人工智慧揭示乳癌隱藏線索，幫助更聰明、更個人化的治療

Researchers at the University of Southampton have developed a new artificial intelligence tool that can find hidden patterns inside breast cancer cells. Working with the local hospital, the team used this technology to study over 330,000 cells from 127 patients. Their findings could help doctors spot patients who are at higher risk and create more targeted treatments for them. Dr Salah Elias, who led the research, said that this discovery could lead to new ways to identify cancer and make treatments more personal for each patient.

The scientists used a microscope to look at samples of breast cancer tissue. They used special stains to color different parts of the cells, which helped them see details that are usually hard to spot. They found two types of changes in the cancer cells. One type was when cells had too many centrosomes, which are parts of the cell that help it divide. The other type was when centrosomes became much larger than normal. These two changes can happen in different areas of the same tumor and might affect how the cancer grows or spreads.

To study these changes, the team created an AI platform called CenSegNet. This tool can examine tumor samples much faster and more accurately than before. The researchers discovered that tumors with more enlarged centrosomes were linked to more aggressive cancer. These tumors were more likely to have spread to nearby lymph nodes and had certain genetic changes. Patients who had fewer enlarged centrosomes in the center of their tumors tended to survive longer. The scientists believe that CenSegNet could help doctors understand which cancers are more likely to grow, spread, or resist treatment. It could also help researchers develop new drugs that target specific weaknesses in tumors. Although the technology is not yet used in hospitals, it is available for free to other scientists, and it can also be used to study other types of cancer, such as kidney, colon, and appendix cancers.