Cosmo N.V. (COPN) Earnings Call Transcript & Summary

Ladies and gentlemen, welcome to the GI Genius Q&A session conference call. I am Paul, the Chorus Call operator. [Operator Instructions] The conference is being recorded. [Operator Instructions] The conference must not be recorded for publication or broadcast. At this time, it's my pleasure to hand over to CEO, Alessandro Della Cha, of Cosmo Pharmaceuticals. Please go ahead, sir.

Good morning, and good afternoon to everyone. Thank you very much for joining our conference call, which, as it has been announced, is particularly crafted in order to allow a Q&A session. We would like to brief you on the potential. We would like to introduce you on the background and future perspective of GI Genius and entertain you on the business potential. We're going to have 4 speakers today here. There's going to be Mauro Ajani, our Chairman; Andrea Cherubini, who's the Head of our Artificial Intelligence Business; Nhan Ngo Dinh, who is the Head of Research and Development for AI; and myself. I'm going to pass the word now for the introduction to our Chairman, Mauro Ajani. Thanks.

Good afternoon. I am Mauro Ajani, Founder and the Chairman of the company. First of all, thank you for giving us the opportunity to present GI Genius and the news that came from the market. Starting from the numbers. Every year in the world, there are about 1.2 million new cases of CRC, and we have about 600,000 deaths per year. It's the second cause of cancer mortality. And let me say that despite the other tumor, other cancer, there is a peculiarity that is faced to the early detection because, thanks God, not all the polyps became cancer. There is no cancer that was not a polyp. This means that detection is really, really important just to prevent the disease. And there is no better than endoscopy as a tool that can play in power of this topic. Let me say that the -- what is happening and what's happened in the world with the pandemic is a really disaster for the screening because there was a lot of -- more than 80% in the U.S. in the screening colonoscopy. And the same, let me say, we are now the rate data is more or less the same in all the Europe. This means that the -- not only for the colonoscopy, even for the other screening cancer test, this means that the new pandemia or the next future year can be cancer. When we talk about colonoscopy, we have to take in consideration that it's a very difficult procedures. And let me say that we have to thank all the physicians that save the lives every year, every day because the complexity of the organ, the complexity of the procedure make everything very difficult. And even though there were a lot of investment, massive investment in colonoscope in the last 30 years and let me say that we achieved a very good result because now if you compare the colonoscope of 30 years ago that you cannot compare. So technology is so improved that is even difficult to believe that we can have a new improvement in this field. But despite this improvement and despite the massive investment, we still have a big problem that is the miss rate of the polyps. So as I said, this rate belong to the complexity of the procedures because it's very, very difficult even though the technology has been improved, there is a, let me say, paper now. Colonoscopy is difficult in its own. It's a compass test since the colon, the organ we export is complex. Indeed, it must be said that endoscopists are real magicians in being able to find lesion of the colon while at the same time, having to worry about navigation with a problem along the testing as well as keeping the duration of the procedure as short as possible. Let me give you a sort of a metaphor. You can consider the physicians, the endoscopists as a conductor, an orchestra conductor because there are no doubt for us that with GI Genius, we don't want to substitute the conductor. The conductor remain as still the physician. What we try to do is help the conductor. And thanks to the artificial intelligence, we believe that we can do this. And after Nhan and Andrea can explain the reason why we strongly believe in this topic. Now I leave to Nhan. Please, Nhan?

