Bowel Research UK’s Patient and Public Involvement Manager, Sam Alexandra Rose, is joined by Alicia Toon, Marketing and Communications Executive at Odin Vision. Together, they provide a simple explanation of what AI (Artificial Intelligence) is and what colonoscopies involve. They emphasize the benefits of combining AI with colonoscopies for both doctors and patients. Alicia discusses the research conducted by Odin Vision and highlights that the AI technology they are developing will be used as a tool to assist doctors rather than completely replace them, which has shown to be an important point for patients.
Transcript
Sam
Welcome to Can I Butt In, the Bowel Research UK podcast where we welcome bowel cancer and bowel disease, patients, researchers, healthcare professionals and carers to butt in and share their experiences. We’re picking a topic every episode and getting to the bottom of it.
I’m your host, Sam Alexandra Rose. I’m the patient and public involvement manager at Bowel Research UK. And as a patient myself, I’m excited to bring more patient and researcher voices into the spotlight.
Hi everybody. Today I’m sitting down with Alicia Toon from Odin vision to talk about the use of AI in colonoscopy procedures. Odin Vision is a cloud AI endoscopy company now owned by Olympus. Alicia is the marketing and communications executive at Odin and like me, she’s been sitting in on focus groups around this research. So, we’re excited to explain the research and dig into how it’s all going. Alicia, welcome to the podcast.
Alicia
Thank you. Lovely to be here.
Sam
So, I’m just very quickly going to explain what a colonoscopy is, because that’s the main thing that we’re gonna be talking about today, for anybody who is lucky enough to, to not know what a colonoscopy is, and to have never had this procedure. It is the, basically, it’s the, the camera up, up the bum, I’m doing this off the cuff. I, I have not planned how I would describe what a colonoscopy is. But yes, it’s an investigation to diagnose things like cancer or IBD, diverticulitis, all of that sort of stuff. And yes, it’s, it’s the old camera up at the bottom and I’m sure that most people are aware of it, and many of us are, unfortunate enough to have experienced it. Not to say it’s that bad of a procedure because I don’t want to put anybody off. It is a lifesaving procedure, very important and that’s why the work that Odin vision is doing to improve this procedure is so important. So, let’s jump right into it and tell me about Odin visions research. So how are you using AI to try to improve colonoscopy procedures?
Alicia
Here at Odin vision, we are using cloud and AI to support doctors in the detection and diagnosis of disease during endoscopy procedures. So, from our research, we know that colonoscopies can be uncomfortable for patients, and with our AI, we are looking to have reduced the chances that colonoscopies will have to be repeated multiple times and our aim is to get the doctors to detect any polyps that could be missed. Therefore, cutting down on the chances of these progressing into something more serious, that could affect the standard of living for patients.
Sam
Great thank you. So, yeah it is, an uncomfortable procedure and yeah, difficult when you have to repeat it several times as, as well, I mean some people, you know, have to go for regular as part of perhaps a screening programme. So yeah, anything that that we can do to ensure that anything is, is caught during that time is really important. And I should actually say at this point, I’ve explained what a colonoscopy is, but I haven’t explained what AI is, and I know that’s been in the news quite a bit recently. It’s, it’s really sort of ramped up the interest around that. So AI, again for anybody who doesn’t know, stands for artificial intelligence, and there’s also sort of part of it called machine learning, which is it’s basically your you’re teaching computers or technology to, to do something right? So, it’s really only as good as the data that you’re putting into it. So, you’re giving something lots and lots of data, so that it then learns patterns and then it can tell you about things that you, you then give it later on is. Is that a, a fair kind of way of explaining it?
Alicia
Yeah, definitely. And then, you know, the more we’re putting into it, we train these models to recognise… I touch upon it later. It’s, it’s much like the facial recognition on your phone, you know, it’s been given multiple images and different faces to ensure that it can pick this up and then that of course, leads us to be able to detect the polyps. And that’s what we change our AI on different polyp sizes. To ensure we can detect these with our AI.
Sam
What is the difference then between a kind of regular colonoscopy, if you like, and the one that Odin vision is working on with AI?
Alicia
Yeah, so with the regular colonoscopy, the doctor will manually navigate the scope through the colon. And they visually inspect the colon walls for abnormal… abnormalities such as polyps, you know, inflammations or any signs of diseases, like colorectal cancer in this case. And then if any abnormalities are found, the doctor May perform biopsies or remove polyps during the procedure, which they send off for inspection, and with our AI… so this incorporates the, the AI algorithms and computer vision technologies to help the doctor during the procedure, and with the AI system it can analyse real time images and videos from the colonoscopy and provide immediate feedback to the doctor. So in in real time, it may help identify these abnormalities that could be missed by the human eye, or provide additional information to aid and decision making during the procedure.
