Artificial Intelligence Risk Model Validated for Predicting Breast Cancer

THURSDAY, Jan. 4, 2024 (HealthDay News) -- Across European populations, an artificial intelligence (AI) risk model shows generalizable discriminatory performance for predicting breast cancer among women with a negative mammogram, according to a study published online Dec. 6 in The Lancet Regional Health: Europe.

Mikael Eriksson, Ph.D., from the Karolinska Institutet in Stockholm, and colleagues used four European screening populations in three countries screened between 2009 and 2020 for women aged 45 to 69 years to assess the predictive performance of an AI-based risk model. Data were included for 739 women with incident breast cancers and 7,812 controls. Mammographic features were extracted using AI from negative digital mammograms. The two-year absolute risks for breast cancer were predicted and examined after two-year follow-up.

The researchers found that for breast cancers developed in four screening populations, the overall adjusted area under the receiver operating characteristic curve of the AI risk model was 0.72. Cancers were more likely diagnosed after two years of follow-up in the 6.2 percent of women at high risk using the National Institute of Health and Care Excellence guideline thresholds compared with the 69 percent of women classified as general risk by the model (risk ratio, 6.7). Across levels of mammographic density, similar risk ratios were observed.

"An image-derived AI model is feasible for personalized breast cancer screening to improve the screening outcomes," the authors write.

One author has a patent on a system and method for assessing breast cancer risk using imagery with a license to iCAD.

Abstract/Full Text