II-stethoscope can check the heart for three diseases in 15 seconds.
California-based Eko Health has announced that it has received regulatory approval and is starting to roll out AI-powered technology in the UK that will allow it to check the heart as part of a primary diagnosis. II will allow to detect the presence or absence of three diseases at once in 15 seconds.
The technology will be implemented as part of the TRICORDER programme, funded by the National Institute for Health and Care Research (NIHR). It will be used in primary care practice across the UK. The technology will help specialists in the assessment of heart failure, heart defects and atrial fibrillation.
The Imperial College London team will implement the technology in 100 general practice clinics across the UK. Patients who visit a general practitioner will have a short non-invasive heart examination. This process will use the Eko digital stethoscope and an AI-based application. When AI detects possible signs of heart disease, the GP will be able to quickly initiate further tests and treatment.
Currently, the standard for detecting heart failure is a blood test (NT-proBNP), followed by a lengthy diagnostic procedure.
“Hospitals for heart failure alone cost the UK more than £2 billion a year, and 80% of these diagnoses are made during emergency admissions,” said Dr Nicholas Peters, professor of cardiology at Imperial College London and Imperial College London. Co-authors of the study, Patrick Byachtiger and Mihir Kelshiker, noted that the II stethoscope “will save the health care payer system £2,400 per patient by eliminating the potential need for an emergency department visit.”
Eko’s AI-powered technology combines a digital electrocardiogram (ECG) stethoscope with advanced machine learning algorithms that analyze ECG and heart sound data. TRICORDER uses clinically proven algorithms that detect decreased left ventricular ejection fraction (HFrEF) indicative of heart failure, structural murmurs indicative of heart disease, and atrial fibrillation. The Eko HFrEF algorithm has been independently validated and provides detection sensitivity and specificity approaching 80%.
The structural noise algorithm was also tested in a real-world study presented at the 2023 American Heart Association Scientific Sessions. In the course of the study, it was found that this algorithm more than doubles the detection of heart defects compared to the standard method used by primary care specialists (94.1% versus 41.2%).
Finally, the atrial fibrillation detection algorithm was also clinically validated and found to provide a sensitivity of 100% and a specificity of 96.2%.