Cardiovascular disease is the world’s biggest killer and one of our greatest healthcare burdens; accounting for more than 30% of all deaths globally and incurring an economic cost that is expected to exceed $1tn by 2030. In the UK alone around 425 people die every day: one person every three minutes.
Founded in 2017 with the support of OSI, Ultromics’ can quickly deliver highly accurate assessments of heart function that support the diagnosis of coronary artery disease and heart failure, built-in partnership with the U.K.’s National Health Service (NHS).
Echocardiograms uploaded to the cloud are analysed using Ultromics’ artificial intelligence platforms – they are set to revolutionise cardiac disease treatment pathways; helping more people understand their risk of heart disease earlier and with greater accuracy.
“Echocardiography is the most accessible, lowest cost, least-invasive, safest and most convenient means of imaging the heart. With Ultromics’ AI, we have proven that it can now be as accurate and effective as more expensive modalities that are not available to all patients.” – Dr. Ross Upton, Founder and CEO, Ultromics.
“Ross and the team at Ultromics exemplify the new wave of innovators transforming healthcare, and OSI’s mission to commercialise visionary solutions to the world’s big problems. We have supported the company since it span out from Oxford University and then proved its clinical excellence both in the UK with the NHS and abroad. We are now delighted to participate in this significant funding round alongside new investors who bring not only capital but expertise and networks that will drive the next exciting chapter in Ultromics’ growth as it achieves broad adoption globally.” – Alexis Zervoglos, Chief Investment Officer, OSI.
Ultromics have now raised $33m in Series B funding to substantially scale adoption of AI-enabled echocardiograms so that more people can get the care they need and deserve earlier.
We spoke to Dr. Ross Upton, Founder and CEO, to learn more.
Why and when was Ultromics founded?
Ultromics was officially founded in 2017 but it was a long time in the making, we started working on the technology that underpins the company quite some time before that.
Echocardiography is simply an ultrasound of the heart, and it is a wonderful technology. It is cheap, provides images for analysis in real-time and can plug straight into a phone making it portable. This technology isn’t limited to western society, it is available and accessible worldwide. In my opinion, its global accessibility makes it the best imaging technology available anywhere in the world today.
But it does have some significant limitations. Firstly, it is a subjective test that relies on one individuals’ interpretation and secondly it takes a significant amount of time to conduct the analysis – approximately 15-20 minutes per patient. The individual is also required to take manual measurements, used to determine how well the heart in question is functioning. As you can imagine the measurements and subsequent results vary depending on the experience of the individual. An inaccurate measurement results in an incorrect diagnosis and ultimately the wrong treatment. I trained as a sonographer myself and saw these issues first-hand, they really frustrated me. So, I began a PhD specifically designed to address these problems.
Can you tell us about yourselves and the Oxford scientists behind Ultromics?
Ultromics was co-founded by myself and Professor Paul Leeson, my PhD supervisor.
I actually began my academic career studying Zoology, but I then went on to do two masters, one in clinical biochemistry and one in cardiovascular imaging before embarking on a PhD in Cardiovascular medicine. It was during my second masters that I also trained as a sonographer within the NHS; performing echocardiograms and analysing images whilst also studying the data science that underpins the technology.
Professor Paul Leeson, now also Chief Medical Officer for Ultromics, is Professor of Cardiovascular Medicine at Oxford University, and a practicing consultant Cardiologist. He is the world-leading expert in both the academic and clinical side of echocardiography; he has deep academic knowledge of the science that underpins Echo and holds clinics day in, day out at the hospital.
We set out to tackle the problems associated with echocardiograms together.
What real-world problems or unmet needs are you addressing?
The first problem we were aiming to address is the subjective analysis that leads to misdiagnosis. Accurate diagnoses are vital for patients, doctors and healthcare systems and the consequences of misdiagnosis are huge. There are patients sent home who are actually high risk, returning as emergency cases, and on the flip side patients subjected to further investigation and surgery who just don’t need it.
The second problem we were aiming to address is the unnecessary time required to analyse echocardiograms. If you just look at the NHS in London, there are currently more than 20,000 patients waiting for an echocardiogram and it takes 15-20 minutes to analyse each one. That is time you could be imaging more patients and clearing backlogs. It is a very similar story in the US and the problem has been compounded by Covid worldwide – not to mention the increasing need for echocardiograms because of Covid.
Can you tell us about the scientific research and why it was ground-breaking?
The trial that validated our technology started in 2011 at the John Radcliffe in Oxford, it then expanded to 30 hospitals within the NHS and is still ongoing to this day. We collect data from every single patient that undergoes an echocardiogram and then follow them up a year later. We called the trial Everest, and the immense scale of the trial is one of the most ground-breaking aspect of our scientific research.
But two other aspects of the technology really push the boundaries, the first is how AI can predict clinical outcomes and the second is how AI can be used to automate the workflow. Much of the AI outcomes prediction work was done during my PhD, but most of the automation work should be credited to the incredibly talented engineers that joined the team from Oxford University and Kings College London. They were able to create an AI stack by stitching lots of different AI models together to automatically process and analyse echocardiogram images in a way that has never been done before. As a result, we are the only company in the world that can fully automate the analysis of an echocardiograms from start to finish. Without any need for manual intervention.
The reason we are breaking new ground and need to be at the forefront is because Echo images are not like other medical images, they aren’t static, they are moving videos of hearts beating and there are a lot of artefacts present. So, the images are really difficult for computers to decipher. We had to create AI models that had never been created before to be able to process and analyse the scans.