Sam
And that sounds, again really, really important. It’s like having kind of a robotic second pair of eyes if you like, isn’t it?
Alicia
Yeah, definitely.
Sam
And so, I’ve been sort of sitting in these focus groups. The thing that excites me the most about it is, yes, it’s excellent that this AI can help the doctors to see that things that they might not necessarily have been able to spot before, but also this part of being able to help with diagnostics, as well, as I understand it from focus groups, it helps doctors to give patients answers there and then, a little bit more, more easily rather than having to wait for, say, a biopsy to go off to, to Histology, AI will be able to help the doctor to, to say more definitively this is likely to be cancerous or non-cancerous and that’s, as a patient, that’s really exciting for me because I can see the, the, the real life benefit of this research, because sometimes when you’re going for procedures and you have, what we kind of in the cancer community call, ‘scanxiety’ which, yeah, maybe other people use as well. But it’s, it’s that worry leading up to the procedure, not just of like what the results will be, but actually having the procedure as well and part of that as well is then, well waiting for the results afterwards. And if you can cut down that time then that’s, that’s cutting down that anxiety as well. That’s really exciting for me to think that. You know, maybe we won’t have as much of a weight to find out if something is good or bad. That’s, that’s excellent.
Alicia
Exactly, it’s true. I think definitely with our, our PPI sessions, this has really come through and you know everyone does speak about it is, you know, the waiting. That and that anxiety that patients get from that. That is quite a big factor that plays into, you know, how this sort of AI can really. You know, help patients with that, of course at different other aspects, but that that one, I think everyone shared was quite a big thing for them.
Sam
Absolutely. And I should explain as well that PPI stands for patient and public involvement and that’s my job as patient and public involvement manager at Bowel Research UK, is to help researchers and companies like Odin Vision to get patients involved in, in research and have focus groups and things like that. And I know that we’ll speak a little bit later about what the patient involvement has been on this research specifically. But going back to the AI technology then, so how does the AI work?
Alicia
So, as we touched upon earlier, so it works by inputting a large amount of data intended to train an algorithm to recognise certain features of, let’s say the abnormal tissue. And this like I said is similar to face recognition or social media filter, how it recognises you. And it has been trained on many, many pictures or faces. And facial features in order to actually place a box or filter. So, our technology has been trained and tested on thousands and thousands of images and videos in order to recognise polyps, live during a colonoscopy.
Sam
Right. Yeah. And it just makes me think that AI is kind of like a little robot that goes to school, when you feed it all of this information and, and then yeah, it goes out into the world and it’s kind of given… it leaves school and it’s given a, a job. Yeah, that’s metaphor I’ve just made-up for myself.
Alicia
The more you, yeah, the more you train on it, the more it’s going to learn and, pick up things. I mean, it is quite fascinating, really what these models, AI can do when you train it and when you’re using it in something like healthcare. I think the important thing we always try to say is AI isn’t there to replace doctors, it’s there to support them in their decision making, and that’s something that patients as well, you know when you say AI would just make the decisions, they… they’re not so keen in a way, but we, what we tried to say is; that isn’t our intention at all. We want patients, you know, to use this AI with doctors. It’s, you know, together to try to help combat this disease.
Sam
And there’s that thing as well, isn’t there? Referred to as human in the loop, where there’s always, there’s always going to be like a human in there somewhere. And it again it’s, it’s. AI is, is trained by humans. You know it’s, it’s not like the robots are just like figuring everything out for themselves. It’s the humans that are that are creating it, feeding the information, and sounds like it’s a very robust and thorough process that have to go through to, to feed thousands of, of the data to it, it must be I would imagine it’s quite a regulated thing.
Alicia
Yeah. And of course, our team, you know, they have, we have the data annotators and a very strong team that, you know, they go through multiple things a day and, you know, without… without our team behind it all this AI would not be possible, which is another amazing thing, it’s the people that come together to actually give the AI the information it needs as well, so everyone works extremely hard to ensure we do that so we can push forward with this AI and make it as robust as possible.
Sam
That is a clever bunch of people for sure.
