A new AI-powered tool can predict the risk of another heart attack in patients with cancer, helping doctors tailor care and improve outcomes, according to HDR UK and The Lancet
The ONCO-ACS AI tool predicts the risk of repeat heart attacks, major bleeding, or death in cancer patients within six months. Using linked data from over a million patients, it helps doctors personalise care for a vulnerable group previously underserved by older risk scores.
The findings are reported in The Lancet.
Cancer and heart attacks: A complex risk
Cancer patients who suffer a heart attack are at a higher risk of further heart attacks due to a weakened cardiovascular system. The type of cancer can also affect the risk of bleeding, of arterial blood clotting, or both, requiring different treatments for prevention.
A team of researchers from the University of Leicester has developed a new Artificial Intelligence tool, known as ONCO-ACS, that uses machine learning to estimate a patient’s risk of death, major bleeding, or recurrent cardiovascular events within six months of a heart attack, specifically tailored for people with cancer.
AI tool validated with real-world data
The researchers used a linked dataset, Virtual Cardio-Oncology Research Initiative (VICORI), and comparable data from Sweden and Switzerland to analyse outcomes for one million heart attack patients, including over 47,000 with cancer. This comprehensive dataset enabled the development and validation of an AI model uniquely designed to address the added complexity of cancer in heart attack patients, outperforming traditional risk scores that lack cancer-specific insights.
VICORI is a national platform that links electronic health records from across England. Supported by the British Heart Foundation Data Science Centre and Health Data Research UK’s Big Data for Complex Diseases programme, it enables secure, large-scale data analysis for researchers.
ONCO-ACS aims to support cancer patients at a higher risk of a heart attack, using intricate technology that helps clinicians make personalised treatment decisions. The AI tool could help healthcare professionals better identify patients who may require more intensive monitoring or tailored therapies, whilst avoiding unnecessary risks for others.
Senior author Professor David Adlam, an interventional cardiologist from the University of Leicester’s Department of Cardiovascular Sciences and a member of the BDCD consortium, said:“Significant advances in the management of heart disease and cancer alike have created new opportunities for these conditions to coexist. As a result, the growing overlap between cancer and heart attacks will confront cardiologists and oncologists with an increasingly complex patient population. We are addressing this pressing issue through a real-world data perspective.”
