An international study led by the University of Zurich reports that Artificial Intelligence can assess risk in patients with the most common form of heart attack more accurately than existing tools, potentially enabling more personalized care. The work, published in The Lancet Digital Health, focuses on non-ST-elevation acute coronary syndrome (NSTE-ACS) and suggests that current stratification approaches may miss important nuances in patient risk and treatment benefit. The researchers say the new model, termed GRACE 3.0, could help clinicians decide which patients should receive early invasive procedures and when.
Today, clinicians often rely on the GRACE score to estimate the risk of patients with NSTE-ACS and to determine the optimal timing for catheter-based treatment. While widely used and embedded in international guidelines, this standardized approach does not always capture the complexity of individual patients. The new findings indicate that many patients may be misclassified under current methods, with direct implications for the timing and selection of angiography and stenting.
In the largest NSTE-ACS risk modeling study to date, the team analyzed health data from more than 600,000 patients across 10 countries. The researchers used Artificial Intelligence to re-examine clinical trial data from the landmark VERDICT study, training a model to identify which patients are most likely to benefit from early invasive treatment. According to first author Florian A. Wenzl of the University of Zurich’s Center for Molecular Cardiology, the results were striking: some patients gained substantial benefit from early intervention, while others showed little or no benefit, underscoring the need for a major re-stratification of patient care.
Senior author Thomas F. Lüscher, who conducts research in Zurich and at the Royal Brompton and Harefield hospitals in London, describes GRACE 3.0 as the most advanced and practical tool yet for treating patients with the most common type of heart attacks. The updated score aims to deliver more accurate risk prediction and to guide treatment decisions at the individual level. The authors argue that this approach could reshape future clinical guidelines and help save lives. They hope GRACE 3.0 will provide a simple, validated and Artificial Intelligence powered tool for routine clinical practice, enabling more effective and personalized care for heart attack patients.
