Researchers at the University of Copenhagen have developed an AI technology that can improve the detection of senescent cells (also known as "zombie cells") in breast tissue, offering a better assessment of breast cancer risk than current clinical models. This development could potentially enhance the treatment and prognosis for women worldwide.

Published in The Lancet Digital Health, the study demonstrates that AI outperforms the traditional Gail model, the current standard for breast cancer risk assessment, in predicting cancer development.

Breast Cancer Facts
Prevalence One of the most common cancers, causing 670,000 deaths globally in 2022
Current Risk Models Primarily use the Gail model for risk prediction
AI Advantages Can detect cellular changes more accurately and offer improved treatment plans

How the AI Predicts Cancer Risk

The new AI technology uses deep learning to analyze mammary tissue biopsies, searching for senescent cells. These cells are metabolically active but no longer divide. Although they can suppress cancer development, they may also contribute to inflammation and tumor growth.

"Millions of biopsies are taken every year, and this technology can help us better identify risks and give women better treatment," says Associate Professor Morten Scheibye-Knudsen, senior author of the study.

Improved Cancer Risk Prediction

By analyzing tissue biopsies for senescent cells, the AI was able to predict cancer risk more effectively than the Gail model. One combination of AI models achieved an odds ratio of 4.70, indicating a fivefold increase in cancer risk prediction accuracy.

"One model combination gave us an odds ratio of 4.70, and that is huge," says Indra Heckenbach, the study's first author.

The Role of 'Zombie Cells'

The AI was trained on senescent cells, which the researchers call "zombie cells." These cells lose some function but are not dead and are closely associated with cancer development. The algorithm assesses cell nuclei shapes, which become more irregular in senescent cells, to predict cancer risk.

Key Features of the AI Model Description
Detection of Senescent Cells Identifies irregularly shaped nuclei linked to cancer development
Better Risk Stratification Improves prediction and allows for tailored treatment protocols
Global Applicability Uses standard biopsy images; can be applied in clinics worldwide

Potential for Improved Screening and Treatment

While not yet available in clinics, the AI has potential for widespread use. It can be used to stratify patients by risk, allowing high-risk individuals to receive more frequent screening while reducing the burden on those at low risk.

"We will be able to use this information to stratify patients by risk and improve treatment and screening protocols," says Scheibye-Knudsen.

Implications for Clinical Practice Potential Benefits
High-Risk Monitoring More frequent mammograms and biopsies to catch cancer earlier
Low-Risk Reduction Reduced frequency of biopsies for low-risk individuals

More Information

  • Study: Deep learning assessment of senescence-associated nuclear morphologies in mammary tissue from healthy female donors to predict future risk of breast cancer: a retrospective cohort study.
  • Authors: Indra Heckenbach et al.
  • Journal: The Lancet Digital Health (2024).
  • DOI: [10.1016/S2589-7500(24)00150-X](www.thelancet.com/journals/lan … (24)00150-X/fulltext)

Provided by the University of Copenhagen