AI Can Predict Breast Cancer Risk Years Before Diagnosis, Study Finds
Artificial intelligence has demonstrated the ability to identify women at an elevated risk of developing breast cancer several years before diagnosis, according to a study by the Norwegian Institute of Public Health (FHI).
Researchers from FHI, the University of California, and the University of Washington utilized a commercially available AI algorithm to analyze mammograms from 116,495 women who participated in a Norwegian breast cancer screening program between 2004 and 2018. Out of these, 1,607 women were later diagnosed with breast cancer.
The AI model successfully predicted which women were more likely to develop breast cancer and even identified which breast was at higher risk, with predictions made four to six years in advance.
“We noted that the breast which developed cancer had an AI score about twice as high as the other breast,” said Solveig Hofvind, head of the detection program and the AI research project. She emphasized that existing AI algorithms could enhance personalized cancer screening programs and improve early detection strategies.
The findings suggest that AI could revolutionize breast cancer detection, reducing healthcare costs and allowing for better targeting of at-risk populations.
According to the World Health Organization (WHO), 670,000 women died of breast cancer in 2022, making it the most common cancer among women worldwide.
The study, published in the Journal of the American Medical Association (JAMA) Network, underscores AI’s growing role in early cancer detection and precision medicine. Additionally, Norway’s breast cancer screening program has initiated a new study involving 140,000 women to evaluate whether AI can match or even surpass radiologists in diagnosing breast cancer.
