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.

ARTIFICIAL INTELLIGENCE

Deep learning links histology, molecular signatures and prognosis in cancer

Nature Cancer volume 1, pages 755–757 (2020)Cite this article

Subjects

Deep learning can be used to predict genomic alterations on the basis of morphological features learned from digital histopathology. Two independent pan-cancer studies now show that automated learning from digital pathology slides and genomics can potentially delineate broader classes of molecular signatures and prognostic associations across cancer types.

In cancer, histologic diagnosis of the disease determines to a large extent the course of treatment or indicates the need for additional testing, such as molecular profiling of the tumor1. Recent studies have demonstrated that histologic subtyping of tumors can potentially be automated through the use of the latest advances in the field of artificial intelligence2,3,4,5,6, more specifically deep learning, although several technical challenges still need to be overcome7. However, in addition to the standard histologic diagnosis, there is a growing demand for molecular profiling of the patients’ tumors, to take advantage of targeted or biomarker-based therapies. For example, targeted treatments have been approved by the US Food and Drug Administration to treat patients with lung cancer who have EGFR mutations8 or patients with melanoma who have BRAF mutations8. Molecular assays can accurately characterize the genotype of a given tumor, but these are not available or accessible to all patients, their turnaround time is relatively long (days to weeks) and their cost can be substantial. Meanwhile, it has been shown that molecular alterations can change the morphology of cells and their microenvironment, and that those morphological differences can be identified visually on whole-slide images9. It has also been shown that specific mutations can potentially be identified in lung cancer3,10, prostate cancer11 and uveal melanoma12, but the extent to which this approach can be generalized in any cancer type has not been studied. In this issue of Nature Cancer, two independent studies by Kather et al.13 and Fu et al.14 provide evidence that the application of deep-learning techniques to histopathology slides obtained from big pan-cancer databases may provide an inexpensive alternative method for quickly assessing a wide range of molecular features, signatures and prognosis.

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

Relevant articles

Open Access articles citing this article.

Access options

Rent or buy this article

Prices vary by article type

from$1.95

to$39.95

Learn more

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Comparison of the two pan-cancer deep-learning-based approaches to unveil genotypes information from whole slide images.

References

  1. El-Deiry, W. S. et al. CA Cancer J. Clin. 69, 305–343 (2019).

    PubMed  PubMed Central  Google Scholar 

  2. Mobadersany, P. et al. Proc. Natl Acad. Sci. USA 115, E2970–E2979 (2018).

    Article  CAS  Google Scholar 

  3. Coudray, N. et al. Nat. Med. 24, 1559–1567 (2018).

    Article  CAS  Google Scholar 

  4. Campanella, G. et al. Nat. Med. 25, 1301–1309 (2019).

    Article  CAS  Google Scholar 

  5. Ehteshami Bejnordi, B. et al. J. Am. Med. Assoc. 318, 2199–2210 (2017).

    Article  Google Scholar 

  6. Topol, E. J. Nat. Med. 25, 44–56 (2019).

    Article  CAS  Google Scholar 

  7. Tizhoosh, H. R. & Pantanowitz, L. J. Pathol. Inform. 9, 38 (2018).

    Article  Google Scholar 

  8. Kazandjian, D. et al. Clin. Cancer Res. 22, 1307–1312 (2016).

    Article  CAS  Google Scholar 

  9. Kather, J. N. et al. Nat. Med. 25, 1054–1056 (2019).

    Article  CAS  Google Scholar 

  10. Sha, L. et al. J. Pathol. Inform. 10, 24 (2019).

    Article  Google Scholar 

  11. Schaumberg, A. J., Rubin, M. A. & Fuchs, T. J. bioRxiv https://doi.org/10.1101/064279 (2018).

    Article  Google Scholar 

  12. Sun, M. et al. Cancers (Basel) 11, 1579 (2019).

    Article  CAS  Google Scholar 

  13. Kather, J. N. et al. Nat. Can. https://doi.org/10.1038/s43018-020-0087-6 (2020).

    Article  Google Scholar 

  14. Fu, Y. et al. Nat. Can. https://doi.org/10.1038/s43018-020-0085-8 (2020).

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Applied Bioinformatics Laboratories, NYU Grossman School of Medicine, New York, NY, USA

    Nicolas Coudray & Aristotelis Tsirigos

  2. Skirball Institute, NYU Grossman School of Medicine, New York, NY, USA

    Nicolas Coudray

  3. Institute for Computational Medicine, NYU Langone Health, New York, NY, USA

    Aristotelis Tsirigos

  4. Department of Pathology, NYU Grossman School of Medicine, New York, NY, USA

    Aristotelis Tsirigos

Authors
  1. Nicolas Coudray
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Aristotelis Tsirigos
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Nicolas Coudray or Aristotelis Tsirigos.

Ethics declarations

Competing interests

The authors declare no competing interests.

Rights and permissions

Reprints and Permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Coudray, N., Tsirigos, A. Deep learning links histology, molecular signatures and prognosis in cancer. Nat Cancer 1, 755–757 (2020). https://doi.org/10.1038/s43018-020-0099-2

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s43018-020-0099-2

This article is cited by

Search

Advanced search

Quick links

Nature Briefing: Cancer

Sign up for the Nature Briefing: Cancer newsletter — what matters in cancer research, free to your inbox weekly.

Get what matters in cancer research, free to your inbox weekly. Sign up for Nature Briefing: Cancer