Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Deep learning of circulating tumour cells

Abstract

Circulating tumour cells (CTCs) found in the blood of cancer patients are a promising biomarker in precision medicine. However, their use is currently hindered by their low frequency, tedious manual scoring and extensive cell heterogeneities. Those challenges limit the effectiveness of classical machine-learning methods for automated CTC analysis. Here, we combine autoencoding convolutional neural networks with advanced visualization techniques. This provides a very informative view on the data that opens the way for new biomedical research questions. We unravel hidden information in the raw image data of fluorescent images of blood samples enriched for CTCs. Our network classifies fluorescent images of single cells in five different classes with an accuracy, sensitivity and specificity of over 96%, and the obtained CTC counts predict the overall survival of cancer patients as well as state-of-the-art manual counts. Moreover, our network excelled in identifying different important subclasses of objects. Deep learning was faster and superior to classical image analysis approaches and enabled the identification of new biological phenomena.

Your institute does not have access to this article

Relevant articles

Open Access articles citing this article.

Access options

Buy article

Get time limited or full article access on ReadCube.

$32.00

All prices are NET prices.

Fig. 1: Overview of the analysis workflow used.
Fig. 2: Comparison of the t-SNE maps of the latent space for a standard and an autoencoder CNN.
Fig. 3: Identification of new cell classes based on the 2D t-SNE map.
Fig. 4: Validation of the classification.
Fig. 5: Morphological heterogeneity of CTCs.
Fig. 6: Kaplan–Meier plots comparing the predictive value of CellSearch, standard CNN and autoencoder CNN scores.
Fig. 7: Overview of the different results obtained with our framework.
Fig. 8: Visualization of the CNN architecture used.

Data availability

A subset of our training and validation set based only on cell line samples and the corresponding weight set is available at https://github.com/LeonieZ/DLofCTCs. Patient data from clinical studies cannot be shared publicly. Raw data to reproduce Figs. 2, 3, 4, 5 and 6 can be shared upon request.

Code availability

The accompanying code for this manuscript is publicly available at https://github.com/LeonieZ/DLofCTCs.

References

  1. Hayes, D. F. et al. Circulating tumor cells at each follow-up time point during therapy of metastatic breast cancer patients predict progression-free and overall survival. Clin. Cancer Res. 12, 4218–4224 (2006).

    Article  Google Scholar 

  2. Cohen, S. J. et al. Relationship of circulating tumor cells to tumor response, progression-free survival, and overall survival in patients with metastatic colorectal cancer. J. Clin. Oncol. 26, 3213–3221 (2008).

    Article  Google Scholar 

  3. de Bono, J. S. et al. Circulating tumor cells predict survival benefit from treatment in metastatic castration-resistant prostate cancer. Clin. Cancer Res. 14, 6302–6309 (2008).

    Article  Google Scholar 

  4. Tol, J. et al. Circulating tumour cells early predict progression-free and overall survival in advanced colorectal cancer patients treated with chemotherapy and targeted agents. Ann. Oncol. 21, 1006–1012 (2010).

    Article  Google Scholar 

  5. Coumans, F. A., Ligthart, S. T. & Terstappen, L. W. Interpretation of changes in circulating tumor cell counts. Transl. Oncol. 5, 486–491 (2012).

    Article  Google Scholar 

  6. Ignatiadis, M. et al. International study on inter-reader variability for circulating tumor cells in breast cancer. Breast Cancer Res. 16, R43 (2014).

    Article  Google Scholar 

  7. Kraan, J. et al. External quality assurance of circulating tumor cell enumeration using the CellSearch system: a feasibility study. Cytometry B 80, 112–118 (2011).

    Article  Google Scholar 

  8. Zeune, L. et al. Quantifying HER-2 expression on circulating tumor cells by ACCEPT. PLoS One 12, e0186562 (2017).

    Article  Google Scholar 

  9. Zeune, L. L. et al. How to agree on a CTC: evaluating the consensus in circulating tumor cell scoring. Cytometry A 93, 1202–1206 (2018).

    Article  Google Scholar 

  10. Zeune, L. ACCEPT (GitHub Repository, 2017); https://github.com/LeonieZ/ACCEPT

  11. Zeune, L., van Dalum, G., Terstappen, L. W., van Gils, S. A. & Brune, C. Multiscale segmentation via Bregman distances and nonlinear spectral analysis. SIAM J. Imaging Sci. 10, 111–146 (2017).

    MathSciNet  Article  Google Scholar 

  12. Nanou, A. et al. Circulating tumor cells, tumor-derived extracellular vesicles and plasma cytokeratins in castration-resistant prostate cancer patients. Oncotarget 9, 19283–19293 (2018).

    Article  Google Scholar 

  13. de Wit, S. et al. Classification of cells in CTC-enriched samples by advanced image analysis. Cancers 10, 377 (2018).

