A NEW artificial intelligence (AI) tool could be set to improve clinical decision-making for patients with cancer who experienced a heart attack, according to a large international study.
Patients with cancer who present with acute coronary syndrome represent one of the highest-risk groups in cardiovascular medicine. These patients are more likely to die, experience major bleeding, or suffer recurrent ischaemic events than patients without cancer. Despite this, individuals with cancer have historically been excluded from cardiovascular clinical trials, leaving clinicians without validated tools to balance treatment benefits and harms.
AI Tool Addresses Critical Evidence Gap
The newly developed AI tool, known as ONCO-ACS, was designed specifically to predict six-month mortality, major bleeding, and ischaemic events in patients with cancer and acute coronary syndrome. The model was derived from a large real-world dataset comprising 1,017,759 patients treated in England, Sweden, and Switzerland between 2004 and 2023, including more than 47,000 patients with cancer.
Outcomes in this population were strikingly poor. Nearly one in three patients with cancer died within six months of their heart attack. Major bleeding occurred in 7.3% of patients, while 16.1% experienced a recurrent cardiovascular event, defined as myocardial infarction, ischaemic stroke, or cardiovascular death.
ONCO-ACS incorporated tumour type, time since cancer diagnosis, and metastatic disease alongside established cardiovascular variables, acknowledging the heterogeneity of patients with cancer presenting with acute coronary syndrome
Machine learning models demonstrated strong predictive accuracy, with time-dependent area under the receiver operating characteristic curve values at six months of 0.84 for all-cause mortality, 0.70 for major bleeding, and 0.79 for ischaemic events. Importantly, these findings were externally validated across geographically distinct cohorts in all three countries.
Dr Florian A Wenzl (National Health Service England and University of Zurich, Zurich, Switzerland) said: “Cancer patients with heart attacks have long been neglected in clinical research, despite being one of the most challenging groups we see in cardiology. This new tool gives doctors reliable information to tailor treatment and balance the benefits and harms.”
Implications for Cardio-Oncology Practice
This new AI tool could support more individualised decisions regarding invasive coronary procedures and antiplatelet therapy, particularly in patients at competing risk of thrombosis and bleeding. Decision curve analyses suggested favourable clinical utility, and application of the model to current guidelines indicated that many patients with cancer may still benefit from invasive management and prolonged dual antiplatelet therapy.
Researchers suggested that ONCO-ACS could also inform the design of future clinical trials and support the implementation of guideline-recommended care in this historically underserved population.
Reference
Wenzl FA et al. Prediction of mortality, bleeding, and ischaemic events in patients with cancer and acute coronary syndrome. Lancet. 2026;407(10527):515-28
