FDA clears AI stethoscope for heart failure detection
Eko Health (CA,USA), a leading innovator in digital health technology for heart and lung disease detection, has announced FDA clearance for its low ejection refraction (EF) detection AI. This cutting-edge AI is integrated into Eko’s digital stethoscopes, allowing healthcare providers to identify low EF—a key indicator of heart failure—in just 15 seconds during a standard physical exam. This advancement promises to significantly improve the early detection of heart failure, a critical factor in enabling effective treatment.
In the US, heart failure affects over six million individuals, with a significant number experiencing heart failure with reduced ejection fraction, characterized by the heart’s decreased efficiency in pumping blood. Traditional methods for detecting heart failure, such as echocardiography, are often not practical in primary care settings due to their cost, the necessity for specialized training, and the time they require, leading to potential delays in diagnosis.
The introduction of Eko’s Low EF AI into routine stethoscope exams could will help to change this by making the detection of low EF more accessible and timely. This method could particularly benefit primary care and frontline health providers by integrating advanced detection capabilities into a familiar and widely used tool.
Paul Friedman, Chair of the Department of Cardiovascular Medicine at Mayo Clinic (UK), noted the utility of such a tool in both urban and remote areas, potentially aiding in the early identification of heart conditions that, if untreated, could lead to hospitalizations or severe health events.
The technology will be part of Eko’s SENSORA™ Cardiac Early Detection Platform, which also features algorithms for identifying atrial fibrillation and structural heart murmurs, often precursors to valvular heart disease. By enabling the detection of Low EF during primary care exams, the platform could expedite the referral process for further diagnostic testing and treatment in cardiology, potentially leading to earlier intervention.
Connor Landgraf, co-founder and CEO of Eko Health, emphasized the transformative potential of incorporating AI into the traditional stethoscope, making it a more powerful tool in detecting heart failure early on.
How it works
Eko’s digital stethoscope produces electrocardiogram recordings from two electrodes on the patient-facing side of the device. These recordings are synchronised with a cloud-based convolutional neural network, which handles the AI interpretation.
The convolutional neural network outputs a continuous value between 0 and 1, indicating the probability of the condition of interest being present (ie, low EF). This information is then relayed to the user interface to aid decision making.
Development and Validation
Eko states their AI was developed using a dataset of over 100,000 electrocardiograms and echocardiograms. It has also been validated through a multi-site, prospective clinical study involving 3,456 patients. It demonstrated an ability to differentiate between low and normal EF levels with an AUROC of 0.835, alongside solid sensitivity and specificity rates.
An independent evaluation by Imperial College London (UK), which was published in Lancet Digital Health, confirmed the AI’s effectiveness, with results prompting its use in over 100 clinics across London and Wales through a collaboration with the UK’s NHS.
The AI’s impact was further highlighted in a study with nearly 1,200 pregnant women in Nigeria, showing its potential in identifying pregnancy-related cardiomyopathy more effectively than standard care.
The FDA clearance of Eko’s Low EF AI is a significant step in integrating advanced diagnostic capabilities into primary care, potentially streamlining the detection and treatment of heart failure.
