AI Picks Out ‘Shapeshifting’ Cancer Cells, Revealing Potential New Drug Targets

Researchers in the UK have used artificial intelligence (AI) tools to detect tiny clusters of abnormal cells in biopsy samples from women with ovarian cancer—the presence of these indicating that the disease is more aggressive.


The study published in Nature Communications used AI to analyze ovarian cancer tissue samples from 514 women, looking at the shape of almost 150 million cells in total.

“We have developed a simple new computer test that can identify women with very aggressive ovarian cancer so treatment can be tailored for their needs,” said Dr Yinyin Yuan, team leader in computational pathology at the Institute for Cancer Research and senior author of the study.

The AI tools scans the shape and content of tumor cells, particularly looking at the nuclei, the part of the cell which contains DNA. In most cells, nuclei are vaguely circular or oval-shaped, but the researchers found small patches with misshapen nuclei in some of the ovarian cancer samples.

Women who had these clusters of shapeshifting cells in their tumors had very aggressive disease, with only a 15% five-year survival rate, compared to 53% for patients without these misshapen nuclei.