Good afternoon to all of you. My name is Nhan Ngo Dinh. I'm the Head of Research and Development for AI and Biomedical IT at Cosmo. Today, I would like to provide you a quick overview of the development process beyond our intelligent medical devices and of course, focusing even on GI Genius. So we started the development of GI Genius many years ago when we started collecting data back in 2013. The data collection initiative was conducted during the Methylene Blue MMX Phase III study. And at that time, the difference was made by the management vision on how data would then be used. In fact, data has been collected in a way no other company is collecting them even today. Data is composed by clinical data and video. Clinical data has been collected with very tight control in a worldwide cycle, more than -- was even -- was due even for a Phase III study in endoscopy as the Methylene Blue MMX study was, bringing the drug standards in this field for the first time. And then secondly, full video has been recorded without losing any detail in a mode, which is technically called lossless. A recording with special feature was not even available at the time, and the team has to develop and design and develop it to allow this recording capability. Then the study was involving key opinion leaders in Europe and the United States, the best endoscopist you could find at the time. So we had a privilege of being able to fully record their performances. Their ability to find lesions in the colon was, therefore, key to allow us to build a huge database of all types of lesions from those that are easy to be found to those that are very difficult. So we have top-class data from top-class endoscopists. So what's then next? We had to build a hardware platform. Here again, the vision has been crucial. We were not thinking to build an experiment like a literal experiment or a nice tool for endoscopists because we understood that we have to bring the break-through innovation all over. So we decided to build a platform that was able to deliver best-in-class performances. To do so, we decided to get the most powerful and flexible technology to give us full trust to the algorithm that then we would have implemented. And we design it and then develop a solution that could deliver all this in the endoscopist room conveniently. And I can say none existed at the time. So we have top-class data from top-class endoscopists, and we have a breakthrough platform. So what's next then? So Cosmo has a deep background on GI -- in GI and has always been looking for improving the standard of care. Most importantly, while its detection has been at the center of the Methylene Blue MMX effort, differently from the drug development program, the platform engineering is something that you design and build comprehensively. And this cannot be done without understanding what are the 2 mechanisms involving in the process of discovering polyps in a colonoscopy. So to give you a full overview of all this, I'm handing over to Andrea Cherubini, Head of Artificial Intelligence, who will discuss about the journey we follow to get to an algorithm.

Thank you. Good afternoon to all of you. My name is Andrea Cherubini, and I'm the Head of Artificial Intelligence of Cosmo. Mauro Ajani, in his introduction, has mentioned the clinical problem that is at the core of GI Genius. So colonoscopy is a fantastic tool to prevent colon cancer, and the introduction of cleaning polyps has helped in reducing the incidence of this type of cancer. However, colonoscopy is less than perfect. And it is known by the GI community, that lesion can be missed. This is proven by the fact that the ADR, a parameter that measures the ability of the endoscopist in finding adenomas in patients is highly variable from physician to physician and even from morning to evening for the same physician. Since it has been shown that a 1% increase in ADR correspond to a 3% decrease in interval counter, it is easily understandable how this ADR variability translates to create a problem at patient level. I believe that it is possible to understand the difference between our vision of an intelligent medical device from a simplistic application of artificial intelligence to care if we consider in some detail, the clinical problem of missed lesions in colonoscopy. As Nhan Ngo Dinh mentioned before, in Cosmo, we do have excellent data. We do know AI science, and we have assembled wonderful chemos with highly skilled scientists. But we would have failed if we try to build an AI that tries on its own to identify lesions in colonoscopy, replacing the physician. Such an AI would always be outperformed by physicians who are called every day to a very different job. We know that the endoscopy during the procedure is doing several tasks at the same time. He needs to navigate through the colon, which is a very complex task while looking for lesion and doing it very quickly to minimize the patient's discomfort. We believe that is simply not realistic to replace the physician. Our aim was to design an algorithm able to work together with endoscopist rather dynamic an algorithm which does not interfere with the endoscopic battle again cancer. In order to reach the same, our approach was to design and our science-driven intelligence system. So what does it mean? We started asking ourselves how does the endoscopist bring work. Only understanding this, we would be able to identify the real bottlenecks. Briefly put from a narrow science perspective, the physician performing a colonoscopy faces a problem of resource optimization, in this case of brain resources. As we said, we need to perform an ligation while looking for lesion, a task, which is called [ viso ] search. We analyzed very carefully where this topical fail. We modeled the process, and we observed which situation during the video procedure could lead to a potentially missed the lesion. Subsequently, we specifically designed and trained an AI able to deal with this potential problematic solution. In other words, we first finalized the requirements and only after we designed the technology. I would like to mention here that real-time AI able to present on the display indication for the endoscopist with a 0 delay, a super reactive algorithm, able to anticipate the moment of endoscopist perception. And in particular, we designed an AI, which is able to work even with the fast-moving, blurry and dark lesions. The result of all this, we believe that we obtained a clear game changer, an intelligent device, which is able to work together with endoscopist and which blends in with everyday clinical activity. Our success is demonstrated by clinical results. Our sensitivity is 99.7%. So GI Genius sees all regions, and this was a basic endpoint. Most importantly, we are also able to anticipate endoscopist perception, thus minimizing the chance of missing the lesions. In a clinical study on 700 patients, the ADR increased from 40.4% to 54.8%. And we have seen what an increase of only 1% in ADR mean as you can do the math and understand why we feel that this is a game changer. Let me hand back to Nhan Ngo Dinh, who can summarize what we just said.