Can you tell us how you got from scientific research to usable technology?
My PhD combined my clinical understanding of Echo and patient outcomes, with my understanding of data science to create an automated diagnostic technology that could fix the inaccuracies of diagnosis.
We wanted a physician to be able to take an image using Echo and then immediately receive a diagnosis. This led to a series of inventive steps, we had to work out how to extract the features from an Echo that would be used to make a diagnosis – we did this using artificial intelligence – and then we had to create the automated measurements needed.
But it was Paul’s incredible database that allowed us to get off the starting blocks, huge volumes of data like that are incredible gifts for PhD students. That database was the foundation of Ultromics and to this day it is still growing and unique in terms of its size and data quality.
I was then presenting our research at a conference when OSI approached and asked how I felt about spinning out a company, they believed our research had real-world potential. It was working with OSI that made us realise what we were creating wasn’t just for publication, it could be applied clinically, if we had the right funding and the right team involved.
The database was our foundation and the inventive steps we took to solve the problem created the IP that formed our company. We were then able to turn our IP into a usable technology in the real-world because of the support and investment from OSI.
How does the cardiovascular risk prediction technology work in practice?
We offer a fully automated software-as-a-service platform, powered by AI, that allows clinicians to make fast, accurate decisions. We can deliver the analysis in minutes, with zero variability, without any disruption to workflow, and can even predict a patient’s risk of developing coronary artery disease.
There are two parts to are our technology, firstly it is an automated measurements platform capable of measuring heart function, and secondly it is a diagnostics platform designed specifically for the diagnosis of coronary artery disease. We are working on creating a second diagnostics platform specific to heart failure and amyloidosis as we speak. In summary, the platform provides a two-pronged approach, it provides automated measurements and automated diagnosis. The automated measurements can be used for any cardiac-related disease, the diagnostic element is specifically designed for coronary artery disease.
What will be the ultimate impact of this technology?
Prevention is better than cure not just for patients but also for health systems like the NHS, by helping to provide the right patients with the right interventions earlier and reducing the volume of unnecessary procedures we will be increasing efficiencies across the board.
I would also love to see echocardiogram and ultrasound used more in the community and earlier in the care pathway, in GP surgeries, pharmacies and even people’s homes, instead of just in hospitals. We can and should make this diagnostic tool more and more accessible – especially as it can just be plugged into a phone.
If we think about cardiovascular disease; 50% of males and 33% of females will suffer at some point in their lives. The prevalence is huge but there is no great mechanism for screening for it, usually the first symptom is shortness of breath and when you do visit your Dr you have to wait ages for a heart scan. And you may not even ever get symptoms before experiencing cardiac problems. Therefore, testing earlier in the pathway is crucial.
That would save the NHS a lot of money; heart failure alone and emergency admissions costs the NHS 2% of its total budget – that is huge. If we could get the technology into the community, we could triage instantly, sending those identified as high risk to a cardiologist earlier and earlier and putting patients minds at ease who are just suffering from something like acid reflux. Our technology has huge potential, if we can get it right.
What are the most important trends in healthcare that Ultromics contributes to?
Digital health and software-as-a-service via the cloud and of course artificial intelligence as a growing technology trend. Pre-covid a lot of hospitals were very averse to sharing data through the cloud and adamant that data must remain within the confines of the hospital. This was a regular battle that we came up against, despite the obvious benefits to both patient outcomes and hospital resources. We now live in a different, world where remote working and remote health monitoring has become more the norm and we rarely hear this objection anymore, furthermore hospitals are struggling to cope with backlogs from the pandemic and consequently it is speeding up the adoption of digital health and software-as-a-service based diagnostic products.
What are Ultromics’ key achievements and major milestones to date?
In the early days, our major successes were the massive trials that were conducted in the UK and US to validate our technology. Every trial completed provides stronger validation that our automated measurements trump the manual clinical workflows and result in better patient outcomes. Those trials seem insignificant now considering everything we have achieved since, but the data was key to securing our follow-on investments.
Our biggest achievements since have been receiving two FDA clearances in the US and two CE marks in Europe. We also have partnerships with leading research institutions, including the Mayo Clinic and the American Society of Echocardiograph, and were awarded a NHSX grant that provided funding to support rollout of our technology across the NHS.
What does the future hold for Ultromics, what’s next?
We have the platform, and we will continue to roll it out across hospitals in the UK, US and beyond. The next stage is to expand the features included within that platform and make it applicable to other uses of Echo in as many ways as possible. We want to be able to diagnose beyond just coronary artery disease and heart failure. My second goal is to expand the use of ultrasound within the community and support that rollout with access to our platform. We need Echo to be used earlier and more accurately in care pathways, so that we can identify and diagnose patients earlier and more accurately. I am dedicated to creating features in the software that will enable this.
What are you personally most excited for?
Everything that we do at Ultromics is designed to improve outcomes for patients, I am excited to see the impact that results from access to more accurate diagnostic tools. I began my PhD because I trained as sonographer, and I wanted to fix the problems I encountered. I am also excited about future advancements in our technology, we are already doing things that no one else can do but I am excited to see how far we can push it. There will be things that we create that we haven’t thought of yet.
I have also loved building a team, we started with just me and now Ultromics employs more than 100 people across the UK and US. It is getting to the point where I am struggling to remember all the names, but I love the team we have built, and I am excited to watch it continue to grow.