Cardiovascular disease is the world’s biggest killer, accounting for more than 30% of all deaths globally and incurring an economic burden that is expected to exceed $1tn by 2030. In the UK alone around 425 people die every day, that is one person every three minutes.
Using artificial intelligence and deep learning image processing, Caristo Diagnostics can identify “ticking-time-bomb” arteries, predicting a potentially fatal heart attack years in advance.
Based on Oxford scientific research, funded by the British Heart Foundation, and a ground-breaking discovery that rewrote the textbooks and transformed understanding of the biological mechanisms behind a heart attack, the technology is beautifully simple – only requiring information that has always been present in routine CT scans.
The game-changing technology is now CE marked and available for use across Europe and can be rolled out across the NHS.
We spoke to founders Dr. Cheerag Shirodaria, CEO, and Prof. Charalambos Antoniades, CSO, Caristo Diagnostics to learn more.
Why and when was Caristo Diagnostics founded?
Dr. Cheerag Shirodaria: Caristo was founded in 2018 by four Oxford cardiologists, including myself and Charis. We have all worked together for many years and our research has been dedicated to investigating the causes & consequences of coronary artery disease. But the primary driver for Caristo came from the research and scientific discoveries led by Charis.
Prof. Charalambos Antoniades: For many years people believed that changes to the fat surrounding arteries was responsible for driving changes inside that lead to a heart attack. But a few years ago, we discovered that this is wrong, and it is actually the reverse; the artery is responsible for driving changes in the fat. This was the breakthrough discovery that led to the development of technology that led to the creation of Caristo. The IP was filed in 2014 and the company followed a few years later in 2018.
Can you tell us about yourselves and the Oxford scientists behind Caristo?
Dr. Cheerag Shirodaria: Charis and I actually began working together 17 years ago as research fellows and we have remained very close ever since. Our other two founders, Prof. Stefan Neubauer and Prof. Keith Channon, were my research supervisors going back many years ago and we have all been consultant colleagues for the last 11 years.
I actually deviated from the conventional medical career; I began as a cardiology registrar, completed my doctoral thesis in Oxford but then received an MBA from the London Business School. I came back to Oxford as a consultant cardiologist and helped to set up the CT heart scanning service before moving into industry, running the medical devices, diagnostic and cardiovascular business for Covance, a large global CRO.
It has been fantastic to found a company with people you know so well, we are completely aligned on the science, the journey and our vision. Our skills are complementary, with a strong focus on academic and Industry experience.
Prof. Charalambos Antoniades: I am a Professor of Cardiovascular Medicine at the University of Oxford and Director of the Acute Vascular Imaging Centre at the John Radcliffe Hospital. I ran a program that collects biopsies from patients who have had cardiac surgery, this includes biopsies of arteries and of the surrounding fat. My aim was to understand how the two communicate and to develop better tools and imaging applications that can detect an inflamed coronary artery. We have always known that inflammation in the arteries drives cardiovascular disease, heart attacks and cardiac deaths and, until recently, cardiology textbooks deemed the detection of inflammation as the holy grail for cardiovascular diagnosis. We believed that if we could detect inflammation, we could predict a future heart attack.
That day has now come, Caristo’s technology is able to non-invasively detect inflamed arteries and predict a patient’s risk of a heart attack.
What real-world problems or unmet needs are you aiming to address?
Dr. Cheerag Shirodaria: Despite the huge amount of time and money we spend trying to predict and treat heart disease, heart attacks and heart disease are still the world’s biggest killer. We have known for over 50 years that a process called atherosclerosis leads to heart attacks. Atherosclerosis is the narrowing of arteries, caused by the build-up of fats and cholesterol (called plaques) within an artery. We have also known for many years that atherosclerosis is an inflammatory process and that inflammation plays a key role in plaque rupture, which causes heart attacks.
Unfortunately, we have never had a straightforward method to detect inflammation within an artery. Blood tests are not sensitive enough; they can tell you about inflammation in the body but not specifically in the heart, and imaging tests can only detect the degree of narrowing and the associated consequences, such as restricted blood flow.
