Skip to main content

Reliable segmentation of 2D cardiac magnetic resonance perfusion image sequences using time as the 3rd dimension

European Radiology volume 31, pages 3941–3950 (2021)Cite this article

Abstract

Objectives

Cardiac magnetic resonance (CMR) first-pass perfusion is an established noninvasive diagnostic imaging modality for detecting myocardial ischemia. A CMR perfusion sequence provides a time series of 2D images for dynamic contrast enhancement of the heart. Accurate myocardial segmentation of the perfusion images is essential for quantitative analysis and it can facilitate automated pixel-wise myocardial perfusion quantification.

Methods

In this study, we compared different deep learning methodologies for CMR perfusion image segmentation. We evaluated the performance of several image segmentation methods using convolutional neural networks, such as the U-Net in 2D and 3D (2D plus time) implementations, with and without additional motion correction image processing step. We also present a modified U-Net architecture with a novel type of temporal pooling layer which results in improved performance.

Results

The best DICE scores were 0.86 and 0.90 for LV myocardium and LV cavity, while the best Hausdorff distances were 2.3 and 2.1 pixels for LV myocardium and LV cavity using 5-fold cross-validation. The methods were corroborated in a second independent test set of 20 patients with similar performance (best DICE scores 0.84 for LV myocardium).

Conclusions

Our results showed that the LV myocardial segmentation of CMR perfusion images is best performed using a combination of motion correction and 3D convolutional networks which significantly outperformed all tested 2D approaches. Reliable frame-by-frame segmentation will facilitate new and improved quantification methods for CMR perfusion imaging.

Key Points

• Reliable segmentation of the myocardium offers the potential to perform pixel level perfusion assessment.

• A deep learning approach in combination with motion correction, 3D (2D + time) methods, and a deep temporal connection module produced reliable segmentation results.

This is a preview of subscription content, access via your institution.

Access options

Buy single article

Instant access to the full article PDF.

43,34 €

Price includes VAT (Finland)
Tax calculation will be finalised during checkout.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Abbreviations

CMR:

Cardiac magnetic resonance imaging

DTC:

Deeply temporally connected pooling layer

GPU:

Graphics processing unit

LV:

Left ventricle

MOCO:

Motion corrected

References

  1. Danad I, Szymonifka J, Twisk et al (2016) Diagnostic performance of cardiac imaging methods to diagnose ischaemia-causing coronary artery disease when directly compared with fractional flow reserve as a reference standard: a meta-analysis. Eur Heart J 095. https://doi.org/10.1093/eurheartj/ehw095

  2. Benovoy M, Jacobs M, Cheriet F, Dahdah N, Arai AE, Hsu LY (2017) Robust universal nonrigid motion correction framework for first-pass cardiac MR perfusion imaging. J Magn Reson Imaging 46(4):1060–1072. https://doi.org/10.1002/jmri.25659

    Article  PubMed  PubMed Central  Google Scholar 

  3. Hsu L-Y, Jacobs M, Benovoy M et al (2018) Diagnostic performance of fully automated pixel-wise quantitative myocardial perfusion imaging by cardiovascular magnetic resonance. JACC Cardiovasc Imaging 11(5):697–707. https://doi.org/10.1016/j.jcmg.2018.01.005

    Article  PubMed  Google Scholar 

  4. Huang H-H, Huang C-Y, Chen C-N, Wang Y-W, Huang T-Y (2017) Automatic regional analysis of myocardial native t1 values: left ventricle segmentation and AHA parcellations. Int J Cardiovasc Imaging 34(1):131–140. https://doi.org/10.1007/s10554-017-1216-x

    Article  PubMed  Google Scholar 

  5. Sahiner B, Pezeshk A, Hadjiiski et al (2018) Deep learning in medical imaging and radiation therapy. Med Phys 46(1):1–36. https://doi.org/10.1002/mp.13264

    Article  Google Scholar 

  6. Scannell CM, Veta M, Villa ADM et al (2019) Deep-learning-based preprocessing for quantitative myocardial perfusion MRI. J Magn Reson Imaging. https://doi.org/10.1002/jmri.26983

  7. Fahmy AS, El-Rewaidy H, Nezafat M, Nakamori S, Nezafat R (2019) Automated analysis of cardiovascular magnetic resonance myocardial native t1 mapping images using fully convolutional neural networks. J Cardiovasc Magn Reson 21(1). https://doi.org/10.1186/s12968-0180516-1

  8. Klein S, Staring M, Murphy K, Viergever MA, Pluim J (2010) elastix: A toolbox for intensity-based medical image registration. IEEE Trans Med Imaging 29(1):196–205. https://doi.org/10.1109/tmi.2009.2035616

    Article  PubMed  Google Scholar 

  9. Jacobs M, Benovoy M, Chang L-C, Arai AE, Hsu L-Y (2016) Evaluation of an automated method for arterial input function detection for first pass myocardial perfusion cardiovascular magnetic resonance. J Cardiovasc Magn Reson 18(1). https://doi.org/10.1186/s12968-0160239-0

