Anmol Satija – Hi everyone, welcome back to the unthinkable tech podcast, the go to source for the pulse on technology shaping our future. I’m your host Anmol Satija and today we are diving into a topic that touches every single one of us. I’m talking about cancer prevention. But today we are not just talking about medicine here. We are talking about how technology and data are helping us catch cancer earlier from starting from personalized risk assessment and maybe just maybe change the way we think about prevention altogether. 

So to help us unpack this, I am thrilled to be joined by Professor Gordon Wishart, a renowned cancer surgeon, researcher, and the founder of Check 4 Cancer, an award-winning early cancer detection company.

He’s also the mind behind the globally used Predict model for breast cancer outcomes. And he’s got some exciting insights to share around AI screening and what the future holds for proactive healthcare. So let’s get into it. Hi Gordon, welcome to the show.

Gordon Wishart – Thank you, Anmol Thank you for asking me to join your podcast.

Anmol Satija – It’s a pleasure to have you at the podcast, Gordon. So you you’ve had a remarkable journey. Like we just discussed a few back, like in few calls earlier from leading clinical teams in breast cancer surgery to founding Check 4 Cancer. And now driving some AI led innovations in early detection.

So what really stands out is how you have consistently been at the forefront of rethinking how we approach cancer and not just treating it, but in predicting and preventing it altogether.

Gordon Wishart on building a cancer prevention startup

So to kick off this conversation, I am curious and really want to know what inspired that shift for you, like from shifting from a traditional clinical setting to founding a tech company for cancer prevention.

Gordon Wishart – Yeah, well, lots of different questions there. think even during my surgical training in Scotland, we were really encouraged to understand research. We were all encouraged to do additional research degrees alongside our medical degrees. And I think that really sets you up for not only being able to develop your own interest in research and changing things for the better, but also it makes you much better understanding research that other people do. Is this a one-off thing? Could it be replicated in lots of different countries? 

So I spent most of my career not just doing breast cancer surgery but really doing clinical breast cancer research and what I mean by that is not really laboratory-based research. Clinical things that you can do. I’ve always been interested in why different countries have different cancer survival. What’s driving that? And so I think, you know, it’s always been a big part of my career and it’s sometimes difficult to balance clinical and research, but I’ve always really enjoyed it. 

And then you mentioned the PREDICT model. So when my colleague Paul Farrow and I launched that in 2010, what we wanted to do was to make a model freely available to patients and their oncologists to help choose the best treatment after their breast cancer surgery. I think this is where I really began to understand the quality of data and that high quality data is really important, whatever tool or model you’re building, because we actually built that model based on data from only 5,000 patients in the East of England.

We validated it in another 5,000 patients in a different part of England. And if you now think that’s now used all around the world by nearly a million patients and doctors a year now, ⁓ and that shows the importance of high quality data. 

So, know, that’s been a big part of my career and it still is a big part. We’ve just published a big study, US study showing that it’s really effective in US patients. And we’ve also just presented some data usin machine learning to try and improve the model so that we get better predictions in certain minority groups in the US like non-Hispanic Asian patients and non-Hispanic black patients and we’ve been able to improve the prediction. So you know that’s really the background and so I think in 2010 I took a sabbatical from the NHS. It’s always difficult to leave the NHS, it’s been your kind of anchor for a long career. Having taken a sabbatical, after six months I thought, maybe I could do something different. Maybe I could use all the things that I’ve learned both in clinical work and research and set up an early cancer detection company in the private sector. And that’s really what’s happened. So we’ve been going for 10 years now and it’s been an exciting journey.

Anmol Satija

Yeah. Sure about it. So Gordon, you have such an incredible journey and I love how you have taken years of your clinical insights and you know, the importance that you have shared for data and you have channeled it into something so forward thinking with your work that you are doing now. 

So it is really remarkable and it is interesting though because for so long, know, the spotlight in cancer care has been focused on treatment itself often after the symptoms appear. But what you are doing really flips that narrative. what do you believe?

Why early detection is more effective than treatment?

So my question to you is, why do you believe prevention is the next big frontier in cancer care? And in your view, how ready is the wider healthcare industry to actually make that sort of shift?

Gordon Wishart – Yes, I think you’re right in saying that for the last 10 or 20 years, the real focus has been on cancer treatment and the introduction of really expensive drugs that make really quite small differences to how long a patient is going to live after their cancer treatment. And actually, if you can pick up the cancer at an early stage, then you can make much more of a difference for that patient. You can ensure that they have less treatment, you know that they’re going to live longer. So you get much more impact from focusing on that early start of the cancer pathway. And you know, there is a lot of publication now that show that almost of cancers are preventable. And that’s really shocking. mean, if we just look at the UK, there are a thousand cancers diagnosed every day. 