Alicia
Yeah, I know! I just have the pleasure of being able to speak about, but definitely the credit goes to the team behind it all 100%.
Sam
Absolutely. So, I mean, I touched on the kind of benefits for, for patients earlier and kind of what, what my impression is of that. But what are the benefits that you see for not just patients but also doctors in the NHS as well?
Alicia
Yeah. So right now, of course doctors, and they are manually entering… So, they have multiple procedures a day and they have to manually, you know, that’s a lot of procedures to remember exactly, you know, what they saw, and they have to review these procedures once they’re finished. And this can, you know, takes a lot of time. They may even maybe forget some things. But with this sort of AI, it can help with the automation of tasks to ensure doctors can, you know, spend more time with patients so, you know, it’s not. It’s something we’re working on in the background, but you know, with the idea of reporting, so reporting tools that can help take information and automatically put this in, you know, a form for a doctor. This is something that we, we definitely believe will help make the whole procedure a lot quicker, but also ensure that that time with patients isn’t being taken up by the doctor having to go to their computer, filling in whilst they’re going, but also saving, saving the NHS a lot of time because, if we can narrow down the amount of, you know, if someone goes for a colonoscopy and let’s say they didn’t find something, it was missed or overlooked, then they’re going to have one in another few years and then if that something’s found then, then that time, that was wasted not only for the patient that maybe they could have possibly found something earlier, but also for the NHS. I’m not sure how much, but I think you know for these procedures. I’m sure there’s a cost. I think there’s quite a big cost on the NHS for this whole process. So, we believe that with this it will save a lot of time and money spent as well for the NHS.
Sam
Yeah. And that benefits, yeah, benefits the patients as well and, maybe we’ll even get to a point where it saves so much time that doctors, we will be able to see an extra patient in their clinic. Yeah, that would be great.
Alicia
And yes, exactly.
Sam
And it really feels to me, because I work with researchers, obviously a lot and I’m seeing more and more of studies around AI, and AI being researched, for all sorts of different things. And yeah, automating different processes and, and using things like ChatGPT type things as, as well in other studies to automate these more administrative processes that take up a lot of doctor’s time because that’s something that that you can automate and then maybe the doctor can just have like a quick look over it. And then that gives the doctor more time to spend face to face with the patients, talking to them, reassuring them. And then, yeah, doing the clinical work that is actually their, their expertise. So yeah, I can definitely see like the benefit of AI in the NHS, and I think it’s just gonna grow and grow. So how far along are you with this specific research and what’s what’s next?
Alicia
So, our, our medical devices live and available to use and it has been rigorously tested in UK and European hospitals currently as well. We’ve completed a large clinical trial for AI that helps doctors detect adenomas, which are the more the cancerous sort of polyps, which supports the use of our technology. And within this we found that, that adenoma detection rate was significantly higher with the use of AI as opposed to without, so it was again emphasising and showing our AI was there as a second pair of eyes and live during this procedure, and we are also planning further exciting collaborations with UK hospitals to support the use of our medical devices and prove the benefit to patients.
Sam
Yeah, I think if I was going for a colonoscopy, I would and I had the option for having one with AI, I would. I would be really interested in in this. Yeah. I would want to see it. And yeah, it sounds like it would, it would definitely be beneficial, yeah. And it’s again, it’s a robust process, isn’t it? Because you have these clinical trials to, to test everything. So yeah, really rigorous, checking everything all the way through.
Alicia
Yeah. And of course, then from that to, to sort of showing people. Because it’s one thing us telling everyone, you know, what the potential AI has, but when you have the figures there, because we have multiple studies and reports that that are out there even now. I think having the data and the evidence behind it, as well is definitely surely reassuring to patients to show that our AI is doing what we’re saying, and what we’ve set it out to do.
Sam
And what about the clinicians then? What do they think about it?
Alicia
So, you can imagine they have been very positive about it, and our medical devices are intuitive to use and to easily integrated into current clinical workflows. With little change to how things are normally done, so it doesn’t disrupt doctor’s current workflows, which is great, and they have been very positive about the use of AI and how it aids real time decision making. And we have our AI and it has been adopted by world leading cancer hospitals such as; The Royal Marsden in London and some of the largest gastroenterology centres in Europe, and they have all been so positive about it, and I think again the fact that they’ve, they want to keep our AI there and they’re seeing the benefits is definitely testament to show that. You know the benefits that this AI does bring in, world leading cancer hospitals and normal hospitals worldwide.
Sam
Guess in an ideal world, every hospital would be using this.
