Machine Learning Diagnoses Heart Attack In A Fast And Accurate Way

Heart attack symptoms are sometimes similar to non-heart-related conditions, making diagnosis tricky. UK researchers have turned to machine learning to provide doctors with a fast and accurate way of diagnosing heart attacks that has the potential to shorten the time needed to make a diagnosis and provide more efficient and effective treatment to patients.

Currently, the gold-star method for diagnosing a heart attack is to measure levels of the protein troponin in the blood. Troponin is released when the heart muscle is damaged; levels usually increase sharply within three to 12 hours after a heart attack, peaking after about 24 hours. Many hospitals worldwide have adopted diagnostic pathways that include assessing troponin levels when someone is admitted with a suspected heart attack. But they have some limitations: they require the fixed-time collection of blood samples which can be a challenge in the emergency department setting; they only categorize patients as being a low, intermediate or high risk of a heart attack without considering other important information such as when the symptoms began or electrocardiogram (ECG) findings; and, they don’t take into account the influence of sex, age and comorbidities.

Now UK researchers have developed an AI-based machine learning algorithm that is fast and accurate. Named the Collaboration for the Diagnosis and Evaluation of Acute Coronary Syndrome (CoDE-ACS), the algorithm was designed to calculate the probability of a heart attack for an individual patient. The researchers used data from 10,286 patients who presented with possible heart attacks across six countries worldwide. The machine learning algorithm was “taught” using the patient’s sex, age, ECG findings and medical history, in addition to troponin levels, to identify the probability that a heart attack had occurred. Compared to existing methods, the researchers found that CoDE-ACS could rule out a heart attack in more than double the number of patients, with an accuracy of 99.6%.

The researchers say that their CoDE-ACS algorithm could prevent unnecessary hospital admissions in patients unlikely to have had a heart attack or those at low risk of suffering heart muscle damage or death following a heart attack. They say this would make emergency treatment more efficient and effective, identifying which patients are safe to go home and which need to stay for further tests.

For patients with acute chest pain due to a heart attack, early diagnosis and treatment saves lives,” said Nicholas Mills, corresponding author of the study. “Unfortunately, many conditions cause these common symptoms, and the diagnosis is not always straightforward. Harnessing data and artificial intelligence to support clinical decisions has enormous potential to improve care for patients and efficiency in our busy Emergency Departments.