AI enabled echocardiogram software
The EchoGo suite
The EchoGo suite provides AI enabled echocardiogram software to support physicians in detecting cardiovascular disease. Echocardiogram images are automatically sent to Ultromics for analysis and artificial intelligence assists in identifying suspected heart disease, sending a report back within minutes. The system is zero click and has zero variability.
Automated calculations for EF, GLS and LV Volumes
Automated prediction for coronary artery disease
EF, GLS and LV Volumes with Zero Variability
EchoGo Core is the automated solution for cardiovascular findings – calculating Ejection Fraction, Global Longitudinal Strain and Left Ventricular Volumes.
Provides physicians at any skill level with highly reproducible, vendor-neutral analysis which can enhance confidence in making diagnostic decisions.
Performance testing and validation
Accuracy and precision of EchoGo were assessed against EF, GLS and LV Volumes produced by other commercially available software that required manual contouring. Images were processed by five trained professionally accredited operators*. EchoGo was shown to provide accurate and robust measurements of EF, GLS and LV volumes. It also showed zero variability.
*BSE accredited echocardiographers & board-certified cardiologists
Strain reimbursed from January 2020
EchoGo Core allows physicians to measure strain in an automated means with zero variability, which could increase reimbursement and save time.
Speak to us for assistance with setting up strain
Coronary Artery Disease Prediction
Outcomes-based AI system that predicts coronary artery disease, before it becomes acute.
Performance testing and validation
EchoGo Pro was trialled in the UK and US and achieved a diagnostic performance of over 90% (AUROC), significantly reducing the number of misdiagnoses compared to reports of routine clinical practice. The technology has been validated and developed through its partnership in the EVAREST trial with the NHS, one of the largest ultrasound programmes in the world. It’s trained on past clinical examples, to spot thousands of features, compared to a handful of indicators in traditional visual inspection. The EVAREST validation trial continues to run in over 30 NHS hospitals and has recruited over 6000 patients to date.
Performance of the coronary artery disease classifier:
EchoGo Pro uses novel features derived from Stress Echo images and machine learning algorithms. The system was tested on an independent validation dataset demonstrating an area under the ROC curve of 0.927.
In a reader study to evaluate the benefit of EchoGo Pro in clinical settings, clinicians performed significantly better by diagnostic accuracy, with the aid of EchoGo Pro.
The pipeline of AI algorithms underpinning EchoGo was built in partnership with the NHS in the United Kingdom and leading cardiology clinical centres in the United States, all of which have expertise in echocardiography, vascular assessment and clinical physiology. These have provided access to large imaging archives suitable for training and validation of EchoGo.
The system consists of several fundamental computational phases:
Convolutional neural networks (CNN) applied to the automation of left ventricle (LV) segmentation
Models were created using thousands of echocardiographic images in the two and four chamber views
Cardiac cycles are determined using a complex algorithm to detect the correct end-diastolic (ED) and end-systolic (ES) frames
Learn how EchoGo could benefit your medical practice
Get started with zero up front costs
EchoGo Pro is limited by United States law to investigational use
EchoGo Core is not CE Marked for use in the European Union