AI-Powered Mammography Shows Promise in Predicting Women’s Cardiovascular Risk
LONDON – A new artificial intelligence algorithm analyzing routine mammography images, combined wiht a woman’s age, demonstrates cardiovascular risk prediction capabilities comparable to established clinical assessment tools, researchers announced today. The findings, stemming from a nine-year study, suggest mammograms could evolve into a dual-purpose screening tool, together detecting breast cancer risk and flagging potential heart problems.
The study, involving a large cohort of women, tracked 3,392 participants who experienced a frist cardiovascular event – including coronary artery disease, heart attack, stroke, or heart failure – over approximately nine years. Researchers found the AI model’s predictive accuracy matched that of widely used risk calculators like the “PREDICT” tool from New Zealand and the American Heart Association’s “ASCVD Risk Estimator Plus” calculator. This breakthrough is important because it offers a non-invasive, readily available method for identifying women at increased risk of cardiovascular disease, a leading cause of death globally.
Prior to the analysis, participants provided detailed medical histories encompassing factors like alcohol consumption, body mass index (BMI), history of diabetes, hypertension, hypercholesterolemia, and anticoagulant use. They also shared information regarding menopause, reproductive history, hormone therapy, and prior treatments like radiotherapy or surgery for cancer – all factors that can influence breast tissue structure. However, the AI model itself operated independently of this detailed clinical data, relying solely on mammography images and age.
“One of the main advantages of the mammography model that we have developed is that it does not require anamnesis or data from additional medical records and that it is indeed based on an existing risk detection process widely used by women,” explained the study authors. “Mammography has the potential to be a double-faced risk assessment tool, offering efficiency gains for both patients and for the health system.”
the research team believes this approach could streamline cardiovascular risk assessment, particularly for women who may not regularly engage with conventional preventative care.Further research will focus on refining the algorithm and exploring its potential for widespread clinical implementation.
