Monitoring Acute Lymphoblastic Leukemia Therapy with Stacked Denoising Autoencoders

  • Jakob Scheithe
  • Roxane LicandroEmail author
  • Paolo Rota
  • Michael Reiter
  • Markus Diem
  • Martin Kampel
Conference paper
Part of the Lecture Notes in Computational Vision and Biomechanics book series (LNCVB, volume 31)


For acute lymphoblastic leukemia treatment monitoring, the ratio of cancerous blood cells, called Minimal Residual Disease (MRD), is in practice assessed manually by experts. Using flow cytometry, single cells are classified as cancerous or healthy, based on a number of measured parameters. This task allows application of machine learning techniques, such as Stacked Denoising Autoencoders (DSAE). Seven different models’ performance in assessing MRD was evaluated. Higher model complexity does not guarantee better performance. For all models, a high number of false positives biases the predicted MRD value. Therefore, cost-sensitive learning is proposed as a way of improving classification performance.



This work was supported by the European Commission FP7-PEOPLE-2013-IAPP 610872 and by ZIT Life Sciences 2014 (1207843). We want to thank Michael Dworzak and Angela Schumich at the Children Cancer Research Center in Vienna for annotating and providing the FCM data.


  1. 1.
    Theunissen P, Mejstrikova E et al (2017) Standardized flow cytometry for highly sensitive MRD measurements in B-cell acute lymphoblastic leukemia. Blood 129(3):347–357CrossRefGoogle Scholar
  2. 2.
    Brüggemann M, Schrauder A, Raff T et al (2008) Standardized MRD quantification in European ALL trials: proceedings of the second international symposium on MRD assessment in Kiel, Germany, 18–20 Sept 2008. (2010) Leuk Off J Leuk Soc Am Leuk Res Fund 24(3):521–535CrossRefGoogle Scholar
  3. 3.
    Greenspan H, van Ginneken B, Summers RM (2016) Guest editorial deep learning in medical imaging: overview and future promise of an exciting new technique. IEEE Trans Med Imaging 35(5):1153–1159CrossRefGoogle Scholar
  4. 4.
    Guo Z, Li X, Xiang et al (2017) Medical image segmentation based on multi-modal convolutional neural network: study on image fusion schemes. arXiv:1711.00049
  5. 5.
    Moeskops P, Viergever M, Mendrik A et al (2016) Automatic segmentation of MR brain images with a convolutional neural network. IEEE Trans Med Imag 35(5):1252–1261CrossRefGoogle Scholar
  6. 6.
    He K, Zhang X, Ren S, Sun J (2014) Spatial pyramid pooling in deep convolutional networks for visual recognition. In: ECCV. SpringerGoogle Scholar
  7. 7.
    Zhang Y, Sohn K, Villegas R, Pan G, Lee H (2015) Improving object detection with deep convolutional networks via bayesian optimization and structured prediction. In: CVPR. IEEE (2015)Google Scholar
  8. 8.
    Schlegl T, Waldstein S, Vogl WD et al (2015) Predicting semantic descriptions from medical images with convolutional neural networks. In: Ourselin S, Alexander DC, Westin CF, Cardoso M (eds) IPMI. Springer International Publishing, vol 9123. pp 437–448 (2015)Google Scholar
  9. 9.
    Vincent P, Larochelle H, Lajoie I et al (2010) Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion. J Mach Learn Res 11:3371–3408MathSciNetzbMATHGoogle Scholar
  10. 10.
    Bengio Y (2009) Learning Deep Architectures for AI. Mach Learn 2(1):1–127MathSciNetCrossRefGoogle Scholar
  11. 11.
    Licandro R, Rota P, Reiter M, Kampel M (2016) Flow Cytometry based automatic MRD assessment in Acute Lymphoblastic Leukaemia: Longitudinal evaluation of time-specific cell population models. In: 14th international workshop on content-based multimedia indexing (CBMI)Google Scholar
  12. 12.
    Ng A (2011) Sparse autoencoder. CS294A Lect Notes 72(2011):1–19Google Scholar
  13. 13.
    Bengio Y, Lamblin P, Popovici D et al (2007) Greedy layer-wise training of deep networks. In: Advances in neural information processing systems, pp 153–160Google Scholar
  14. 14.
    Bishop C (2006) Pattern recognition and machine learning. Information science and statistics, Springer, New YorkGoogle Scholar
  15. 15.
    Vincent P, Larochelle H, Lajoie I et al (2010) Stacked denoising autoencoders: learning useful representations in a deep network with a local denoising criterion. J Machi Learn Res 11(Dec):3371–3408Google Scholar
  16. 16.
    Bergstra J (2018) Hyperopt—Distributed asynchronous hyperparameter optimization in python. Accessed 5 Mar 2018
  17. 17.
    Bergstra J, Yamins D, Cox D Making a science of model search. arXiv:1209.5111
  18. 18.
    Bergstra J, Bardenet R, Bengio Y, Kgl B (2011) Algorithms for hyper-parameter optimization. In: Advances in neural information processing systems pp. 2546–2554Google Scholar
  19. 19.
    Han J, Kamber M (2011) Data mining: concepts and techniques, 3rd edn. Elsevier, Burlington, MAGoogle Scholar
  20. 20.
    Karsa M, Dalla Pozza L, Venn N et al (2013) Improving the identification of high risk precursor B acute lymphoblastic leukemia patients with earlier quantification of minimal residual disease. PLoS ONE 8(10):e76455CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Jakob Scheithe
    • 1
  • Roxane Licandro
    • 1
    • 2
    Email author
  • Paolo Rota
    • 3
  • Michael Reiter
    • 1
  • Markus Diem
    • 1
  • Martin Kampel
    • 1
  1. 1.Computer Vision LabInstitute of Visual Computing and Human-Centered Technology, TU WienViennaAustria
  2. 2.Computational Imaging Research Lab, Department Biomedical Imaging and Image-guided TherapyMedical University of ViennaViennaAustria
  3. 3.Central Research Lab, Pattern Analysis and Computer VisionIstituto Italiano di TecnologiaGenovaItaly

Personalised recommendations