Thank you, Andrea. And so at the end of it, we have GI Genius. So this is a top-class intelligent medical device that changes the game in colonoscopy case screening. And you already know about it and how it can be effective and also thanks to Andrea giving you some numbers. However, let's focus on what we've built with GI Genius. A full stack of competencies ranging from clinical expertise to data harvesting and management to establish the clinical partnerships, to hardware and software engineering, to regulatory capabilities, to production and much more. So endoscopist it's not -- it's more than just colonoscopy screening, of course. So we have go colonoscopy as a whole, and we have a gastroscopy, endoscopy and combating techniques, including surgical procedures. However, there is a common ground, the probe. There's always a probe here, the room, endoscopy room; the physician, the endoscopist conducting the procedure and, of course, the subject of the investigation, which is then the internal mucosa. So let's see those items here. So we have the probe. So we engineered the product that is compatible with all the end of endoscopes. And those that are not compatible today, they are going to be set as compatible in the near future. Then about the room, we designed our product to be -- for the endoscopy room and it perfectly fix this use case. And then the physician. So we developed a full technology about the endoscopy, how the endoscopists behave during procedures, as Andrea was showing you. So we define it the relative cognitive tasks and understood the full landscape. And we developed an asset to work together with the physician. And then as a final item, the internal mucosa. We gather a huge database of videos, and we are continuously acquiring new data about internal mucosa and related pathologies. So summarizing, As you can see, we have an outstanding platform of knowledge and technology that enables us to support endoscopy comprehensively. And there is where we are headed to. Thank you.

