A revolutionary AI model can predict heart disease risk in women using mammograms, promising earlier and more accessible screening methods for cardiovascular disease.
A groundbreaking machine learning model developed by The George Institute for Global Health now predicts heart disease risk in women by analyzing mammograms. Published in the journal Heart, this innovative algorithm is a collaboration between The George Institute, the University of New South Wales and the University of Sydney.
Cardiovascular disease (CVD) is the leading cause of death among women worldwide, causing approximately 9 million deaths annually. Historically, diagnosis and treatment of CVD in women have lagged behind men, partly due to the misconception that heart disease predominantly affects men. This disparity has resulted in fewer diagnostic tests, specialist referrals and prescriptions for women.
“It’s a common misconception that CVD predominantly affects men, resulting in underdiagnosis and undertreatment of the condition in women,” Clare Arnott, an associate professor and global director of the cardiovascular program at The George Institute, said in a news release. “By integrating CV risk screening with breast screening through the use of mammograms — something many women already engage with at a stage in life when their cardiovascular risk increases — we can identify and potentially prevent two major causes of illness and death at the same time.”
The deep learning algorithm was trained and validated using routine mammograms from over 49,000 women in both urban and rural areas of Victoria, Australia, which linked to hospital and death records, to assess cardiovascular outcomes.
The researchers then compared the AI model’s performance against traditional risk calculators that require various data points like blood pressure and cholesterol levels.
“We found that our model performed just as well without the need for extensive clinical and medical data,” Arnott added.
This model’s accuracy and efficiency lie in its reliance on various mammographic features combined simply with the patient’s age. Unlike previous methods that focus solely on breast arterial calcification (BAC), which is less accurate in older women, this model uses a range of mammographic features, making it a resource-efficient but highly accurate tool.
“We hope this technology will one day provide greater and more equitable access to screening in rural areas, as many women already benefit from mobile mammography units free of charge,” added Jennifer Barraclough, a research fellow at The George Institute.
Mammography-based screening programs already engage a significant portion of the female population in countries like the United States and the UK, with over 67% participation. Integrating cardiovascular risk prediction into this existing framework could revolutionize preventive health care for women globally.
The researchers behind this groundbreaking algorithm look forward to testing the model in diverse populations to further validate its effectiveness and explore potential barriers to its implementation, according to Barraclough.