AI algorithm ‘could improve heart attack diagnosis’
Researchers in the UK have said that an algorithm developed using artificial intelligence could help doctors decide more quickly if a patient has suffered a heart attack, and alleviate pressure on busy emergency departments.
The team from the University of Edinburgh tested the algorithm – called CoDE-ACS – in more than 10,000 patients presenting at hospital with chest pain, to see how well it performed compared to current testing methods.
The main finding? CoDE-ACS ruled out a heart attack in more than double the number of patients, with an accuracy of 99.6%, which its developers say could greatly reduce hospital admissions. The study is published in the journal Nature Medicine.
The algorithm draws on the cardiac troponin testing routinely carried out when a patient presents with a suspected myocardial infarction, but layers in clinical features such as age, sex or other illnesses that may affect cardiac troponin, as well as the time since symptoms started and electrocardiogram (ECG) readings. The result is expressed as a probability score from 0 to 100 for each patient.
The AI tool performed well regardless of age, sex, or pre-existing health conditions, according to the researchers, with 71% of patients ruled out of having a heart attack with a single cardiac troponin test. Its conclusions were double-checked by experienced clinicians.
“For patients with acute chest pain due to a heart attack, early diagnosis and treatment saves lives,” commented Prof Nicholas Mills, British Heart Foundation (BHF) professor of cardiology at the University of Edinburgh, who led the research.
“Unfortunately, many conditions cause these common symptoms, and the diagnosis is not always straightforward,” he added. “Harnessing data and artificial intelligence to support clinical decisions has enormous potential to improve care for patients and efficiency in our busy emergency departments.”
The team are now running a clinical trial of the algorithm in Scotland – backed by the Wellcome Leap programme for funding “unconventional” research efforts – to see how it may work in routine clinical practice.
Another algorithm used to identify acute myocardial infarction called MI3 was recently put through its paces in a study involving around 11,000 patients, and according to its developers had 100% sensitivity for a heart attack at 30 minutes and also successfully identified patients who could be considered for early discharge.
The team behind CoDE-ACS say, however, that MI3 requires at least two cardiac troponin measurements to estimate probability, which could limit its usefulness, and only takes into account the levels of the biomarker and the age and sex of the patient.
Emergency medicine specialist Prof Steve Goodacre of the University of Sheffield – who was not involved in the CoDE-ACS trial – described the findings as “intriguing,” showing how AI can use complex analysis rather than a simple rule to improve diagnosis.
“This doesn’t (yet) show that we can replace doctors with computers. Experienced clinicians know that diagnosis is a complex business,” he remarked.
“It will be interesting to see how clinicians in the emergency department use this algorithm. What will they do if they think the algorithm has got it wrong? The next stage of the research will hopefully answer that question.”