Breast cancer: AI detects women at risk several years before the disease
A team of five FHI researchers and universities from Washington and California has shown that AI can identify women at increased risk of breast cancer, several years before the diagnosis of the disease.
A diagnosis of breast cancer via AI
The study was published in October 2024 in Jama Network and relies on the retrospective analysis of mammography of 116,495 women who participated in a screening program in Norway, between 2004 and 2018. Among the participants, 1,607 had breast cancer. The researchers then used an AI available commercially to analyze this data.
The researchers then set up a risk rating system from these mammograms. Solveig Hofvind, director of the screening program and project manager, explains that AI identifies the risk over a period of four to ten years before the diagnosis but also the breast concerned. “The breast that developed cancer obtained an AI score about twice as high as the other breast”specifies the researcher.
It is therefore an advance for very early and personalized screening for breast cancer. The AI will also allow a reduction in costs and better targeting of populations at risk, according to FHI. For researchers, it is a precious help in the face of the magnitude of this disease.
According to the World Health Organization, it is responsible for 670,000 deaths in 2022. It is the most widespread form of cancer in women in several countries. But note that men are not spared by this disease, this AI advance concerns everyone, without exception.
The Norwegian screening program continues with the launch, in 2024, of a new study with 140,000 women. The research aims to determine if AI can match, even surpass, radiologists to detect breast cancer.
