A groundbreaking advancement in cardiovascular risk detection has emerged from a collaboration between Edith Cowan University (ECU) and the University of Manitoba, unveiling an automated machine learning algorithm capable of rapidly assessing cardiovascular risk through routinely acquired bone density scans. This innovation leverages vertebral fracture assessment (VFA) imagery, customarily used in osteoporosis screenings, to analyze the presence and severity of abdominal aortic calcification (AAC)—a critical marker strongly correlated with cardiovascular disease risk. The integration of machine learning with existing diagnostic imaging modalities marks a transformative step towards efficient, non-invasive, and scalable cardiovascular risk stratification.
Abdominal aortic calcification, a dense build-up of calcium deposits within the abdominal aorta, is an underdiagnosed predictor of major adverse cardiovascular events such as heart attacks and strokes. Despite its clinical significance, AAC screening remains rare due to the need for specialized imaging and expert evaluation. The novel algorithm developed by the research team significantly reduces the time burden on clinicians: it can analyze thousands of images in under a minute, whereas manual scoring by experienced readers typically takes five to six minutes per single image. This remarkable acceleration positions AAC evaluation as a feasible adjunct in widespread osteoporosis screening regimens, allowing dual-purpose clinical imaging with minimal added resource demand.
The algorithm was trained on a large dataset of VFA images, predominantly from older women undergoing routine bone density testing. These individuals frequently undergo scans to assess osteoporosis-related fracture risk, but until now, these images have not been systematically analyzed for vascular health markers. The researchers’ innovative approach repurposes existing diagnostic tools, highlighting how cross-disciplinary technological applications can unlock hidden prognostic information without additional patient burden or cost.
Findings from ECU research fellow Dr. Cassandra Smith revealed striking epidemiological insights: more than half (58%) of older adults screened during routine bone density assessments exhibited moderate to high AAC levels. Perhaps even more alarmingly, approximately 25% of these individuals were completely unaware of their elevated cardiovascular risk status, underscoring an urgent need for improved screening pathways. This under-recognition of cardiovascular risk factors in women, in particular, continues to fuel disparities in prevention strategies and outcomes.
"Incorporating AAC evaluation into bone density testing presents a compelling opportunity to bridge cardiovascular risk assessment gaps in women, who are traditionally under-screened and under-treated for heart disease," Dr. Smith explains. Cardiovascular disease remains the leading cause of death globally, yet female patients often receive less aggressive preventative care. By embedding vascular health evaluation within routine osteoporosis management, clinicians can proactively identify those at highest risk and initiate timely interventions.
A critical clinical challenge in cardiovascular risk management is that AAC accrual is symptomless until severe complications occur. People with calcified abdominal arteries frequently exhibit no overt manifestations, escaping detection during standard health evaluations. The algorithm’s ability to automate and expedite AAC detection directly confronts this clinical blind spot, enabling earlier diagnosis and stratification of patients based on quantifiable calcification scores.
Beyond cardiovascular implications, Dr. Marc Sim, ECU senior research fellow, emphasized the broader prognostic potential of AAC quantification. His work demonstrated that higher AAC scores not only correlate with cardiovascular pathology but also predict heightened risk of fall-induced hospitalizations and fragility fractures. Traditionally, assessments of fall and fracture risk include prior fall history, bone mineral density (BMD), and medication usage, but vascular health seldom factors into this equation. The emerging data suggests a paradigm shift is warranted.
Dr. Sim elaborated, “Our analysis unveiled that vascular calcification is a potent and previously overlooked contributor to fall-related risks, surpassing conventional clinical predictors.” This revelation implies that the vascular system’s integrity significantly influences musculoskeletal stability and injury susceptibility, potentially through mechanisms involving compromised perfusion and microvascular dysfunction in bone and muscle tissues.
The unveiling of this algorithm is notable not only for its technical innovation but also for its potential to reshape clinical workflows. Because VFA imaging is already embedded in osteoporosis screening programs, integrating AAC analysis requires no additional scans or radiation exposure. This seamless incorporation minimizes patient inconvenience while maximizing diagnostic yield, embodying a precision medicine approach that harmonizes cost-effectiveness with clinical impact.
Technically, the underlying machine learning model employs advanced image processing and pattern recognition techniques to delineate calcified regions within the aortic silhouette captured on bone density scans. By training on annotated datasets, the model learns to quantify calcification extent with high reproducibility and accuracy. The algorithm outputs AAC scores that have been strongly validated against expert human interpretation and correlate robustly with known cardiovascular risk markers.
This integration of artificial intelligence into diagnostic radiology represents a cutting-edge trend that stands to democratize access to sophisticated risk assessments worldwide. By automating labor-intensive scoring tasks, the system alleviates bottlenecks in specialist availability, facilitating faster clinical decision-making. Moreover, such algorithms can continuously improve through ongoing machine learning with accumulating datasets, enhancing precision over time.
From a public health perspective, widespread adoption of this technology could dramatically alter the landscape of cardiovascular disease prevention. Early identification of high-risk individuals, particularly among older women who undergo routine osteoporosis screening, enables targeted preventative therapies such as lifestyle modification, pharmacological intervention, and closer cardiovascular monitoring. This proactive strategy promises to reduce the burden of heart attacks and strokes, thereby improving life quality and reducing healthcare costs.
In summary, the development of a machine learning algorithm capable of extracting clinically actionable AAC scores from routine bone density scans epitomizes the transformative potential of artificial intelligence in medicine. This dual-purpose diagnostic approach enhances cardiovascular risk detection without increasing patient burden, addresses the silent progression of vascular calcification, and offers novel insights into the intersection between vascular health and musculoskeletal stability. Ongoing research will clarify its applicability across diverse populations and clinical settings, but its current validation in a robust Manitoba Bone Mineral Density Registry cohort indicates substantial promise.
As cardiovascular disease remains a paramount global health challenge, innovations like this hold the key to refining personalized risk stratification and ultimately saving lives. The integration of machine learning tools into daily clinical practice heralds a new era where sophisticated diagnostic insights can be garnered "at the click of a button," transforming opportunistic imaging into comprehensive health assessments.
Subject of Research: People
Article Title: Automated abdominal aortic calcification and major adverse cardiovascular events in people undergoing osteoporosis screening: the Manitoba Bone Mineral Density Registry
News Publication Date: 3-Jan-2025
Web References:
- Journal of Bone and Mineral Research article
- Potential cardiovascular incidents
- Fall and fracture risks
References:
Smith, C., Sim, M., et al. (2025). Automated abdominal aortic calcification and major adverse cardiovascular events in people undergoing osteoporosis screening: the Manitoba Bone Mineral Density Registry. Journal of Bone and Mineral Research. DOI: 10.1093/jbmr/zjae208
Keywords: Cardiovascular disease, Risk factors, Algorithms, Medical treatments
