Direct Estimation of Pharmacokinetic Parameters from DCE-MRI Using Deep CNN with Forward Physical Model Loss

  • Cagdas UlasEmail author
  • Giles Tetteh
  • Michael J. Thrippleton
  • Paul A. Armitage
  • Stephen D. Makin
  • Joanna M. Wardlaw
  • Mike E. Davies
  • Bjoern H. Menze
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11070)


Dynamic contrast-enhanced (DCE) MRI is an evolving imaging technique that provides a quantitative measure of pharmacokinetic (PK) parameters in body tissues, in which series of \(T_1\)-weighted images are collected following the administration of a paramagnetic contrast agent. Unfortunately, in many applications, conventional clinical DCE-MRI suffers from low spatiotemporal resolution and insufficient volume coverage. In this paper, we propose a novel deep learning based approach to directly estimate the PK parameters from undersampled DCE-MRI data. Specifically, we design a custom loss function where we incorporate a forward physical model that relates the PK parameters to corrupted image-time series obtained due to subsampling in k-space. This allows the network to directly exploit the knowledge of true contrast agent kinetics in the training phase, and hence provide more accurate restoration of PK parameters. Experiments on clinical brain DCE datasets demonstrate the efficacy of our approach in terms of fidelity of PK parameter reconstruction and significantly faster parameter inference compared to a model-based iterative reconstruction method.



The research leading to these results has received funding from the European Unions H2020 Framework Programme (H2020-MSCA-ITN-2014) under grant agreement no 642685 MacSeNet. We acknowledge Wellcome Trust (Grant 353 088134/Z/09/A) for recruitment and MRI scanning costs. We also gratefully acknowledge the support of NVIDIA Corporation with the donation of the GeForce Titan Xp GPU used for this research.


  1. 1.
    Fang, R., et al.: Direct estimation of permeability maps for low-dose CT perfusion. In: IEEE ISBI, pp. 739–742, April 2016Google Scholar
  2. 2.
    Guo, Y., et al.: High-resolution whole-brain DCE-MRI using constrained reconstruction. Med. Phys. 43(5), 2013–2023 (2016)CrossRefGoogle Scholar
  3. 3.
    Guo, Y., et al.: Direct estimation of tracer-kinetic parameter maps from highly undersampled brain dynamic contrast enhanced MRI. MRM 78(4), 1566–1578 (2017)CrossRefGoogle Scholar
  4. 4.
    Heye, A.K., et al.: Tracer kinetic modelling for DCE-MRI quantification of subtle bloodbrain barrier permeability. NeuroImage 125, 446–455 (2016)CrossRefGoogle Scholar
  5. 5.
    Kamnitsas, K., et al.: Efficient multi-scale 3d CNN with fully connected CRF for accurate brain lesion segmentation. Med. Image Anal. 36, 61–78 (2017)CrossRefGoogle Scholar
  6. 6.
    Lebel, R.M., et al.: Highly accelerated dynamic contrast enhanced imaging. MRM 71(2), 635–644 (2014)CrossRefGoogle Scholar
  7. 7.
    O’Connor, J.P.B., et al.: Dynamic contrast-enhanced MRI in clinical trials of antivascular therapies. Nat. Rev. Clin. Oncol. 9(3), 167–77 (2012)CrossRefGoogle Scholar
  8. 8.
    Patlak, C.S., et al.: Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. J. Cereb. Blood Flow Metab. 3(1), 1–7 (1983)CrossRefGoogle Scholar
  9. 9.
    Sourbron, S.P., Buckley, D.L.: Classic models for dynamic contrast-enhanced MRI. NMR Biomed. 26(8), 1004–1027 (2013)CrossRefGoogle Scholar
  10. 10.
    Zhang, K., et al.: Beyond a gaussian denoiser: residual learning of deep CNN for image denoising. IEEE Trans. Image Process. 26(7), 3142–3155 (2017)MathSciNetCrossRefGoogle Scholar
  11. 11.
    Zhang, T., et al.: Fast pediatric 3d free-breathing abdominal dynamic contrast enhanced MRI with high spatiotemporal resolution. JMRI 41(2), 460–473 (2015)CrossRefGoogle Scholar
  12. 12.
    Zhu, Y., et al.: GOCART: GOlden-angle CArtesian randomized time-resolved 3D MRI. Magn. Reson. Imag. 34(7), 940–950 (2016)CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Cagdas Ulas
    • 1
    Email author
  • Giles Tetteh
    • 1
  • Michael J. Thrippleton
    • 2
  • Paul A. Armitage
    • 4
  • Stephen D. Makin
    • 2
  • Joanna M. Wardlaw
    • 2
  • Mike E. Davies
    • 3
  • Bjoern H. Menze
    • 1
  1. 1.Department of Computer ScienceTechnische Universität MünchenMunichGermany
  2. 2.Department of Neuroimaging SciencesUniversity of EdinburghEdinburghUK
  3. 3.Institute for Digital CommunicationsUniversity of EdinburghEdinburghUK
  4. 4.Department of Cardiovascular SciencesUniversity of SheffieldSheffieldUK

Personalised recommendations