But approximately 50% of heart attacks occur in people who do not have narrowing’s. Instead, they have minor narrowing’s that are highly inflamed. that rupture. These are people at risk who are missed by current diagnostic tools because we couldn’t detect inflammation. We have always known this, but we have lacked the technology to identify and do something that could save them. Caristo’s technology can detect inflammation and it can do so from a routine CT heart scan. This is a first line investigation for someone with chest pain, for example when exercising or walking.
Prof. Charalambos Antoniades: To echo Cheerag, half of all heart attacks happen to people whose CT scans appear normal by conventional analysis and show no evidence of disease. This is the big unmet need that Caristo solves, we can detect patients that are going to have a heart attack when there is no other way of knowing.
Can you tell us about the scientific research and why it was ground-breaking?
Prof. Charalambos Antoniades: We found that when an artery is inflamed it secretes inflammatory molecules. The artery is surrounded by fatty tissue (perivascular fat), and this fat tissue changes its composition and structure in response to the inflammation in a way that is not visible to the naked eye.
Perivascular fat does not affect arterial inflammation, inflammation within the artery affects the fat. The exact opposite of what people used to believe. It actually took us many years to believe it ourselves because it was rewriting textbooks, changing the established knowledge that we believed to be true. We had to be very comfortable with our data before showing the world that its thinking was upside down.
In fact, when we started collecting fat biopsies from patients who had had cardiac surgery, I had a DPhil student working with me. He was trying to understand whether protective molecules that are generated in perivascular fat are correlated with a better inflammatory artery profile. He came to me with the first 100 patients and the results were the opposite. The more “protective” molecules there were present in the fat the more damage there was in the artery. It didn’t make sense, so I dismissed the results, sent him back and said this is nonsense. I told him to increase the numbers in the research group, so he came back with 200 patients and the results were stronger. I still dismissed the results and refused to believe it. He then came to me with 640 patients and the result was hugely significant and striking. We then built a whole programme to explore this new hypothesis and confirmed our discovery.
Can you tell us how you got from scientific discovery to usable technology?
Prof. Charalambos Antoniades: Our discoveries were made by analysing the biology of patient biopsies, so the next step was to analyse images of arteries and perivascular fat on CT scans and as a result we developed a method called Fat Attenuation Indexing (FAITM). This new biomarker measurement analyses changes in the perivascular fat around the arteries, detects inflamed arteries and is able to accurately quantify the extent of inflammation.
The challenge is to visualise these changes. If you look at any kind of CT heart scan, the arteries appear white and the fat surrounding black. It is black because it has been removed. All algorithms developed by big software focus on presenting nice, clear pictures of arteries and until now fat has been considered an artefact, so it is suppressed and removed during reconstruction of the image. However, the digital information remains available, hidden in the background.
Our technology uses AI and deep learning imaging processing to delve into this digital background and bring the hidden information to the surface, reconstructing the image and then quantifying the changes that correspond to the biological changes we see in the tissue when we take biopsies.
Our technology was validated in a landmark study involving around 4,000 patients, followed up to nine years after examination. Published in the Lancet in 2018, we found that when you quantify the degree of inflammation in the arteries you have a striking ability to predict who is going to have a heart attack, many years before that heart attack happens and before plaque has even begun to develop.
How does the cardiovascular risk prediction technology work in practice?
Dr. Cheerag Shirodaria: In practice our technology fits perfectly with a cardiologist’s workflow. Clinicians can send us their patient’s routine CT scans; we can run the analysis and return the report directly to them. Hospitals don’t need to change equipment and patients don’t need another test. We are hoping to introduce it into the NHS as soon as possible and have already held positive discussions with senior figures who are excited by the potential impact of the technology.
Prof. Charalambos Antoniades: You can identify the arteries, that appear entirely normal right now, that will develop plaque as a result of inflammation and go on to rupture and cause a fatal heart attack. So very early on in the disease process we will be able to intervene and change things, because if you wait for the plaque to develop in the arteries then it is impossible to go back to normal. But you can treat inflammation and prevent plaque from developing and that is the big strength of the technology.
