Article of the Month – January 2023
January 1, 2023
FDA Updates
January 1, 2023Eko
Overview
Eko puts together specialized stethoscopes, apps for patients and providers, and AI-powered analysis, improving how cardiovascular disease is diagnosed and tracked. Through this cohesive forum, their mission is to help clinicians push cardiovascular management to the next level. FDA-cleared AI analysis algorithms empower clinicians to confidently identify early signs of structural heart disease and atrial fibrillation (AF) without a more costly and time-consuming workup.
Healthcare, Networked
The Eko Platform is an efficient, realistic screening tool that improves cardiopulmonary treatment and adapts to modern needs.
Cutting-edge medicine
Make the most of each exam: Eko broadens the stethoscope’s capabilities to increase the potential for early detection of life-threatening diseases, allowing for the best possible treatment.
Diagnosis is wireless and secure
Using high-fidelity sound, visualized waveforms, wireless auscultation, and ECG, and the most detailed view of cardiopulmonary function, this device allows for greater insight into each patient’s condition.
Capture it
The equipment from Eko produces precise auscultation. Eko allows clinicians access to powerful software and research resources by digitizing the gold-standard test.
Analyze it
The machine learning algorithm suite from Eko equips providers and health systems with a strong new ally for heart disease detection. The software is able to analyze the heart sound and ECG data for heart murmurs, AF, and other hemodynamic parameters suggestive of heart disease.
Share it
As open as it is efficient, AI should be. Providers conveniently handle sounds and ECG tracings with Eko’s smartphone and web apps, share a second opinion, and live stream during telemedicine visits. (1)
Eko Health IT Overview
Eko’s AI is a cloud-based software application that analyzes ECG and heart sound/ phonocardiogram data. The software uses several methods to interpret the acquired signals, including signal processing and convolutional neural networks.
The software is intended to support the physician in evaluating the patient’s heart sounds and ECGs. The software simultaneously analyzes ECG and heart sounds and will detect the presence of suspected murmurs in the heart sounds. It also detects the presence of AF and normal sinus rhythm from the ECG signal. In addition, it calculates certain cardiac time intervals such as heart rate, QRS duration, and electromechanical activation time. However, the software does not differentiate between different kinds of murmurs and does not identify other arrhythmias. The interpretations of heart sounds and ECG offered by the software are only significant when used in conjunction with physician over-read and when used on adults.
Clinical studies showed that Eko AI detected AF with 99% sensitivity and 97% specificity and identified heart murmurs with 87% sensitivity and 87% specificity. Eko AI was trained on real-world heart murmurs sounds and ECGs and validated through multi-site clinical studies.
Eko Telehealth brings the high-fidelity auscultation sounds and ECG that a physician needs to conduct a more comprehensive exam. In addition, Eko AI analysis identifies heart disease by bringing a cardiologist’s ears into every exam. Videoconferencing, collaboration, and sharing tools help the physician to provide best-in-class patient care. The Eko platform is HIPAA compliant and SOC 2 certified to keep patients’ data safe and offers an enlarged administrative control, including single sign-on to guard enterprise patient data.
The platform supports iOS, Android, and Windows. Record, save, and share. Chronicle each patient’s exam history, with ongoing access to stored sounds, waveforms, and ECGs. Share for a second opinion or referral as easily as sending an email.
Eko brings together advanced stethoscopes, patient and provider software, and AI-powered analysis.(2)
Eko health's financial aspects
Eko Health announced $65 million in Series C funding led by Highland Capital Partners and Questa Capital, with participation from Artis Ventures, DigiTx Partners, NTTVC, 3M Ventures, and other new and existing investors. The new funding will expand in-clinic uses of the company’s telehealth platform and AI algorithms for disease screening and launch a monitoring program for cardiopulmonary patients at home. This funding comes on the heels of collaborations the company announced this year with AstraZeneca and 3M, as well as the achievement of product milestones, including FDA clearance of its AI suite and the launch of its telehealth platform.
Currently, its FDA-cleared platform is used by thousands of clinicians treating millions of patients worldwide, in-person and through telehealth. The company is headquartered in Oakland, California.(3)
Handheld Wireless Digital Phonocardiography for Machine Learning-Based Detection of Aortic Stenosis
Aortic stenosis (AS) is a medical condition that can be detected as a murmur on auscultation. Studies show that around 80% of primary care physicians cannot detect AS murmurs confirmed by a transthoracic echocardiogram (TTE). The Eko CORE is a digital stethoscope paired with the Eko mobile application allowing the recording and analysis of phonocardiograms (PCG). The data is analyzed by ML system to identify clinically significant AS. For this study, patients undergoing TTE underwent PCG recording by the Eko CORE. The recordings were obtained at four standard auscultation positions. 161 patients with 639 recordings have been enrolled, with 14 patients (8.7%) having a significant AS on TTE. The Receiving-operating characteristic curve had an area of 0.964, with a sensitivity of 97.2% and a specificity of 86.4% for the detection of AS. (4)
Refining Auscultation of Left Ventricular Assist Devices: Insights from Phonography
Heart auscultation is a dominant part of the physical examination in patients with left ventricular assist devices (LVAD). There have been over 15,000 continuous-flow LVADs implanted in the United States. LVAD’s sound may help clinicians identify possible underlying pathology at the bedside. The HeartMate II is a continuous flow LVAD that uses axial flow technology. The PGN shows a predominantly steady sound without fluctuations in sound amplitude. Respectively, the HeartMate III device is a continuous flow LVAD that operates a centrifugal technology that utilizes artificial pulses every 2 sec. The PGN shows a baseline of crescendo decrescendo sound interspersed by a higher amplitude sound produced by the artificial pulse cycle. (5)
Analysis of Utility of Signals from an ECG-Enabled Stethoscope to Automatically Detect a Low Ejection Fraction Using AI
ECG-enabled stethoscopes (ECG-steth) can register single lead ECG during cardiac auscultation, with the possibility of real-time screening for cardiac pathologies during a regular physical examination. This study aimed to demonstrate that an AI algorithm trained using ECG-12 can be applied to ECG-steth to detect low ejection fraction. One hundred patients referred for echocardiography were included. ECG-steth with the patient supine or sitting were taken in standard positions where cardiac auscultation was done via a hand-held equivalent. The AI system was trained using 35,970 patients with pairs of ECG-12, and echocardiograms were retrained using a single lead and validated against ECG-steth to determine the accuracy for low EF (≤35% or <50%). Results showed that among the 100 patients, 7 had EF ≤35% and 7 had EF 35-50%. The Area Under the Curve was 0.906 for EF ≤35% and 0.841 for EF <50%. (6)
Handheld Wireless Digital Phonocardiography for ML-Based Detection of Mitral Regurgitation
Mitral Regurgitation (MR) is a condition that can be spotted as a murmur on auscultation. However, primary care physicians find it challenging to detect MR at the bedside, requiring the diagnosis to be confirmed by transthoracic echocardiography (TTE). Patients undergoing TTE underwent PGN recordings by an Eko CORE device. The recordings were obtained at four standard auscultation positions. The ML system assessed the presence or absence of murmur with dominant localization to the cardiac apex, indicating clinically significant MR. 234 patients with 626 recordings were enrolled, with 32 patients (13.7%) found to have a significant MR on TTE. The receiver-operating characteristics curve had an area of 0.764, with a sensitivity of 61.5% and a specificity of 86.3%. (7)
References