    Article  Google Scholar 

  14. Ko, J. et al. Machine learning to detect signatures of disease in liquid biopsies—a user’s guide. Lab Chip 18, 395–405 (2018).

    Article  Google Scholar 

  15. Raub, C. B. & Nehmetallah, G. Holography, machine learning, and cancer cells. Cytometry A 91, 754–756 (2017).

    Article  Google Scholar 

  16. Lannin, T. B., Thege, F. I. & Kirby, B. J. Comparison and optimization of machine learning methods for automated classification of circulating tumor cells. Cytometry A 89, 922–931 (2016).

    Article  Google Scholar 

  17. Svensson, C.-M., Hübler, R. & Figge, M. T. Automated classification of circulating tumor cells and the impact of interobserver variability on classifier training and performance. J. Immunol. Res. 2015, 1–9 (2015).

    Article  Google Scholar 

  18. Mao, Y., Yin, Z. & Schober, J. M. Iteratively training classifiers for circulating tumor cell detection. In 2015 IEEE 12th International Symposium on Biomedical Imaging (ISBI) 190–194 (IEEE, 2015).

  19. Svensson, C.-M., Krusekopf, S., Lücke, J. & Thilo Figge, M. Automated detection of circulating tumor cells with naive Bayesian classifiers. Cytometry A 85, 501–511 (2014).

    Article  Google Scholar 

  20. Chen, C. L. et al. Deep learning in label-free cell classification. Sci. Rep. 6, 21471 (2016).

    Article  Google Scholar 

  21. Mao, Y., Yin, Z. & Schober, J. A deep convolutional neural network trained on representative samples for circulating tumor cell detection. In 2016 IEEE Winter Conference on Applications of Computer Vision (WACV) 1–6 (IEEE, 2016).

  22. LeCun, Y. A., Bengio, Y. & Hinton, G. E. Deep learning. Nature 521, 436–444 (2015).

    Article  Google Scholar 

  23. Litjens, G. et al. A survey on deep learning in medical image analysis. Med. Image Anal. 42, 60–88 (2017).

    Article  Google Scholar 

  24. Ching, T. et al. Opportunities and obstacles for deep learning in biology and medicine. J. R. Soc. Interface 15, 20170387 (2018).

    Article  Google Scholar 

  25. Eulenberg, P. et al. Reconstructing cell cycle and disease progression using deep learning. Nat. Commun. 8, 463 (2017).

    Article  Google Scholar 

  26. Vidal, R., Bruna, J., Giryes, R. & Soatto, S. Mathematics of deep learning. Preprint at https://arxiv.org/pdf/1712.04741.pdf (2017).

  27. Montavon, G., Samek, W. & Müller, K.-R. Methods for interpreting and understanding deep neural networks. Digital Signal Process. 73, 1–15 (2018).

    MathSciNet  Article  Google Scholar 

  28. Van Der Maaten, L. J. P. & Hinton, G. E. Visualizing high-dimensional data using t-SNE. J. Mach. Learn. Res. 9, 2579–2605 (2008).

    MATH  Google Scholar 

  29. Zeiler, M. D. & Fergus, R. Visualizing and understanding convolutional networks. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol. 8689, 818–833 (Springer, 2014).

  30. Bengio, Y., Courville, A. & Vincent, P. Representation learning: a review and new perspectives. IEEE Trans. Pattern Anal. Mach. Intell. 35, 1798–1828 (2013).

    Article  Google Scholar 

  31. Le, L., Patterson, A. & White, M. Supervised autoencoders: improving generalization performance with unsupervised regularizers. In Proceedings of the 32nd International Conference on Neural Information Processing Systems 107–117 (NIPS, 2018).

  32. Zhang, Y., Lee, K. & Lee, H. Augmenting supervised neural networks with unsupervised objectives for large-scale image classification. In Proceedings of the 33rd International Conference on Machine Learning 48, 612–621 (JMLR, 2016).

  33. Cristofanilli, M. et al. Circulating tumor cells, disease progression, and survival in metastatic breast cancer. New Eng. J. Med. 351, 781–791 (2004).

    Article  Google Scholar 

  34. Hinton, G. et al. Deep neural networks for acoustic modeling in speech recognition. IEEE Signal Process. Mag. 29, 82–97 (2012).

    Article  Google Scholar 

  35. Krizhevsky, A., Sutskever, I. & Hinton, G. E. ImageNet classification with deep convolutional neural networks. Adv. Neural Inform. Process. Syst. 25, 1097–1105 (2012).

    Google Scholar 

  36. Ronneberger, O., Fischer, P. & Brox, T. U-Net: convolutional networks for biomedical image segmentation. In International Conference on Medical Image Computing and Computer-Assisted Intervention Vol. 9351, 234–241 (LNCS, 2015).

  37. Vincent, P., Larochelle, H., Bengio, Y. & Manzagol, P.-A. Extracting and composing robust features with denoising autoencoders. In Proceedings of the 25th International Conference on Machine Learning 1096–1103 (ACM, 2008).

