Deep learning-based prediction of kinetic parameters from myocardial perfusion MRI

Cian M. Scannell, Piet van den Bosch, Amedeo Chiribiri, Jack Lee, Marcel Breeuwer, Mitko Veta

Research output: Contribution to journalArticleAcademic

5 Downloads (Pure)

Abstract

The quantification of myocardial perfusion MRI has the potential to provide a fast, automated and user-independent assessment of myocardial ischaemia. However, due to the relatively high noise level and low temporal resolution of the acquired data and the complexity of the tracer-kinetic models, the model fitting can yield unreliable parameter estimates. A solution to this problem is the use of Bayesian inference which can incorporate prior knowledge and improve the reliability of the parameter estimation. This, however, uses Markov chain Monte Carlo sampling to approximate the posterior distribution of the kinetic parameters which is extremely time intensive. This work proposes training convolutional networks to directly predict the kinetic parameters from the signal-intensity curves that are trained using estimates obtained from the Bayesian inference. This allows fast estimation of the kinetic parameters with a similar performance to the Bayesian inference.
Original languageEnglish
Article number1907.11899v1
Number of pages4
JournalarXiv
Publication statusPublished - 27 Jul 2019

Fingerprint

Kinetic parameters
Magnetic resonance imaging
Parameter estimation
Markov processes
Sampling
Kinetics
Deep learning

Bibliographical note

Medical Imaging with Deep Learning: MIDL 2019 Extended Abstract Track. MIDL 2019

Cite this

Scannell, C. M., Bosch, P. V. D., Chiribiri, A., Lee, J., Breeuwer, M., & Veta, M. (2019). Deep learning-based prediction of kinetic parameters from myocardial perfusion MRI. arXiv, [1907.11899v1].
Scannell, Cian M. ; Bosch, Piet van den ; Chiribiri, Amedeo ; Lee, Jack ; Breeuwer, Marcel ; Veta, Mitko. / Deep learning-based prediction of kinetic parameters from myocardial perfusion MRI. In: arXiv. 2019.
@article{877e80768d104c53bea2013a05e43dd8,
title = "Deep learning-based prediction of kinetic parameters from myocardial perfusion MRI",
abstract = "The quantification of myocardial perfusion MRI has the potential to provide a fast, automated and user-independent assessment of myocardial ischaemia. However, due to the relatively high noise level and low temporal resolution of the acquired data and the complexity of the tracer-kinetic models, the model fitting can yield unreliable parameter estimates. A solution to this problem is the use of Bayesian inference which can incorporate prior knowledge and improve the reliability of the parameter estimation. This, however, uses Markov chain Monte Carlo sampling to approximate the posterior distribution of the kinetic parameters which is extremely time intensive. This work proposes training convolutional networks to directly predict the kinetic parameters from the signal-intensity curves that are trained using estimates obtained from the Bayesian inference. This allows fast estimation of the kinetic parameters with a similar performance to the Bayesian inference.",
keywords = "eess.IV, cs.CV, physics.med-ph, q-bio.QM",
author = "Scannell, {Cian M.} and Bosch, {Piet van den} and Amedeo Chiribiri and Jack Lee and Marcel Breeuwer and Mitko Veta",
note = "Medical Imaging with Deep Learning: MIDL 2019 Extended Abstract Track. MIDL 2019",
year = "2019",
month = "7",
day = "27",
language = "English",
journal = "arXiv",
publisher = "Cornell University Library",

}

Deep learning-based prediction of kinetic parameters from myocardial perfusion MRI. / Scannell, Cian M.; Bosch, Piet van den; Chiribiri, Amedeo; Lee, Jack; Breeuwer, Marcel; Veta, Mitko.

In: arXiv, 27.07.2019.

Research output: Contribution to journalArticleAcademic

TY - JOUR

T1 - Deep learning-based prediction of kinetic parameters from myocardial perfusion MRI

AU - Scannell, Cian M.

AU - Bosch, Piet van den

AU - Chiribiri, Amedeo

AU - Lee, Jack

AU - Breeuwer, Marcel

AU - Veta, Mitko

N1 - Medical Imaging with Deep Learning: MIDL 2019 Extended Abstract Track. MIDL 2019

PY - 2019/7/27

Y1 - 2019/7/27

N2 - The quantification of myocardial perfusion MRI has the potential to provide a fast, automated and user-independent assessment of myocardial ischaemia. However, due to the relatively high noise level and low temporal resolution of the acquired data and the complexity of the tracer-kinetic models, the model fitting can yield unreliable parameter estimates. A solution to this problem is the use of Bayesian inference which can incorporate prior knowledge and improve the reliability of the parameter estimation. This, however, uses Markov chain Monte Carlo sampling to approximate the posterior distribution of the kinetic parameters which is extremely time intensive. This work proposes training convolutional networks to directly predict the kinetic parameters from the signal-intensity curves that are trained using estimates obtained from the Bayesian inference. This allows fast estimation of the kinetic parameters with a similar performance to the Bayesian inference.

AB - The quantification of myocardial perfusion MRI has the potential to provide a fast, automated and user-independent assessment of myocardial ischaemia. However, due to the relatively high noise level and low temporal resolution of the acquired data and the complexity of the tracer-kinetic models, the model fitting can yield unreliable parameter estimates. A solution to this problem is the use of Bayesian inference which can incorporate prior knowledge and improve the reliability of the parameter estimation. This, however, uses Markov chain Monte Carlo sampling to approximate the posterior distribution of the kinetic parameters which is extremely time intensive. This work proposes training convolutional networks to directly predict the kinetic parameters from the signal-intensity curves that are trained using estimates obtained from the Bayesian inference. This allows fast estimation of the kinetic parameters with a similar performance to the Bayesian inference.

KW - eess.IV

KW - cs.CV

KW - physics.med-ph

KW - q-bio.QM

M3 - Article

JO - arXiv

JF - arXiv

M1 - 1907.11899v1

ER -

Scannell CM, Bosch PVD, Chiribiri A, Lee J, Breeuwer M, Veta M. Deep learning-based prediction of kinetic parameters from myocardial perfusion MRI. arXiv. 2019 Jul 27. 1907.11899v1.