution. However, the long reconstruction times have limited the widespread clinical use of CS in FPP-CMR. Deep learning techniques based on supervised learning have emerged as alternatives for speeding up reconstructions. However, these approaches require fully sampled data for training, which is not possible to obtain, particularly high-resolution FPP-CMR images. Here, we propose a physics-informed self-supervised deep learning FPP-CMR reconstruction approach for accelerating FPP-CMR scans and hence facilitate high spatial resolution imaging. The proposed method provides high-quality FPP-CMR images from 10x undersampled data without using fully sampled reference data.

Original language	English
Title of host publication	Machine Learning for Medical Image Reconstruction - 4th International Workshop, MLMIR 2021, Held in Conjunction with MICCAI 2021, Proceedings
Editors	Nandinee Haq, Patricia Johnson, Andreas Maier, Tobias Würfl, Jaejun Yoo
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	86-95
Number of pages	10
ISBN (Print)	9783030885519
DOIs	https://doi.org/10.1007/978-3-030-88552-6_9
Publication status	Published - 2021
Event	4th International Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2021 held in Conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021 - Virtual, Online Duration: 1 Oct 2021 → 1 Oct 2021

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12964 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	4th International Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2021 held in Conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021
City	Virtual, Online
Period	1/10/2021 → 1/10/2021

Keywords

Deep learning reconstruction
Model-based reconstruction
Quantitative perfusion cardiac MRI

Access to Document

10.1007/978-3-030-88552-6_9

Cite this

Martín-González, E., Alskaf, E., Chiribiri, A., Casaseca-de-la-Higuera, P., Alberola-López, C., Nunes, R. G., & Correia, T. (2021). Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI. In N. Haq, P. Johnson, A. Maier, T. Würfl, & J. Yoo (Eds.), Machine Learning for Medical Image Reconstruction - 4th International Workshop, MLMIR 2021, Held in Conjunction with MICCAI 2021, Proceedings (pp. 86-95). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12964 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-88552-6_9

Martín-González, Elena ; Alskaf, Ebraham ; Chiribiri, Amedeo et al. / Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI. Machine Learning for Medical Image Reconstruction - 4th International Workshop, MLMIR 2021, Held in Conjunction with MICCAI 2021, Proceedings. editor / Nandinee Haq ; Patricia Johnson ; Andreas Maier ; Tobias Würfl ; Jaejun Yoo. Springer Science and Business Media Deutschland GmbH, 2021. pp. 86-95 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inbook{0382fa256e93406cb3f0e28ae6ab0814,

title = "Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI",

abstract = "First-pass perfusion cardiac magnetic resonance (FPP-CMR) is becoming an essential non-invasive imaging method for detecting deficits of myocardial blood flow, allowing the assessment of coronary heart disease. Nevertheless, acquisitions suffer from relatively low spatial resolution and limited heart coverage. Compressed sensing (CS) methods have been proposed to accelerate FPP-CMR and achieve higher spatial resolution. However, the long reconstruction times have limited the widespread clinical use of CS in FPP-CMR. Deep learning techniques based on supervised learning have emerged as alternatives for speeding up reconstructions. However, these approaches require fully sampled data for training, which is not possible to obtain, particularly high-resolution FPP-CMR images. Here, we propose a physics-informed self-supervised deep learning FPP-CMR reconstruction approach for accelerating FPP-CMR scans and hence facilitate high spatial resolution imaging. The proposed method provides high-quality FPP-CMR images from 10x undersampled data without using fully sampled reference data.",

keywords = "Deep learning reconstruction, Model-based reconstruction, Quantitative perfusion cardiac MRI",

author = "Elena Mart{\'i}n-Gonz{\'a}lez and Ebraham Alskaf and Amedeo Chiribiri and Pablo Casaseca-de-la-Higuera and Carlos Alberola-L{\'o}pez and Nunes, {Rita G.} and Teresa Correia",

note = "Funding Information: Acknowledgements. This work is part of a project that has received funding from the European Union{\textquoteright}s Horizon 2020 research and innovation programme under grant agreement No 867450. Authors also thank European Social Fund, Operational Programme of Castilla y Le{\'o}n, and the Junta de Castilla y Le{\'o}n. This work has also been supported by Agencia Estatal de Investigaci{\'o}n through grant TEC2017-82408-R and by Funda¸c{\~a}o para a Ci{\^e}ncia e Tecnologia (FCT) through grant UIDP/50009/2020. Funding Information: This work is part of a project that has received funding from the European Union?s Horizon 2020 research and innovation programme under grant agreement No 867450. Authors also thank European Social Fund, Operational Programme of Castilla y Le?n, and the Junta de Castilla y Le?n. This work has also been supported by Agencia Estatal de Investigaci?n through grant TEC2017-82408-R and by Funda??o para a Ci?ncia e Tecnologia (FCT) through grant UIDP/50009/2020. Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; 4th International Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2021 held in Conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021 ; Conference date: 01-10-2021 Through 01-10-2021",

year = "2021",

doi = "10.1007/978-3-030-88552-6_9",

language = "English",

isbn = "9783030885519",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "86--95",

editor = "Nandinee Haq and Patricia Johnson and Andreas Maier and Tobias W{\"u}rfl and Jaejun Yoo",

booktitle = "Machine Learning for Medical Image Reconstruction - 4th International Workshop, MLMIR 2021, Held in Conjunction with MICCAI 2021, Proceedings",

address = "Germany",

}

Martín-González, E, Alskaf, E , Chiribiri, A, Casaseca-de-la-Higuera, P, Alberola-López, C, Nunes, RG & Correia, T 2021, Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI. in N Haq, P Johnson, A Maier, T Würfl & J Yoo (eds), Machine Learning for Medical Image Reconstruction - 4th International Workshop, MLMIR 2021, Held in Conjunction with MICCAI 2021, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12964 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 86-95, 4th International Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2021 held in Conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021, Virtual, Online, 1/10/2021. https://doi.org/10.1007/978-3-030-88552-6_9

Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI. / Martín-González, Elena; Alskaf, Ebraham ; Chiribiri, Amedeo et al.
Machine Learning for Medical Image Reconstruction - 4th International Workshop, MLMIR 2021, Held in Conjunction with MICCAI 2021, Proceedings. ed. / Nandinee Haq; Patricia Johnson; Andreas Maier; Tobias Würfl; Jaejun Yoo. Springer Science and Business Media Deutschland GmbH, 2021. p. 86-95 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12964 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

TY - CHAP

T1 - Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI

AU - Martín-González, Elena

AU - Alskaf, Ebraham

AU - Chiribiri, Amedeo

AU - Casaseca-de-la-Higuera, Pablo

AU - Alberola-López, Carlos

AU - Nunes, Rita G.

AU - Correia, Teresa

N1 - Funding Information: Acknowledgements. This work is part of a project that has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 867450. Authors also thank European Social Fund, Operational Programme of Castilla y León, and the Junta de Castilla y León. This work has also been supported by Agencia Estatal de Investigación through grant TEC2017-82408-R and by Funda¸cão para a Ciência e Tecnologia (FCT) through grant UIDP/50009/2020. Funding Information: This work is part of a project that has received funding from the European Union?s Horizon 2020 research and innovation programme under grant agreement No 867450. Authors also thank European Social Fund, Operational Programme of Castilla y Le?n, and the Junta de Castilla y Le?n. This work has also been supported by Agencia Estatal de Investigaci?n through grant TEC2017-82408-R and by Funda??o para a Ci?ncia e Tecnologia (FCT) through grant UIDP/50009/2020. Publisher Copyright: © 2021, Springer Nature Switzerland AG.

PY - 2021

Y1 - 2021

N2 - First-pass perfusion cardiac magnetic resonance (FPP-CMR) is becoming an essential non-invasive imaging method for detecting deficits of myocardial blood flow, allowing the assessment of coronary heart disease. Nevertheless, acquisitions suffer from relatively low spatial resolution and limited heart coverage. Compressed sensing (CS) methods have been proposed to accelerate FPP-CMR and achieve higher spatial resolution. However, the long reconstruction times have limited the widespread clinical use of CS in FPP-CMR. Deep learning techniques based on supervised learning have emerged as alternatives for speeding up reconstructions. However, these approaches require fully sampled data for training, which is not possible to obtain, particularly high-resolution FPP-CMR images. Here, we propose a physics-informed self-supervised deep learning FPP-CMR reconstruction approach for accelerating FPP-CMR scans and hence facilitate high spatial resolution imaging. The proposed method provides high-quality FPP-CMR images from 10x undersampled data without using fully sampled reference data.

AB - First-pass perfusion cardiac magnetic resonance (FPP-CMR) is becoming an essential non-invasive imaging method for detecting deficits of myocardial blood flow, allowing the assessment of coronary heart disease. Nevertheless, acquisitions suffer from relatively low spatial resolution and limited heart coverage. Compressed sensing (CS) methods have been proposed to accelerate FPP-CMR and achieve higher spatial resolution. However, the long reconstruction times have limited the widespread clinical use of CS in FPP-CMR. Deep learning techniques based on supervised learning have emerged as alternatives for speeding up reconstructions. However, these approaches require fully sampled data for training, which is not possible to obtain, particularly high-resolution FPP-CMR images. Here, we propose a physics-informed self-supervised deep learning FPP-CMR reconstruction approach for accelerating FPP-CMR scans and hence facilitate high spatial resolution imaging. The proposed method provides high-quality FPP-CMR images from 10x undersampled data without using fully sampled reference data.

KW - Deep learning reconstruction

KW - Model-based reconstruction

KW - Quantitative perfusion cardiac MRI

UR - http://www.scopus.com/inward/record.url?scp=85116855436&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-88552-6_9

DO - 10.1007/978-3-030-88552-6_9

M3 - Conference paper

AN - SCOPUS:85116855436

SN - 9783030885519

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 86

EP - 95

BT - Machine Learning for Medical Image Reconstruction - 4th International Workshop, MLMIR 2021, Held in Conjunction with MICCAI 2021, Proceedings

A2 - Haq, Nandinee

A2 - Johnson, Patricia

A2 - Maier, Andreas

A2 - Würfl, Tobias

A2 - Yoo, Jaejun

PB - Springer Science and Business Media Deutschland GmbH

T2 - 4th International Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2021 held in Conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021

Y2 - 1 October 2021 through 1 October 2021

ER -

Martín-González E, Alskaf E , Chiribiri A, Casaseca-de-la-Higuera P, Alberola-López C, Nunes RG et al. Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI. In Haq N, Johnson P, Maier A, Würfl T, Yoo J, editors, Machine Learning for Medical Image Reconstruction - 4th International Workshop, MLMIR 2021, Held in Conjunction with MICCAI 2021, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 86-95. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-88552-6_9

Physics-Informed Self-supervised Deep Learning Reconstruction for Accelerated First-Pass Perfusion Cardiac MRI

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this