What will be the ultimate impact of this technology?
Dr. Cheerag Shirodaria: The important thing to realise here is that Caristo’s four founders are all cardiologists. We see patient’s day-in and day-out that slip through the cracks, whose CT scans hide vital information. Our ultimate aim is to give our fellow clinicians the best information to help them make the right decisions for their patients.
As a first step, Caristo will impact those patients who are already having CT scans, these are already taking place, so we are not asking patients to undergo any extra diagnostic tests, we are just trying to extract the maximum information available.
However, Caristo’s technology provides a window into heart health which raises some very interesting questions in terms of screening. Should we be looking to scan patients when they are younger, before they have developed symptoms to quantify risk? We believe so, this would be a whole new level of real-world impact and opens up a totally different approach to prevention and meaningful intervention.
If we can identify patients who are at risk earlier then we can intervene in a targeted and personalised manner to modify and reduce that individual’s lifetime risk of having a heart attack.
There are ongoing studies looking into CT heart scanning as a screening tool in a number of countries, Oxford and Caristo are involved in a number of these and we hope to have answers in the next few years.
What are the most important trends in healthcare that Caristo contributes to?
Dr. Cheerag Shirodaria: Prevention is better than a cure and COVID has accelerated the focus here. We need to intervene earlier to keep people out of hospital and in order to do that we need intelligent tools to identify the individuals most at risk. We can’t treat populations, we can’t just give statins or aspirin to everyone, it just doesn’t work. Targeted tools allow us to identify the at-risk patients and to focus our time, effort and resources on modifying their risks. Caristo is all about prevention and identifying high-risk patients. Using data that is already there, we can support physicians to make better decisions for their patients.
What are Caristo’s key achievements and major milestones to date?
Dr. Cheerag Shirodaria: We went from founding a company to FDA clearance in 18 months, this is almost unheard of, and we have now gone from founding a company to CE mark within two and a half years. Nature voted us “one of their top four spinouts” across all science sectors, OBN named us “best med tech startup” in 2019 and both the British Heart Foundation and the European Society of Cardiology have hailed our scientific breakthroughs as “game-changing”.
From four founders we have built a great team of software developers, image analysts, operations and regulatory people growing now to a company of 23 people.
Our first product CaRi-Heart™ was accelerated by an Innovate UK’ Biomedical Catalyst Award. Together with the University of Oxford we have been awarded a British Heart Foundation Translation Award and a Horizon 2020 Award to help develop our second product, which focuses on stroke prediction. We recently received an NHS AI Award to develop our third product which can predict a patient’s risk of diabetes by looking at fat health.
There has also been a lot of interest in what we are doing from the biopharmaceutical industry and we are very lucky to be working with some of these companies who are using our technology as biomarkers in their clinical development programs as they focus on developing new drugs to target heart disease.
What does the future hold for Caristo, what’s next?
Prof. Charalambos Antoniades: At a very high level you see the same biological changes in other diseases, inflammation drives disease and changes can be seen in the fat tissue that surrounds blood vessels. As Cheerag noted, our first product predicts heart attacks, but from the same scan we can predict who is going to develop a stroke and who is going to develop diabetes. We have also developed a new technology which tells you during an acute infection from COVID-19 which arteries are affected and if a patient is likely to die as a result of the vascular complications. Information again that is present in routine scans taking place when COVID-19 patients are admitted to hospital.
Our flagship programme with the NIHR and the British Heart Foundation aims to analyse tens of thousands of scans from NHS patients admitted with COVID-19 to see whether we have the ability though this technology to predict cardiac events, heart attacks and strokes years after an acute COVID-19 infection.
What are you personally most excited for?
Dr. Cheerag Shirodaria: From my perspective, it is all about bringing our technology to patients. We have translated a tremendous scientific breakthrough into a meaningful tool that will support our physician colleagues to make personalised treatment decisions and improve outcomes for patients. This is all about patients.
Prof. Charalambos Antoniades: Same for me, all the research and work we do is about saving lives. Seeing our research translated into something that is widely used in the clinic to save lives and improve the quality of life of our patients. That is the vision.