Alicia
And now of course, with our newfound support of Olympus, that’s what they’re enabling us to do, to get our, our AI out on a global scale so we can see survival rates, you know, improve massively.
Sam
And patient feedback has been an important thing throughout this process, which we briefly touched on before, so how, how have you involved patients in the research, what’s their feedback been?
Alicia
So patient safety and user satisfaction are key goals in the development of Odin vision medical devices. So, we want to include the patients… that our technology will be used on, in the design of our medical devices and, effectively their, their feedback shapes the future of our medical devices. So, as you know from our PPI sessions, we found that patients were positive about the use of AI and its benefits. It’s but also as part of our collaborative study with Cardiff University and NIHR we run a Patient Steering Committee where we discuss our medical devices and future plans and we also use this opportunity to gain insight into how the public feels about artificial intelligence being used by doctors for treatment and screening of conditions.
Sam
Great to see that people are so positive about it, and I think it’s obviously when it’s, when it’s something that’s going to impact your healthcare, that’s something that you really need to think about. And yeah, I think for me, it’s yeah, anything that, that would help my doctor to diagnose me and see me quickly and give me my results quickly, definitely a bonus, but again, it’s that making sure that the doctor is still very much present. And again, it’s not just, farming off all sorts of healthcare work to robots and then not sort of seeing your, your doctor. It’s very much still the doctor is there overseeing everything, checking everything and then communicating. With you as, as well, so you don’t lose that human touch, do you? Which is so important to people.
Alicia
Definitely. And of course, that’s something we found in our PPI sessions which. Yeah, definitely. We don’t want to remove that that factor at all.
Sam
And AI is such a hot topic right now. Opinions are a little bit divided again. I think that when it comes to healthcare, people who sit in this kind of focus group, they get to hear about the benefits and how you’re making sure that it’s all safe and robust and all of that kind of thing. But what do you think of the general public’s attitudes to AI at the moment? And how does that affect the work that Odin vision is doing?
Alicia
So yeah, the general public is becoming more and more aware of how artificial intelligence works, and they have more interaction with new technologies in everyday life. As I’m sure we’ll, we’ll see in our day-to-day lives now. And because of this greater understanding, the general attitude is definitely positive. You know about the use of AI to lend a helping hand, you know, in in this industry to doctors lending a helping hand to doctors and the use of our medical devices and there have been a lot of news reports recently about, you know, the potential negatives of AI. So, it is definitely important to us to work with the public so they understand. How you know what we are doing has a positive impact and as I said before, you know we do not want to replace doctors with AI, but just have it there as a second pair of eyes and a support system for doctors.
Sam
Yeah. A safety net kind of thing.
Alicia
Yes, exactly.
Sam
Absolutely. Well, Alicia, it’s been great digging into all of this with you. Just one final question. What do you hope that the impact of AI will be on patient care in the future?
Alicia
So, we would definitely like to see how the combination of AI and doctors can help revolutionise patient care and help to significantly improve survival rates globally and give access to all patients, and we want, you know, I know a lot of people know about this, but we want more people to know that this AI is here, because I’m not sure how many people know because yourself, did you know about it before we had done this sort of PPI session?
Sam
I mean, I knew about AI as in kind of chat bots and that kind of thing.
Alicia
Yeah, but with the colonoscopy and…
Sam
Absolutely not. No, I didn’t really know anything about kind of AI in, in, in healthcare. I mean, I’m from a digital marketing background. So similar to yourself and yeah, so I kind of associated AI with digital marketing and how it can be used there. But yeah, I didn’t really have a clue about how it could be used in, in other applications. And I think that, that Healthcare is an area that could really benefit from it, it sounds like.
Alicia
Definitely. It’s. It’s exciting to see the impact that this will have, let’s say in, in a few years from now when it is out there. And yeah, definitely just to improve patient care for all is what we’re aiming to do.
Sam
Very exciting stuff and excited to keep an eye on this and yeah, see, see what happens in the future and where it all goes next. Alicia, thank you so much for chatting to me today. Lovely to have you.
Alicia
Well, thank you so much for having me.
Sam
Thank you for listening to Can I Butt In. This podcast was brought to you by Bowel Research UK. Find out more about the charity, our work and how you can get involved. Visit bowelresearchuk.org where you can join our People and Research Together network or PaRT, read about our research campaigns and fundraising, or make a donation to support the vital work we do. Let’s end bowel cancer and bowel disease.