How artificial intelligence can help prevent heart disease misdiagnosis

A US company is implementing artificial intelligence practices to help decrease the number of deaths stemming from not treating heart disease-related incidents accurately and on time.

Coronary heart disease is the leading cause of death in Cyprus, with pre-pandemic data by the World Health Organisation showing that 22.69 per cent of deaths in Cyprus were the result of CAD.

Misdiagnoses of heart disease-related incidents are unfortunately not uncommon, with differing symptoms, inaccurate or vague descriptions by the patients themselves and the presence of comorbidities hindering the process.

In addition, the various tests needed to reach a diagnosis can take time, which might ultimately put the patient in greater discomfort and raise costs for either the patient, the health care provider or both.

Pittsburgh-based company HEARTio seeks to prevent misdiagnoses by applying artificial intelligence to the diagnostic process.

“These people [doctors and nurses] are making life-and-death decisions and they should really be given the maximum amount of information possible for situations like that,” HEARTio cofounder and CEO Utkars Jain told one of their partners in this venture Oracle.

Oracle are the cloud host of the ECGio platform, with the company also working with NSF Innovation Corps, Nvidia’s Inception Programme, and the University of Pittsburgh’s Innovation Insitute.

The ECGio platform is rooted in the study of electrocardiograms (ECG), a common and simple test to check a patient’s heart rhythm and electrical activity produced each time it beats.

Through the study of ECG results, HEARTio developed deep learning algorithms which concentrate on making an individualised prediction by trying to detect arterial stenosis.

While this platform is not yet cleared for commercial use in the United States, the company have stated that they are building a cloud-based tool that can be used by healthcare providers without disrupting their normal operating procedures.

“The HEARTio solution, called ECGio, can take any 12-lead electrocardiogram (ECG) result and, using artificial intelligence (AI), determine whether a patient experiencing pain is at low risk or high risk of coronary artery disease,” senior director and writer at Oracle Margaret Lindquist wrote.

“Armed with this information, physicians are better able to determine the appropriate level and type of treatment for the almost 70 per cent of patients who are misdiagnosed,” Lindquist added.