Thanks, Nhan, and thanks, Andrea. A few quick words on the business potential. First of all, it is important that it realized that our partner in the global distribution is Medtronic. Medtronic is the world #1 Medtech player. Its might is unparalleled. And from what we can see right now, they have thrown all of their weight in order to press for the success of GI Genius, which is, in their expectation, going to be also a very significant contributor over time in their revenue lines. One of the things that we mentioned to give you an outline on what is happening is that the news about GI Genius and the importance of the GI Genius has been reflected in the publication of 64 different articles that have been occurring since the approval of the device. Now I think the device has been approved only very recently, 20 days ago, on the 12th of April, and it's already 64 article out talking of this. And the reason why everyone is talking about this that this is the first time ever that artificial intelligence hits a mass sector in health care. There's nothing as big as the colonoscopy market, if one thinks a bit very carefully with 19 million procedures in the U.S. only. You can imagine the intense activity of Medtronic on the social media and on the social networks, LinkedIn, Twitter and so on. The reason why I am lighting this is because they have reported that as of today, they have received 152 million contacts, which was mind-boggling for them as well. They wouldn't have expected such an answer from the audience, which has translated in 9,000 e-mails just received last week in order to inquire about GI Genius and how to get it. Now let's talk about the colonoscopy market. Again, in order to understand what is the potential here that Medtronic is dealing with. I've been telling you 19 million colonoscopies, number are relatively approximate. It's very difficult to receive -- to find out the precise number, but this is what we have according to the latest statistics. Now 19 million colonoscopies in the United States, according to the information that has been provided by Medtronic, are executed through around 32,000 endoscopy towers in the whole of the U.S. And keep in mind that we're talking only about the U.S. market potential now, whereas the distribution agreement that we have with Medtronic is global. So it's 32,000 towers. Clearly, not all towers perform the same number of colonoscopy every day or every year. But potentially, every tower is acceptable to be integrated with GI Genius. So that is what we're looking at in terms of our overall market. Medtronic, I cannot be more precise here, but Medtronic is foreseeing -- as you remember, we have discussed this already in the past, is for foreseeing user subscription fee model. And they are assuming that the yearly subscription fee will be no less than USD 40,000 per device. So if you multiply $40,000 per device for the number of towers, then you have a dimension of the potential market that we're looking at. When we made our statements in the -- in our press release announcing the approval of a market in excess of $1 billion, that was exactly what we were thinking of. It's important that you understand that because of the way the U.S. market is shaped, increasing the ADR and making colonoscopy screening more efficacious is not convenient just for the patient who's the one who comes first. It's also convenient for the operators. Because of the way the U.S. market works, on one side, the increased number of lesions found will translate into an increased number of excisions for which the operator will be paid an increased consideration. On the other side, an increased number of lesion will translate into a higher number of histopathological examination, which are reimbursed separately. And therefore, however, you are going to look at it -- and we cannot get too much into details right now. The increase in the detection is not just extraordinarily beneficial for the patient, but it's also bringing new business for the operators in the field. So to come back to what we had, again, originally stated since we have entered into the agreement with Medtronic, Cosmo will make a net profit, and I'm underlying the word profit, will be making an underlying profit in excess of 20% of whatever Medtronic sales. So you should figure out the market potential, as I have indicated, and you should decide yourself what kind of penetration you're going to give to Medtronic capacity of marketing the device, and then you will have an estimate of what could be our revenues and therefore, our profit going forward. So I think this is enough. As we had announced, we were not going to give any specific financial forecast or financial guidance in this session because we're gathering orders and we're gathering feedback. That is something that we will more likely do once we will discuss our half-year report. So I'd rather stop now and pass the word to the audience for questions. Thank you.

[Operator Instructions] The first question comes from the line of Bob Pooler of ValuationLAB.

Thank you for the excellent presentation and description. A few questions from my side. First of all, on the overall market, do you see recovery of the number of colonoscopies? As you said, it's also affected by the COVID-19. Now in the U.S., you've seen that the vaccination programs have rapidly progressed. Do you see also people coming back and going toward the colonoscopy?

We can answer immediately. Medtronic has reported to us a drop of more than 83% in the overall number of procedures in the United States. They're expecting things go back to normal by 2023.

But right now, there are 70% of the hospital that are ready. So there is reopening in the last 2 months.

Okay. That's good. Of course, the U.S. has just been launched there, just approval so -- but in the other countries, what is the feedback that operators using GI Genius give you? Where did they see improvements? Where do they see also new opportunities?

Let me say that what happened in Europe was a sort of test launch. And unfortunately, the test coincided with the pandemia. Let me say that the GI Genius in Europe will be launched now, even for Europe. The market in Europe is completely different. You have to consider something that is country by country because the reimbursement is different. And in any case, the reimbursement drive the business plan. So I can tell you that, for example, U.K., you can consider U.K. looks like U.S. market. But we have very different country-by-country reimbursement in the colonoscopy procedure. We will see. So the numbers that we have and because there is even a big difference. Now the authorization that we got from FDA is a sort of authorization of pharma because de novo in the U.S., you can compare the novo authorization like a pharma because we have to provide a lot of numbers, clinical trials. So it's something very difficult to get. In Europe, it's easier the pathway -- is a very different market.

Yes. Just on the device itself, is it easy that you can hook it up to colonoscopy stack? And also, is much training needed? And do you have like over-the-air upgrades?

So your question about like what is necessary for an endoscopist to use the device, right? .

Yes.

Yes. Well, actually, the device has demonstrated to be really seamless because actually, there's no special training you have to get except just to read through the instructions for use, which are definitely just describing the context and the fact that you just have to plug in the device and do the colonoscopy. You do not have to do special care because the device is just highlighting polyps as they are visible. And so it was really -- we've got an excellent feedback from the field concerning the usability of the product because no one was complaining really on the way the device was working. So false positives were really low. The system showed very high sensitivity, and this was very compatible with the procedure. So it is not even lengthening the procedure, and this was demonstrated also in the clinical trial. We use it for registration. So this was about the feedback. And your question about the upgrades, right? We are -- in this moment, we are very much focusing on delivering this huge innovation to the market. At the same time, we are definitely working on our innovation pipeline. We do not have, let's say, a program for delivering updates from time to time. We would like to give out the possible new algorithms when they are ready, when they are in the position of changing the game. We don't want to commit to our targets that then will force us to delivering something that then is not going to be useful in the field. So our position would be like that. I believe that this is answering your question.

Okay. Just one question before I go back in the queue. Literature says that there's basically 2 main reasons why there's such a low adenoma detection rate, so the ADR. One is human error and the other is actually that some of the lesions are very difficult to spot. Two questions on that. Any idea which is the sort of main reason, so human error or just difficult to spot? And then how can you improve on that with the lesions that are particular difficult to spot?

Okay. So basically, the clinical effectiveness is testified by the increase in ADR. Lesion, we, of course, cannot know if lesions are difficult to spot because they are themselves particularly in a category of this lesion. But we can say this. We had the luxury of having a data set that was recorded by the -- from the procedure of the top endoscopists in the world. These are the key opinion leader and the top endoscopist because they've got a very high ADR themselves. So this means that in our data set, we had a data set we had doubt that was reached with lesions that were not seen by other endoscopist. Training with this particular data set means that GI Genius has gained a sort of ability to detect these lesions. And in this sense, we can say that Genius is able to spot very difficult lesion. Do you want to add something, Nhan?

Yes. Yes. In general, we attribute the real reason for missing lesions to the fact -- as Mauro was saying at the beginning, to the fact that the endoscopist is very busy in doing something else, okay? So essentially, this is the main burden of doing colonoscopies. So you have to do a lot of things together, and you have just to do all of them at the same time. And GI Genius is basically helping a lot because of that. So that's why GI Genius is working because it's helping the physician in a moment when the physician is doing a lot of things all together. So this is just by completing the answer of Andrea and maybe providing you a key for understanding a little bit better why GI Genius is so helpful for endoscopist.

Okay. If I may squeeze in just one more because many people are also asking, what's the role of Methylene Blue leucomethylene blue combined with GI Genius? How do you see that progressing?

Well, Bob, you remember we discussed this a couple of other times. The 2 things are synergistic with one another and complementary. So we see them playing a very nice roll together, and this is why currently, we intend to proceed in the program, as outlined in the past.

So basically, Methylene Blue enhances the imaging and of course, because the GI Genius is based on imaging that would improve also potentially the ADR rate from GI Genius.

Precisely, and the addition of the contrast agent is definitely not going to do any bad but can only make things better. And this is also in the perspective of other tasks that we assume the device will be performed over time. So for us, the 2 things together are completely synergistic with one another.

The next question comes from the line of Paul Verbraeken from Research Partners.

I have 2 questions. First, concerning the number of towers. I mean now you mentioned 32,000 in the U.S. In the past, I think you mentioned 20,000. Is there a different definition? Or is it an updated number? And maybe connected to that, what's your estimate of the number of towers in Europe?

Yes. Well, now we have more precise numbers because what was difficult at the time is the fact that the tower in the, let me say, private clinical market in the U.S. The problem is -- not a problem, the opportunity is that in the U.S., 50% of the colonoscopies are run in the private sector. It's completely different than in Europe. This is the reason why this number was -- we took this number from Medtronic because they did an extensive research that, at the time, was very difficult for us to consider the, let me say, private bracket market. In Europe, we remain with the same towers that we released just 1 year ago. Because in Europe, private is very small, small, small market. In the general landscape, hospital represent more than 90% in Europe, completely different in the U.S. .

And in Europe, what's your estimate then of the number of towers?

We believe it was between 10,000 and 15,000. It was something like that.

Yes, less than 15,000.

But keep in mind that, unfortunately, you will remember probably we touched this already, there is not even a precise number of colonoscopy in Europe, whereas we have more precise data in terms of number of colonoscopies in the United States, this is definitely way less precise in Europe.

