A new artificial intelligence (AI) tool has been shown to significantly improve the detection of thoracic aortic aneurysms, potentially life-threatening heart vessel enlargements, during routine breast MRI scans, offering a critical opportunity to improve outcomes for women.
Researchers retrospectively validated a fully automated artificial neural network (ANN) using over 5,000 breast MRI scans from both public (Duke University Hospital/EA1141 trial) and in-house (Erlangen University Hospital) datasets. The AI model demonstrated high accuracy in identifying aneurysms, with strong agreement to expert annotations and clinical reports (dice scores between 0.88–0.91; clDice 0.97–0.99) across varying imaging systems.
Aneurysm detection rates improved by 3.5 times compared to routine clinical readings, suggesting the ANN can help catch these silent yet deadly conditions much earlier, particularly important as thoracic aortic aneurysms often go undetected until they rupture.
The study also revealed that women with breast cancer or a history of the disease had over twice the odds (OR = 2.29) of having an aortic aneurysm, strengthening the case for dual-purpose screening.
Given the poorer prognosis for women with aneurysms, integrating this AI tool into standard breast MRI workflows could transform early detection efforts, enabling timely intervention and ultimately saving lives.
Reference
Bounias D et al. AI-Based screening for thoracic aortic aneurysms in routine breast MRI. Nat Commun. 2025;16(1):5299.