So 40% mean we could stop 400 of them and that puts it into perspective. And really a lot of the increase in incidence in recent years has been driven by lifestyle factors. So smoking, alcohol, lack of exercise, poor diet. So those are things that we can do something about or we can try to help people do something about. we’ll come back to that maybe because that is quite a big challenge.

And so alongside that, trying to reduce your lifestyle risk factors, the other thing that we can do is screening. Now, a lot of people just think that screening picks up cancers at an earlier stage, which it usually does. But what they don’t understand is that screening also picks up a lot of high risk lesions, which because they’re removed, stops them developing into cancer. So it’s kind of got a twofold way of working, picking up cancers early and stopping some cancers actually developing. 

How companies are adopting cancer screening for employees?

And that’s really important. Now, the corporate sector is really stepping in and starting to look after their employees in lots of different ways. Health, wellness is a big part now of kind of support packages that people get employee benefits. And the reason that the corporates like cancer screening is that actually there’s a big return on investment for them as a company. If people are going to get diagnosed with cancer, they get diagnosed at an early stage. They’ve got less time of work. There’s a whole lot of less things. So they get it and they’re willing to invest because they know that return on investment will come over maybe start about five years. 

The problem is health insurers work on an annual basis and want to balance their books on an annual basis. So they are struggling as health insurers to invest in prevention now because they might not even have that member. They might have gone to a different insurer in five years time. So they are less guaranteed to get that return. So there’s a real juxtaposition at the moment. Corporates get it, it’s in their interest. The insurers would like to do it. 

And we are working with one insurer in the UK, Vitality Health, who are now focused on cancer prevention. But as an insurer they’re very unique because they’ve got lots of data that shows that for their insured members the amount of exercise they is directly related to their life expectancy. They’ve got fantastic data. But I think it’s going to take a lot of effort to get the other insurers to catch up.

Anmol Satija – Yeah, definitely. So that’s a great initiative that you guys are doing. And I think that the way that you mentioned that the corporate sector is already embracing screening is promising. Of course, they are looking at the bigger picture of what eventually they’ll benefit from the ROI that they’ll get. But for wider adoption, it sounds like we still have some hurdles to clear, right?

So what do you think are the biggest blockers, whether systematic, cultural, or even financial, that are slowing down the shift to proactive care? How can tech maybe help us close this gap faster?

The role of behavior change in cancer prevention

Gordon Wishart – Yeah, well, I hope tech will help us do this. I think one of the biggest hurdles is we can help people identify what their own risk factors might be, but changing someone’s behaviour is incredibly difficult. It needs strong messaging, needs consistent messaging, it needs some kind of ongoing support. And certainly in the UK, we used to have a big public health infrastructure throughout the UK that would get all these public health messages out, but that’s largely been dismantled. We have got an NHS Cancer app now, and I was part of a group of people that recently wrote to the UK government saying, could we use the NHS app to get messaging out about helping people have healthier lifestyles? But I think that’s the biggest challenge because almost 70 % of people in the UK are now overweight or obese.

So, and they all know what’s driving it, but actually, you know, it’s a challenge to get people, it’s like stopping smoking, you know, or even taking more exercise. People just seem reluctant to get off the underground, a stop earlier and walk or use the stairs, you know, things that you can actually just build into your everyday life. You know, not everybody has to go running for an hour every day. 

There are loads of things you can do, but I think changing behaviors probably the biggest hurdle that we have. The good thing is that health trackers are now becoming quite popular. So a lot of apps help you track your steps and even we just partnered with a company called Alula, which has a facial scan and which can actually show up lots of your cardiovascular risk and it also takes information from wearables. 

So that’s incredibly powerful and we’ve teamed up with them because when you look at cancer and cardiovascular risk, they are two of the biggest risks for people and also for insurers. And if you can start to detect those risk factors early, you can do something about them. So I think from a tech point of view, I health trackers are going to be really important.

Anmol Satija – Yeah, I would agree to that and that’s really exciting and the idea of layering cancer and cardiovascular risk using something as accessible as a facial scan. I think it shows how much potential there is when tech and healthcare truly work together. Now, you know, with platforms like My Cancer Risk that you offer, offering personalized insights. 

How do you strike that delicate balance giving people the information they need without really triggering over diagnosis or unnecessary anxiety? You know, it’s a very big disease that we are talking about and people often get really scared while dealing with it or even with, you know, when they’re thinking about it. So how do you balance that out?

How personalized screening reduces overdiagnosis?