  38. Kapil, A. et al. Deep semi supervised generative learning for automated PD-L1 tumor cell scoring on NSCLC tissue needle biopsies. Sci. Rep. 8, 17343 (2018).

    Article  Google Scholar 

  39. Long, J., Shelhamer, E. & Darrell, T. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 3431–3440 (IEEE, 2015).

  40. Erhan, D., Courville, A. & Vincent, P. Why does unsupervised pre-training help deep learning?. J. Mach. Learn. Res. 11, 625–660 (2010).

    MathSciNet  MATH  Google Scholar 

  41. Ghifary, M., Kleijn, W. B., Zhang, M., Balduzzi, D. & Li, W. Deep reconstruction-classification networks for unsupervised domain adaptation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol. 9908, 597–613 (Springer, 2016).

  42. Liu, J., Xu, B., Shen, L., Garibaldi, J & Qiu, G. HEp-2 cell classification based on a deep autoencoding-classification convolutional neural network. In 2017 IEEE 14th International Symposium on Biomedical Imaging 1019–1023 (IEEE, 2017).

  43. Chollet, F. Keras (GitHub Repository, 2015).

  44. Zeiler, M. D. ADADELTA: an adaptive learning rate method. Preprint at https://arxiv.org/pdf/1212.5701.pdf (2012).

  45. Smith, S. L., Kindermans, P.-J., Ying, C. & Le, Q. V. Don’t decay the learning rate, increase the batch size. Preprint at https://arxiv.org/pdf/1711.00489.pdf (2017).

  46. Fehm, T. et al. HER2 status of circulating tumor cells in patients with metastatic breast cancer: a prospective, multicenter trial. Breast Cancer Res. Treat. 124, 403–412 (2010).

    Article  Google Scholar 

  47. Pierga, J. Y. et al. Neoadjuvant bevacizumab, trastuzumab, and chemotherapy for primary inflammatory HER2-positive breast cancer (BEVERLY-2): an open-label, single-arm phase 2 study. Lancet Oncol. 13, 375–384 (2012).

    Article  Google Scholar 

  48. Pierga, J. Y. et al. Pathological response and circulating tumor cell count identifies treated HER2 inflammatory breast cancer patients with excellent prognosis: Beverly-2 survival data. Clin. Cancer Res. 21, 1298–1304 (2015).

    Article  Google Scholar 

  49. Crespo, M. et al. Androgen receptor expression in circulating tumour cells from castration-resistant prostate cancer patients treated with novel endocrine agents. B. J. Cancer 112, 1166–74 (2015).

    Article  Google Scholar 

  50. Fizazi, K. et al. Abiraterone acetate for treatment of metastatic castration-resistant prostate cancer: final overall survival analysis of the COU-AA-301 randomised, double-blind, placebo-controlled phase 3 study. Lancet Oncol. 13, 983–992 (2012).

    Article  Google Scholar 

  51. de Wit, S. et al. The detection of EpCAM+ and EpCAM circulating tumor cells. Sci. Rep. 5, 12270 (2015).

    Article  Google Scholar 

  52. Franken, B. et al. Circulating tumor cells, disease recurrence and survival in newly diagnosed breast cancer. Breast Cancer Res. 14, R133 (2012).

    Article  Google Scholar 

Download references

Acknowledgements

L.Z., S.d.W., K.A. and L.T. acknowledge support by EUFP7 programme #305341 CTCTrap. L.Z., G.v.D., K.A., L.T. and C.B. acknowledge support by IMI EU programme #115749 CANCER-ID. C.B. acknowledges support by EU-H2020 project NoMADS #777826. Y.B. and C.B. acknowledge support by the SACAMIR project of TKI Life Science & Health. A.N. and L.T. acknowledge support by NWO Applied and Engineering Sciences project Cancer-ID #14190.

Author information

Authors and Affiliations

Authors

Contributions

The conception and design of the work described herein was made by L.L.Z., L.W.M.M.T. and C.B. L.L.Z., G.v.D., A.N., S.d.W., K.C.A. and J.F.S. acquired the data and generated the ground truth set, and L.L.Z. and A.N. performed the data analysis and interpretation of the results. The code used in this work was created by L.L.Z. and Y.E.B. The first draft of the paper was written by L.L.Z., L.W.M.M.T. and C.B. L.L.Z., Y.E.B., G.v.D., A.N., S.d.W., K.C.A., S.A.v.G., L.W.M.M.T. and C.B. commented on and edited the manuscript.

Corresponding author

Correspondence to Christoph Brune.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

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

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zeune, L.L., Boink, Y.E., van Dalum, G. et al. Deep learning of circulating tumour cells. Nat Mach Intell 2, 124–133 (2020). https://doi.org/10.1038/s42256-020-0153-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s42256-020-0153-x

Further reading

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing