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.

  • Research Highlight
  • Published:

Deep learning

Artificial intelligence identifies new cancer biomarkers

Nature Reviews Bioengineering volume 1, page 313 (2023)Cite this article

Subjects

Cancer diagnosis and prognosis often rely on the identification of specific tissue biomarkers by pathological analysis of whole slide images. However, only few biomarkers have been clinically validated to show high sensitivity and specificity, in particular, as it takes years of experience for pathologists to discover new biomarkers. Now, writing in Nature Machine Intelligence, Qionghai Dai, Hongfang Yin, Ying Xiao, Lingjie Kong and team developed a deep learning-guided framework that can help pathologists to discover new tissue biomarkers from well-performing deep learning models.

Dai, Yin, Xiao, Kong and team simplified the input of deep learning models by exploring the relationship between multi-class tissue spatial distribution and prognosis, and by disregarding high-resolution information of cell morphology. “We introduced prior medical knowledge to sparsify complex inputs and enable deep learning models to discover biomarkers more accurately and objectively than models that are based on high-density information of pathological images,” says Kong. Using this approach, the researchers have built a biomarker discovery, characterization and verification framework, named PathFinder, based only on the semantic segmentation results of whole slide images. “The PathFinder platform allows AI and pathologists to collaborate in the discovery of new prognostic markers for cancer,” adds Kong.

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

Access options

Buy this article

  • Purchase on SpringerLink
  • Instant access to the full article PDF.

USD 39.95

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

References

Original article

Download references

Author information

Authors and Affiliations

  1. Nature Reviews Bioengineering https://www.nature.com/natrevbioeng/

    Christine-Maria Horejs

Authors
  1. Christine-Maria Horejs
    View author publications

    Search author on:PubMed Google Scholar

Corresponding author

Correspondence to Christine-Maria Horejs.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Horejs, CM. Artificial intelligence identifies new cancer biomarkers. Nat Rev Bioeng 1, 313 (2023). https://doi.org/10.1038/s44222-023-00069-x

Download citation

  • Published:

  • Version of record:

  • Issue date:

  • DOI: https://doi.org/10.1038/s44222-023-00069-x

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