Innovative Machine Learning Tool Enables Rapid Assessment of Cardiovascular Risk Using Bone Density Scans

Researchers from Edith Cowan University (ECU) in collaboration with the University of Manitoba have developed a groundbreaking machine-learning algorithm that can quickly identify individuals at risk of cardiovascular events, falls, and fractures using routine bone density scans. This innovative technology leverages standard vertebral fracture assessment (VFA) images, often performed during osteoporosis evaluations, to evaluate the extent of abdominal aortic calcification (AAC), a key indicator of cardiovascular health.

The new algorithm significantly speeds up the screening process for AAC, providing results in less than a minute across thousands of images—much faster than the five to six minutes typically required by experienced radiologists. This efficiency enables healthcare providers to pinpoint patients with moderate to high AAC levels, which are linked to elevated risks of heart attack and stroke, even among those who are asymptomatic.

Findings from ECU research fellow Dr. Cassandra Smith revealed that approximately 58% of older adults undergoing routine bone density assessments exhibited moderate to severe AAC. Alarmingly, a quarter of these individuals were unaware of their condition, which places them at a notably higher risk of cardiovascular complications.

This approach presents a valuable opportunity to expand cardiovascular screening in populations often underdiagnosed, particularly women. As Dr. Smith noted, leveraging existing low-radiation bone density machines could facilitate early detection, allowing at-risk individuals to seek appropriate treatment. The study emphasizes that many people with AAC do not show symptoms, making these screenings crucial for timely intervention.

Furthermore, the research uncovered a correlation between higher AAC scores and increased risks of falls and fractures. Dr. Marc Sim explained that arterial calcification can impair vascular health, contributing to instability and a greater likelihood of falls. Interestingly, the analysis demonstrated that vascular health markers like AAC were more indicative of fall risk than traditional factors such as previous falls or bone mineral density.

By integrating this machine-learning tool into routine osteoporosis assessments, clinicians gain comprehensive insight into both skeletal and cardiovascular health, improving risk stratification and preventative care strategies for older populations.

For more detailed information, see the original study published in the Journal of Bone and Mineral Research: [DOI: 10.1093/jbmr/zjae208].

source: https://medicalxpress.com/news/2025-04-machine-algorithm-cardiovascular-click-button.html