Ovarian cancer AI can tell how aggressive a woman’s tumour is
Artificial intelligence is helping researchers spot aggressive forms of ovarian cancer.
Yinyin Yuan and colleagues at the Institute of Cancer Research in London built an AI to look for differences in tumour cell shape. It analysed tissue sample images from 514 women with ovarian cancer and found that misshapen nuclei correspond to a more aggressive form of the disease with a survival rate of 15 per cent over five years. That compares with 53 per cent for the standard form.
Human researchers are very good at looking at cells, but it is hard to quantify differences and the process takes a lot of time – hence the use of AI, says Yuan.
However, the test so far is of limited use, says Kevin Elias at the Dana-Farber Cancer Institute in Boston. “It is one thing to tell me a patient is likely to have a poor outcome, but if you are unable to suggest an alternative treatment, it is not that useful,” he says.
AI is increasingly used in cancer research to sift data for patterns that can help us in various ways, like tracking tumour evolution and improving diagnosis.
Yuan and her team will next use AI to look at cancer that resists chemotherapy, to try to develop more targeted treatments.