Artificial intelligence (AI) is opening a new pathway for early detection of cardiovascular disease by analysing images of the eye. Researchers presented a system that assesses cardiovascular risk using retinal images captured during routine eye exams. The results show strong alignment with established risk calculators, suggesting that eye care settings could become an additional entry point for preventive cardiology.
Cardiovascular disease remains the leading cause of death globally. Current risk assessment and diagnosis typically relies on clinical data such as blood pressure, cholesterol levels and lifestyle factors. However, these assessments depend on patients engaging with primary care services. Something that does not happen consistently across populations.
“The missing link is often awareness,” said Michael V. McConnell, clinical professor at Stanford University and lead author of the study. “The retina offers a direct view of vascular health. With AI, we can translate that into actionable insights and help identify people who may otherwise remain undiagnosed.”
From eye exam to risk prediction
The AI system, known as CLAiR and developed by Toku, analyses high-resolution images of the blood vessels at the back of the eye. It has already received Breakthrough Device designation from the U.S. Food and Drug Administration, underlining its potential clinical relevance. The research was presented at the American College of Cardiology Annual Scientific Session (ACC.26).
In a prospective study involving 874 participants aged 40 to 75, the system was evaluated against the standard ASCVD risk calculator. None of the participants had known atherosclerosis or were on lipid-lowering therapy. Using conventional assessment methods, 26 percent were classified as having an elevated 10-year cardiovascular risk (≥7.5%).
The AI-based analysis showed a high level of agreement with these findings. It identified at-risk individuals with a sensitivity of 91.1 percent and a specificity of 86.2 percent. According to the researchers, this demonstrates that retinal imaging, combined with AI, can serve as a reliable, non-invasive screening tool.
Unlike earlier approaches that relied heavily on expert interpretation, CLAiR automates the analysis process. The algorithm has been trained to detect subtle patterns in vascular structure that correlate with cardiovascular risk, enabling scalable deployment in clinical workflows.
Scalable screening, but integration needed
One of the key advantages of this approach is its practicality. Retinal imaging is already widely available in eye care practices and typically takes less than five minutes. The AI system can deliver results in around 30 seconds, with 94 percent of images in the study deemed suitable for analysis across different camera systems.
This opens the door to integrating cardiovascular risk screening into routine eye exams, potentially reaching individuals who do not regularly visit a primary care provider. However, implementation challenges remain.
The researchers emphasise that AI-based screening should complement standard cardiovascular assessments. Equally important is the need for clear referral pathways. Identifying risk in an eye clinic must be followed by appropriate follow-up in primary care to ensure patients receive guideline-based preventive treatment.
Limitations
There are also limitations to consider. The technology is not intended for use in pregnant individuals or patients with advanced eye disease, as these conditions can affect retinal blood vessels. In addition, retinal imaging is not always covered by insurance, which may limit accessibility.
Despite these hurdles, the study highlights the growing role of AI in expanding preventive care beyond traditional settings. By leveraging existing infrastructure in eye care, healthcare systems may be able to identify at-risk populations earlier, and intervene before cardiovascular disease progresses.
Cardiovascular risk assessment from mammography
A couple of weeks ago another, large-scale, study showed that artificial intelligence (AI) can identify women at risk of cardiovascular disease by analysing routine mammography scans. The technology detects breast arterial calcification (BAC), calcium deposits linked to cardiovascular risk, without requiring additional tests. In a cohort of over 123,000 women, higher levels of calcification were strongly associated with increased risk of heart attack, stroke, and other cardiovascular events, even after adjusting for traditional risk factors.
Because mammography is already widely used in breast cancer screening, integrating AI-based BAC analysis could enable large-scale, low-cost cardiovascular risk detection. This approach may help address the underdiagnosis of heart disease in women and support earlier preventive interventions, such as lifestyle changes or medication, using existing healthcare infrastructure.
