Photo: Radovanovic96/Getty Images
Sydney-headquartered The George Institute for Global Health has developed a deep learning algorithm that can predict major cardiac events using mammograms.
WHAT IT'S ABOUT
The AI model, developed in collaboration with the University of New South Wales and the University of Sydney, was trained and validated using routine mammograms collected from more than 49,000 women in Victoria, Australia.
Its prediction is based on various mammographic features, including breast arterial calcification, and age. It is claimed to be the first-of-its-kind model that is less resource-intensive to implement yet remains highly accurate, according to Dr Clare Arnott, associate professor and global director of the Cardiovascular Program at The George Institute.
In a study published in Heart, the deep learning model was found to have comparable accuracy to traditional cardiovascular risk calculators.
"We found that our model performed just as well without the need for extensive clinical and medical data," said Dr Arnott.
WHY IT MATTERS
Cardiovascular disease accounts for a third of all deaths among women worldwide each year. Yet despite this heavy burden, symptoms and risk factors in women are often underrecognised, resulting in fewer diagnostic tests, specialist referrals, and prescriptions than in men. At the same time, participation in breast screening programs has been increasing, with engagement reaching 67% of women in the United States and the United Kingdom.
Since many are still not being reached by or offered cardiovascular risk screening in their communities, Dr Arnott stressed the need for new approaches to identify women at high risk of CVD.
"It's a common misconception that CVD predominantly affects men, resulting in underdiagnosis and undertreatment of the condition in women," she said.
With their latest AI development, Dr Arnott expects that breast diseases and CVD could be identified and potentially prevented at the same time.
The George Institute researchers are now planning to test the AI in more diverse populations and identify potential barriers to its implementation.
MARKET SNAPSHOT
In the area of AI mammography, South Korean company Lunit has emerged as a leader, with supply deals reported in multiple countries. Recently, it signed contracts to provide its AI solution to support breast screening programmes in Sweden, Australia, and Singapore.
Another player in this space, FathomX from Singapore, testbedded its AI tool with the National University Health System.