Okay. Maybe one more if I can, how well protected is the Genius technology? Or how difficult to replicate it? I think you mentioned that you use a special video camera to get to the data? How hard is it for competitors to get their hands on similar equipment?

Well, look, I think that you should see that the real protection comes from the process and the amount of competence that you need together during the process because you need to have the hardware, you need to have the software, you need to have the data and you need to have the people capable of assembling. We're not assuming, of course, that we're going to be unique. There are many ways to play the same task. And even though we have a consistent IP portfolio covering our device that I assume is always something that you can work around in a way or in another. I think that the best evidence now that one should concentrate on is not so much the IP or the protection of the know-how, but it is the fact that this is entrusted in the largest player, which is now playing without competitors in the market. So it's the time advantage that we have versus the competition rather than anything else. And no other clinical trials, at least to our knowledge, are in sight. And because of the de novo application and because of the fact that we have been setting standards with the FDA for AI application in this field, not only we are groundbreaker, but we have the advantage of the first move and therefore, the other have to follow up. So the real protection that comes here is the time gap. I'm assuming that, let's say, in 3, 4 years' time, there will be other players that will be selling similar devices, of course, but important for us is that we are the first one, and we will be the first one for quite some time.

And let me add that in our knowledge at the time, if we consider the 2 major possible player that can concur with against us or the Japanese company, the strategy is completely different because we developed a typical intelligence platform. They develop something different because they are improving once again that colonoscope, means that the real deal for the other company is try to improve the colonoscope and sell the colonoscope. For us, we have developed a platform, an artificial intelligent platform. This is the big difference.

There are no more questions at this time.

Okay. Again, we thank everyone for their participation and every one of us...

Sorry to interrupt, we have a last-minute registration coming from the line of Bob Pooler.

Okay. I have a couple just on maybe potential future upgrades of the GI Genius as you saw the technology platform that you can roll out. What your ideas on, for instance, optical biopsy and also additional GI and other indications with the scope? And what I think you call mandate as is also procedural documentation as you clearly state that the endoscopist has to do a lot of things, but they have to report each individual colonoscopy. And if you would add a procedural documentation to that package would help them too. So just a little bit what is the flavor of potential upgrades based on this technology platform.

Yes. Well, I was saying in my introduction here that we are -- have developed a full set of competencies. And so actually, to bring innovation then is going to be much easier than it was at the beginning to get GI Genius done for the first time. I can tell you, we are working on a lot of innovation content. For the moment, we are definitely focusing on delivering detection and learning whatever lessons we can learn from that. And so I can confirm that definitely optical biopsies in the range of the technologies that we are working on. And that's what I can say at the moment. And for sure, just to reiterate, it will be much easier for us to bring innovation further than it was to build the first technology. So it's going to be the way we are going. So as I was saying, we are interested to support endoscopy comprehensively. This is our strategy and our goal. And as Mauro was saying, we've built a platform for doing so.

So basically, Cosmo remains responsible for these new developments there. And just one follow-on question. What's the need for approval? Now you have the de novo approval, so you have a large advantage there. But for potential add-ons, will this just be based on the platform or clinical trials? How do you envisage that?

Well, actually, the development of new tools would require a regulatory pathway. And this is depending on the type of tools and there's going to be a set of requirements that then the regulatory is requesting to clear the device. And for sure, for some of them, some clinical trials will be needed. And of course, together with whatever is to require then to validate the platform. So it's very much specific to the new tool that are going to be deployed, the type of data that is required by the regulatory.

There are no more further questions at this time.

Well, if there are no further questions, we're thanking everyone for your participation. And I just want to say that for whatever additional question, you can contact me directly or anyone else on the team with an e-mail. Maybe it's preferable if you address the e-mail to me, and I will then forward it is a relevant team member, but happy to answer separately also to any other query you may have. Thank you so much, and have a nice day.

Thank you.

Thank you. Bye-bye.

Ladies and gentlemen, the conference is now over. Thank you for choosing Chorus call, and thank you for participating in the conference. You may now disconnect your lines. Goodbye.