Gordon Wishart –  Yes, so if I explain that my cancer risk is a platform that is partly a cancer risk questionnaire for six common cancers, which identifies people at higher risk and in the top 20 % of risk. But it also has an education hub which tells you all about risk factors, how to reduce your risk and the signs and symptoms of common cancer. it’s a kind of we developed as a whole of workforce solution. The idea being that the questionnaire can be offered to everybody.

So everybody benefits from the education part of it and the company pays for those at highest risk to have screening. And that works really well. there has been a, there’s always a concern about screening that it leads to over diagnosis and over treatment and unnecessary anxiety. But actually, if you do this risk stratified screening, which is what we call it, actually reduces the amount of unnecessary diagnosis and the amount of unnecessary anxiety because in that higher risk group there are a greater proportion of cancers than in the average risk group. So there’s a good example of this. 

There’s been a lot of publicity in the last two or three weeks about saliva DNA test for prostate cancer and it looks at tiny little errors in your DNA and it can identify the top 10 % of risk. And in the study that they did, 25 % of people in that top 10 % of risk were diagnosed with prostate cancer. Most of them significant prostate cancers, harmful ones. So that’s a one in four risk. The average risk for a man in the street of getting prostate cancer in his lifetime is one in eight. So that’s double the risk, double the number being picked up. So that’s a really good example. 

I think we’re going to see a lot more of this, especially in big companies, where it would just cost too much money for them to offer screening to older employees. This is a really good way of offering something to all the employees, but targeting their focus and their spend on those at highest risk. And I think that’s why it’s become really successful in the corporate community in the UK and we’re now getting increasing interest from Europe, from Australia and from the United States about this approach.

Anmol Satija – Okay, right. I think the way you’re combining clinical insights with data driven screening is really impressive and will make a lot of difference, especially when it comes to prioritizing care for those who need it the most. So, one thing that I want to talk about is the importance of genetic data here. So how much do.

I mean, how do you see genetic data providing an upliftment in terms of accuracy while building such solutions? From a tech standpoint, also, how do you ensure data privacy stays airtight?

Gordon Wishart – Yeah, I mean, that’s a really good question, I think. And I think it’s one of the challenges for a lot of tech companies is that information security side of things. When you come from a clinical background, data protection is huge. It’s kind of part of your inner core. But I think it is really important. But we’re lucky in the UK because there is a longstanding agreement between the UK government and health insurers about genetic data to protect it. 

The impact of genetic data on cancer risk prediction

So it’s never made available to insurers. And we’ve used genetic data for a long time to identify people at high risk of breast cancer, the highest risk group, because they’ll benefit from extra screening. A lot of them have risk-reducing surgery, like bilateral mastectomy. And a lot of people in that higher risk group go on drugs to try and reduce their risk.

We’ve used it for a long time, so you can use it in a way that gets the best outcome for the patient choosing the right drug, choosing the right treatment, but protecting that data from being misused by insurers. So I think that is the way forward. I think data security is becoming a big issue and patient safety and risk reduction is a big part of any medical device certification now. So I think there is an infrastructure in place to help protect patients, but at the same time, help them benefit from this real kind of tsunami of genetic data that’s now coming out.

Anmol Satija – Yeah, definitely. that’s really encouraging, especially knowing that we can harness genetic insights without compromising the privacy part and even use them to protect entire families through early intervention. So while we are talking about data on a related note, as you know, AI, let’s talk about AI now, as AI starts playing a bigger role in diagnostics. There’s always that debate that’s going on. Should it serve purely as a decision aid for clinicians or eventually take an autonomous role? So where do you personally think we should draw that line?

Using AI to improve cancer diagnosis

Gordon Wishart – Yeah, well, I’ve got personal experience, you know, in this area now because we’ve run a very successful tele dermatology pathway for patients with suspicious skin lesions for years now. And we’ve used all of those data, the images and all the clinical information to build an AI model. We’ve published quite a lot of the results to date in good journals, but we’ve just done a study. 

We’ve looked at two years of data and we’ve shown that actually our AI model can be used either as a clinical decision aid or for autonomous reporting. The big challenge with autonomous reporting is are you going to miss any cancers if there’s not a doctor involved as part of that process? But in fact, there are precedents now for these types of models where they can be used autonomously. 

And we’re really lucky because most AI models for skin cancer are just built with images and images are great at detecting the cancers. But what you want to do is you don’t want to have to bring lots of people back to the clinic just to diagnose those cancers. And that’s where the clinical information comes in. It helps us be much more sure about the model saying that something is not suspicious and can be and that patient can be reassured. 

