Researchers at the University of Southampton have developed an open-source AI platform called CenSegNet that analyzes microscopic cell structures in breast cancer tumors to help identify high-risk patients and potential targeted treatments.
Key Findings
- Targeted Biomarker: The AI analyzes centrosomes (structures that help cells divide). It discovered two distinct abnormalities—having too many centrosomes versus having unusually large centrosomes—that play different roles in cancer growth.
- Aggressive Indicators: Tumors with a higher number of enlarged centrosomes were linked to more aggressive disease, including higher tumor grades and spread to lymph nodes.
- Prognosis: Patients with fewer enlarged centrosomes at the center of their tumors showed better overall survival rates.
Scope & Availability
- Analyzed over 330,000 cells from 127 patients.
- Released for free as open-source software for global researchers.
- Tested successfully on tissue samples beyond breast cancer, including kidney, colon, and appendix tissues.
Source: ResearchGate