  10. Kellman P, Arai AE (2007) Imaging sequences for first pass perfusion - a review. J Cardiovasc Magn Reson 9(3):525–537. https://doi.org/10.1080/10976640601187604

    Article  PubMed  Google Scholar 

  11. Cicek O, Abdulkadir A, Lienkamp SS, Brox T, Ronneberger O (2016) 3D U-Net: learning dense volumetric segmentation from sparse annotation. ArXiv e-prints. 1606.06650

  12. Kayalibay B, Jensen G, van der Smagt P (2017) CNN-based Segmentation of medical imaging data. ArXiv e-prints. 1701.03056

  13. Wu Y, He K (2018) Group normalization. ArXiv e-prints. 1803.08494

  14. Liu S, Xu D, Zhou SK et al (2018) 3d anisotropic hybrid network: Transferring convolutional features from 2d images to 3d anisotropic volumes. In: Frangi AF, Schnabel JA, Davatzikos C, Alberola-Lo’pez C, Fichtinger G (eds) Medical image computing and computer assisted intervention – MICCAI 2018. Springer, Cham, pp 851–858

    Chapter  Google Scholar 

  15. Luo W, Li Y, Urtasun R, Zemel R (2016) Understanding the effective receptive field in deep convolutional neural networks. In: Lee DD, Sugiyama M, Luxburg UV, Guyon I, Garnett R (eds) Advances in Neural Information Processing Systems 29. Curran Associates, Inc., New York, pp 4898–4906

    Google Scholar 

  16. Zhao H, Shi J, Qi X, Wang X, Jia J (2016) Pyramid scene parsing network. arXiv:1612.01105

  17. MedPy package by Maier, O.: https://pypi.org/project/MedPy/ Accessed 8/10/2020

  18. Petitjean C, Zuluaga MA, Bai W et al (2015) Right ventricle segmentation from cardiac MRI: a collation study. Med Image Anal 19(1):187–202. https://doi.org/10.1016/j.media.2014.10.004

    Article  PubMed  Google Scholar 

  19. Gupta V, Kirişli HA, Hendriks EA et al (2012) Cardiac MR perfusion image processing techniques: a survey. Med Image Anal 16(4):767–785. https://doi.org/10.1016/j.media.2011.12.005

    Article  PubMed  Google Scholar 

  20. Kim YC, Kim KR, Choe YH (2020) Automatic myocardial segmentation in dynamic contrast enhanced perfusion MRI using Monte Carlo dropout in an encoder-decoder convolutional neural network. Comput Methods Programs Biomed 185. https://doi.org/10.1016/j.cmpb.2019.105150

Download references

Acknowledgements

This work used the computational resources of the NIH HPC Biowulf cluster (http://hpc.nih.gov). We are grateful for the support of the NIH Biowulf team.

Funding

This work was supported by the NIH intramural research program of the National Heart, Lung and Blood Institute (ZIA HL006137-08).

Author information

Affiliations

  1. Stanford Medicine, Pasteur Drive 300, Stanford, CA, 94305, USA

    Veit Sandfort

  2. National Heart Lung and Blood Institute, National Institutes of Health, Bethesda, MD, USA

    Veit Sandfort, Matthew Jacobs, Andrew E. Arai & Li-Yueh Hsu

  3. Department of Electrical Engineering and Computer Science, The Catholic University of America, Washington, DC, USA

    Matthew Jacobs

  4. Clinical Center, National Institutes of Health, Bethesda, MD, USA

    Li-Yueh Hsu

Authors
  1. Veit Sandfort
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Matthew Jacobs
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Andrew E. Arai
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Li-Yueh Hsu
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

VS planned and performed experiments, analyzed data, and drafted the manuscript. LH and AEA conceived experiments, evaluated data, and edited manuscript. MJ contributed to data, experiments, and manuscript.

Corresponding author

Correspondence to Veit Sandfort.

Ethics declarations

Guarantor

The scientific guarantor of this publication is Veit Sandfort, MD.

Conflict of interest

The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article.

Statistics and biometry

No complex statistical methods were necessary for this paper.

Informed consent

Written informed consent was obtained from all subjects (patients) in this study.

Ethical approval

Institutional Review Board approval was obtained.

Study subjects or cohorts overlap

Some study subjects or cohorts have been previously reported in “Evaluation of an Automated Method for Arterial Input Function Detection for First-Pass Myocardial Perfusion Cardiovascular Magnetic Resonance” Matthew Jacobs, Mitchel Benovoy, Lin-Ching Chang, Andrew E Arai, Li-Yueh Hsu, 10.1186/s12968-016-0239-0.

Methodology

• Retrospective, performed at one institution

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

ESM 1

(DOCX 19.7 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Sandfort, V., Jacobs, M., Arai, A.E. et al. Reliable segmentation of 2D cardiac magnetic resonance perfusion image sequences using time as the 3rd dimension. Eur Radiol 31, 3941–3950 (2021). https://doi.org/10.1007/s00330-020-07474-5

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00330-020-07474-5

Keywords

  • Deep learning
  • Image segmentation
  • Cardiac magnetic resonance imaging
  • Myocardial perfusion