So we found a way with our model to say we can actually we can discharge about 50 % of the patients on the basis of the AI model alone without a human report. And that means we can speed up skin cancer diagnosis, diagnose it earlier because there’s a big global problem with skin cancer. The incidence is rising and we’ve got international shortages of dermatologists. So we need ways to streamline skin cancer diagnosis. Now, teledermatology does that.

But there’s also a shortage of really good teledermatology reporters. Someone who’s willing to sit at their desk and look at the images and the clinical information and say discharge or recall. And so that’s where the AI model will come in. It can help more inexperienced, less experienced reporters improve their reporting as a decision aid, but then as a whole group, we can streamline the diagnostic process.

That’s really exciting. so I think the biggest challenge is really to get the medical device regulators to catch up with this. We’re really progressing at quite a pace. And it’s very frustrating when we’ve got different medical device regulators all over the world. But actually, we’re now seeing real breakthroughs where, and I think the other challenge is everybody thinks that if AI doesn’t pick up 100 % of something, it’s bad. Doctors don’t pick up 100%. If you look at the data, the average percentage of skin cancers picked up by experienced dermatologists is about 90%. Our AI model is up to 99%.

So we can’t ignore that, you know, this is a, but we have to do it in a way that’s safe and protects patients. And I think you can do that in a pathway where doctors are involved and have oversight and put a check on it. ⁓ I think you can do it.

Anmol Satija – Of course, of course. And that’s really exciting. And I love that approach where, you know, we give clinicians both confidence and the bandwidth that’s the sweet spot tech should aim for. And how is it, how it’s exactly building up the efficiency and detecting more and more patients.

The use case that you shared of using imagery for skin cancer detection. So Gordon, now moving, like stepping aside from all of this discussion. So now you have worn so many hats across clinical and tech settings. So I am curious to know one thing, like.

What’s one common misconception you see business leaders making when it comes to innovating in healthcare space?

Why clinical governance matters in health tech?

Gordon Wishart – Yeah, think when you’re looking at lot of innovators, especially technological innovators, you know, trying to solve health issues, I think I don’t think they understand the importance of clinical governance. You know, it’s it’s just as important as information governance, because as doctors, we have, you know, a responsibility to look after the health of the patient and to reduce any unnecessary risk. 

So I think if you come at this just from a tech background, then you will fall foul of that. And I think that’s why it’s really important that clinicians are involved with our tech partners, both to determine what’s the clinical question we’re trying to solve here? What are we actually trying to do? And then you involve those doctors with their tech counterparts to try and develop something that will solve that problem. So think having clinicians involved and understanding the clinical risks and clinical governance is really important.

Anmol Satija – Right, right. Definitely it is of utmost priority. And today, you know, bringing a product to market in healthcare is a whole different ballgame when we compare it to most of the other industries. And I recently did an entire episode discussing this issue wherein, you know, we discussed certain challenges that specifically comes in picture when we are innovating or building a product in healthcare space. 

So, right, so this was a great conversation that I’ve had with you, Gordon, and I loved your insights. Now, on a concluding note, I would say that looking ahead, what’s your moonshot prediction for cancer tech, something that’s not quite mainstream yet but you believe could gain serious traction by maybe 2026.

Moonshot prediction for cancer tech

Gordon Wishart – Yeah, well the thing I’ve chosen for this and there are lots of exciting things out there but I’ve been working with a company who have managed to train rats to be able to smell a signature from a particular type of cancer in the urine. So what happens is that cancer cells produce a lot of what we call volatile organic compounds of EOC. You can detect them in the breath.

But you can detect them in the urine and they have got really exciting data that shows that they can detect over 90% of lung cancers doing this. that technology, they don’t actually know what all the compounds are yet. Once they are able to do that, then they will be able to stop using those animals and use the animals to train for a different cancer. 

So that’s super exciting because a urine test could be at home, it could be multi-cancer, it could really revolutionise screening. And so I think we’re staying very close to them because I’m super excited about them. And I think in the next three four years that could really accelerate and be available for several cancers.

Anmol Satija – really rats detecting cancer. Now that’s a moonshot that captures the actual spirit of innovation perfectly. So thank you Gordon for sharing all of these exciting insights and informational too. Thank you for your time and for leading the charge in cancer prevention through different innovations and initiatives that you have been taking up. Thank you so much.

Gordon Wishart – Thank you very much.

Anmol Satija – And to our listeners, if this episode got you thinking differently about prevention, risk, or the role of AI in healthcare, don’t forget to follow the Unthinkable Tech podcast. We’ll be back soon with more exciting conversation from the front lines of innovation. Until next time, keep imagining what’s next.