AI Blood Test Predicts Heart Disease Risk 15 Years Early

A single blood test, leveraging advances in artificial intelligence, may offer a window into a person’s cardiovascular future, potentially predicting the onset of heart disease up to 15 years before symptoms appear. Researchers at the University of Hong Kong (HKUMed) have developed a tool called CardiOmicScore that analyzes a range of molecular signals in blood to assess risk for six major cardiovascular diseases (CVDs): coronary artery disease, stroke, heart failure, atrial fibrillation, peripheral artery disease, and venous thromboembolism. The findings, published in Nature Communications, represent a shift towards more proactive and personalized cardiovascular risk assessment.

Beyond Traditional Risk Factors

Cardiovascular diseases remain the leading cause of death globally, responsible for approximately 19.8 million fatalities in 2022. Current clinical practice typically relies on evaluating risk factors like age, blood pressure, and smoking history. While these factors are important, they often miss the subtle, early biological changes that precede clinical disease. This is where CardiOmicScore aims to make a difference. Traditional methods also struggle to account for the dynamic impact of lifestyle and environmental factors on heart health.

Polygenic risk scores, which assess genetic predisposition to disease, have gained traction in recent years. However, these scores reflect a person’s baseline genetic risk, which doesn’t change over time. CardiOmicScore, in contrast, aims to capture a person’s current biological state, offering a more responsive and timely assessment.

Decoding Molecular Signals with AI

The HKUMed team developed CardiOmicScore by applying deep learning techniques to integrate multiomics data – genomics, metabolomics, and proteomics – from a large dataset within the UK Biobank. The study analyzed 2,920 circulating proteins and 168 metabolites found in blood samples. These molecules act as indicators of the body’s current status, reflecting changes in the immune system, metabolism, and vascular health. Metabolomic profiling, in particular, has emerged as a powerful tool for understanding the complex interplay of biological processes.

“Genes determine where we start—they define our baseline health risk,” explains Professor Zhang Qingpeng, Associate Professor in the Department of Pharmacology and Pharmacy at HKUMed. “However, proteins and metabolites reflect our current physical health. Our AI tool is designed to decode these complex molecular signals, enabling doctors and patients to identify risks much earlier, which can potentially change the trajectory of disease through timely lifestyle modifications and early prevention.”

Improved Predictive Accuracy

The research demonstrates that CardiOmicScore significantly improves the accuracy of cardiovascular risk prediction compared to conventional polygenic risk scores. When combined with basic clinical information like age and gender, the model enhances the prediction of six common CVDs and can identify elevated risk up to 15 years before symptoms manifest. The tool’s ability to reclassify risk suggests it could be particularly valuable for identifying individuals who might benefit from early intervention.

A Shift Towards Dynamic, Multiomics-Based Precision Medicine

This study represents a move away from a static, gene-centric approach to precision medicine towards a more dynamic, multiomics-based model. Instead of relying solely on genetic predisposition, CardiOmicScore considers a broader range of biological factors that are constantly changing in response to lifestyle and environmental influences. The potential for a simple blood sample to provide a comprehensive cardiovascular risk profile for multiple diseases is a significant step forward.

Professor Zhang emphasizes the team’s goal: “We aim to leverage technology to identify and prevent diseases before they develop. By shifting health management from reactive treatment to proactive prediction and intervention, we aim to create a lasting impact for both public health and individual patient care.”

Understanding the Limitations and Next Steps

While the findings are promising, it’s important to acknowledge the study’s limitations. The research was based on data from the UK Biobank, a large-scale population database, but the participants are primarily of European ancestry. Further research is needed to determine whether CardiOmicScore performs equally well in diverse populations. The model also requires validation in independent datasets and clinical trials to confirm its effectiveness in real-world settings.

The next steps involve refining the model, exploring its potential for early detection of other diseases, and developing strategies for integrating it into clinical practice. Researchers are also investigating ways to make the test more accessible and affordable. The University of Hong Kong team is actively working on collaborations to expand the scope of the research and translate these findings into tangible benefits for patients. Further development and validation will be crucial to ensure the tool’s reliability and clinical utility.

For individuals concerned about their cardiovascular risk, maintaining a healthy lifestyle – including a balanced diet, regular exercise, and avoiding smoking – remains the cornerstone of prevention. Regular check-ups with a qualified healthcare professional are also essential for monitoring risk factors and receiving personalized guidance.

This research offers a glimpse into a future where proactive, data-driven approaches to healthcare can help us stay ahead of disease and live longer, healthier lives.

Journal information: Nature Communications