A Multi-task Learning Approach for Compartmental Model Parameter Estimation in DCE-CT Sequences

  • Blandine Romain
  • Véronique Letort
  • Olivier Lucidarme
  • Laurence Rouet
  • Florence d’Alché-Buc
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8150)


Today’s follow-up of patients presenting abdominal tumors is generally performed through acquisition of dynamic sequences of contrast-enhanced CT. Estimating parameters of appropriate models of contrast intake diffusion through tissues should help characterizing the tumor physiology, but is impeded by the high level of noise inherent to the acquisition conditions. To improve the quality of estimation, we consider parameter estimation in voxels as a multi-task learning problem (one task per voxel) that takes advantage from the similarity between two tasks. We introduce a temporal similarity between tasks based on a robust distance between observed contrast-intake profiles of intensity. Using synthetic images, we compare multi-task learning using this temporal similarity, a spatial similarity and a single-task learning. The similarities based on temporal profiles are shown to bring significant improvements compared to the spatial one. Results on real CT sequences also confirm the relevance of the approach.


Multi-task learning CT perfusion model parameter estimation 


  1. 1.
    Miles, K., et al.: Current Status and Guidelines for the Assessment of Tumour Vascular Support with Dynamic Contrast-Enhanced Computed Tomography. Eur. Radiol. 22(7), 1430–1441 (2012)CrossRefGoogle Scholar
  2. 2.
    Sourbron, S.P., Buckley, D.L.: Tracer Kinetic Modelling in MRI: Estimating Perfusion and Capillary permeability. Phys. Med. Biol. 57(2), R1–R33 (2012)Google Scholar
  3. 3.
    Schmid, V., Whitcher, B., et al.: Bayesian Methods for Pharmacokinetic Models in Dynamic Contrast-Enhanced Magnetic Resonance Imaging. IEEE Trans. Med. Imaging 25(12), 1627–1636 (2006)CrossRefGoogle Scholar
  4. 4.
    Kelm, B.M., Menze, B.H., et al.: Estimating Kinetic Parameter Maps from Dynamic Contrast-Enhanced MRI using Spatial Prior Knowledge. IEEE Trans. Med. Imaging 28(10), 1534–1547 (2009)CrossRefGoogle Scholar
  5. 5.
    Brix, G., Griebel, J., et al.: Tracer Kinetic Modelling of Tumour Angiogenesis Based on Dynamic Contrast-Enhanced CT and MRI Measurements. Eur. J. Nucl. Med. Mol. Imaging 37(suppl. 1), S30–S51 (2010)Google Scholar
  6. 6.
    Raue, A., Kreutz, C., et al.: Addressing Parameter Identifiability by Model-Based Experimentation. IET Syst. Biol. 5(2), 120–130 (2011)CrossRefGoogle Scholar
  7. 7.
    Evgeniou, T., Pontil, M.: Regularized Multi-Task Learning. In: Proceedings of the Tenth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2004, pp. 109–117. ACM, New York (2004)CrossRefGoogle Scholar
  8. 8.
    Forsgren, A., Gill, P., et al.: Interior Methods for Nonlinear Optimization. SIAM Review 44(4), 525–597 (2002)MathSciNetCrossRefzbMATHGoogle Scholar
  9. 9.
    Tofts, P.S., Brix, G., et al.: Estimating Kinetic Parameters from Dynamic Contrast-Enhanced T1-Weighted MRI of a Diffusable Tracer: Standardized Quantities and Symbols. Journal of Magn. Reson. Imaging 10(3), 223–232 (1999)CrossRefGoogle Scholar
  10. 10.
    Golub, G.H., Heath, M.T., et al.: Generalized Cross-Validation as a Method for Choosing a Good Ridge Parameter. Technometrics 21, 215–223 (1979)MathSciNetCrossRefzbMATHGoogle Scholar
  11. 11.
    Romain, B., Letort, V., Lucidarme, O., d’Alché-Buc, F., Rouet, L.: Registration of Free-Breathing Abdominal 3D Contrast-Enhanced CT. In: Yoshida, H., Hawkes, D., Vannier, M.W. (eds.) MMCP 2012 and CCAAI 2012. LNCS, vol. 7601, pp. 274–282. Springer, Heidelberg (2012)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  • Blandine Romain
    • 1
    • 2
    • 5
  • Véronique Letort
    • 1
  • Olivier Lucidarme
    • 3
  • Laurence Rouet
    • 2
  • Florence d’Alché-Buc
    • 4
    • 5
  1. 1.MASEcole Centrale ParisChatenay-MalabryFrance
  2. 2.Philips Research, SuresnesFrance
  3. 3.Hospital La Pitie-Salpetriere, AP-HPParisFrance
  4. 4.INRIA-Saclay, LRI CNRS 8623OrsayFrance
  5. 5.IBISCUniversity of EvryEvryFrance

Personalised recommendations