Innovation & Entrepreneurship

Innovation & Entrepreneurship

Revolutionising cardiovascular disease diagnostics with new AI technology


Ross Upton, CEO and Co-Academic Founder of Ultromics, discusses the £10 million grant the company has received for its new AI technology, Topological Analysis, and the potential for the future use of AI in hospitals


Medical imaging is frequently used in modern medicine as a diagnostic tool, but before any sort of diagnosis can actually be made, they must be interpreted by clinicians. In the case of heart disease scans (echocardiograms) especially, not every indication of the disease can be seen by the eye, making it difficult for clinicians to accurately diagnose patients. However, recent developments in AI means that echocardiograms can now be analysed much more accurately than by doctors alone.

Coronary heart disease is a huge problem and is the single most common cause of death in the UK. While there are plenty of Government bodies helping to raise awareness of it, particularly by encouraging people to adopt better lifestyles and diets, one way we know that will help reduce this statistic is by improving the diagnostic tools that identify coronary heart disease. At the moment, the diagnostic accuracy is only 80 per cent, so one in five cases are being misdiagnosed, resulting in patients with coronary heart disease being sent home or those without it undergoing unnecessary surgical procedures.

The role of AI in diagnosis

However, we have been developing a technology to make echocardiogram analysis more accurate by bringing machine-learning, a type of AI, into the process. Traditionally, echocardiograms use sound waves to create an image of the heart, which can then pinpoint abnormalities in order to diagnose a patient. The challenge, however, is that not all of the information in a scan is visible to the human eye. Clinicians can usually pick out five to ten data points and then base a diagnosis on this. To expand the level of information that can be attained from a single scan, we have developed an AI system, the first of its kind, to analyse echocardiograms to a much greater level.

Ross Upton

It has been developed at the John Radcliff hospital and overseen by Paul Leeson, Professor of Cardiovascular Medicine at the University of Oxford and Co-founder of Ultromics. The technology uses Topological Analysis, which has the ability to identify thousands of data points from a single echocardiogram and draws on one of the world’s largest imaging databases of over 120,000 echocardiogram images from 5,000 research participants. The AI algorithms are trained by compiling measurements into a pattern to diagnose a disease, based on the scans of previous patients and their outcomes.

Our aim is to significantly increase diagnostic accuracy in tests widely used for cardiovascular disease. To date we’ve reduced diagnostic errors by more than 50 per cent, and we are working on improvements to reduce that margin even further. Topological Analysis technology is a great example of how clinicians and technology can collaborate to effectively improve healthcare services. Not only can AI technology help potentially save thousands of lives by diagnosing heart conditions more accurately and by ensuring the correct course of treatment is provided much earlier, it also has the potential to save the NHS around £300 million per year – the money spent on procedures not actually required.

Why design more than matters in medtech – KOLs discuss the future

In this next article, McKinsey & Co’s Thomas Nilsson and Benedict Sheppard speak with three leaders about how the discipline is improving medical products and driving innovation.


The future of AI in hospitals and the healthcare system

There are some very exciting developments in medtech using AI, including stroke detection software, which analyses CTA images of the brain and alerts clinicians of any irregularities in vessels, and the diagnostic software OsteoDetect which detects bone fractures in wrists, both of which received recent FDA clearance. The trend of AI-implemented healthcare is only set to rise, with support from ministers such as Sir John Bell, Head of Industrial Strategy for Healthcare. He believes in twenty years’ time healthcare will have AI embedded in a whole variety of different levels and a lot of our healthcare will be enabled by smart systems that help identify people at risk, diagnose disease earlier and more precisely and identify those who will benefit from interventions. Professor Bell believes it will change the whole way the paradigm operates.

Our technology, which we are proud to say was selected by the American Society of Echocardiography for the 2018 Echovation Honor Role, is already being tested in a major multi-centre trial involving six UK cardiology units, and we are planning to roll it out to 20 hospitals in early 2019, giving clinicians as early a chance as possible to test the technology. We received a £10 million investment recently which we plan to use to introduce the technology to the USA market. We are also continuing research and development into more diagnostic aid products for cardiology, such as valve disease, heart failure, heart screening technologies, cardiotoxicity, and heart muscle disease. Just like the AI that sits at the centre of our Topological Analysis technology, we aim to keep learning and to continue to make improvements, which we hope we can all benefit from.

About the author

Journalist and editor Kathryn Reilly has worked in consumer, contract and medical writing for more than 20 years.

Expert Insight

The European public-procurement opportunity: delivering value in medtech

The magic of innovation: Interview with Billy Cohn, MD

The MedTech Roadmap: a medical device’s journey from concept to market

An introduction to Beth Susanne – international pitch coach

Be first to know

Key resources


You're the expert! Write for The Engine or share your articles, papers and research

Add your content

Add your content

Keep informed

Sign up for Ignition, our regular, ideas-packed newsletter

Sign in with social media

or with a username