RaySearch - Johan Löf

Thank you very much. Yes, as I said, my name is Johan Löf. I am CEO and founder of RaySearch, and we are a pure software company that develops software to treat cancer. So I will start with our basic problem: cancer is a growing problem. In 2022, there were 20 million new cancer cases in the world and 10 million people died from cancer that same year. And it is growing, as I said. By 2050, it is expected – these are figures from WHO – that 35 million new cases will be diagnosed. And radiotherapy, which has been our main focus so far in the company's history, is a very important treatment method because more than 50% of all cancer patients receive radiotherapy, and in some countries it is much more, and this number is increasing. And to see how large our addressable market is, we can note that there are 8,000 radiotherapy centers in the world, and it is growing because in many parts of the world there is almost no access to radiotherapy. So a little background.

RaySearch's vision is to provide all possible software for cancer treatment, not only radiotherapy but also chemotherapy, surgery, and other applications like immunotherapy. But the three most important today, with which most patients are treated, are chemotherapy, surgery, and radiotherapy. And here we see an illustration of a comprehensive cancer center where one can deliver all types of treatments. Many patients receive a combination of these different methods. But as I said, so far we have been focused on radiotherapy, and we have a dose planning system, our main product today, called RayStation. And the parts we come into contact with in such a comprehensive cancer center are, excuse me, the machines down in the basement because they need steel protection. Then we have a room upstairs which is the treatment planning department, where the actual dose planning takes place elsewhere in the center. To describe a little what dose planning is: it is about optimizing and creating a treatment plan for the patient. Once a cancer tumor has been diagnosed, we need to figure out how to best irradiate it, and that is optimized in this dose system.

The radiotherapy machine is called a linac, commonly in the industry, and it can rotate around the patient and send beams from different directions – a very advanced robot. So now we see a view inside RayStation, our dose system, and we have a 3D model of the machine. We model the machine, how it can move; it has a lot of advanced equipment to shape the beam. It can rotate around the patient and send beams from different directions. To optimize the treatment, we need to understand how the machine can behave mechanically and the physics – how the radiation beam looks.

Medical methods to map the tumor. So RayStation in this case can read in all this image material and build a three-dimensional model of the patient where the doctor can identify where the tumor area is, what should be irradiated, but also healthy tissues around. In this case, you see the spinal cord in purple in the background, for example, which is a structure you are very careful not to give too much dose. So here we see an optimized dose distribution. The goal is to get the dose as low as possible to the healthy tissue. The red you see in the image has almost the same shape as the tumor, so it overlaps the tumor, and then there is a very sharp dose gradient so that the dose decays very quickly to the healthy tissue. And you see it from different slices: a cross-section, frontal, and side view. On the right side, you also see the machine, and in the lower right corner, you see the opening from a certain angle when you see a contour of the patient in the background. The black area is exactly the machine's opening from that angle where the beam will be let through. By radiating from different directions in this way, you can achieve that fine dose distribution.

A little about RaySearch: We have 12 offices, 19 distributors to sell around the world. Today we have 394 employees globally, 38 in the US, 328 in the EU, and 28 in APAC. 285 people work at our headquarters here in Stockholm. You see that 203 people – which is almost everyone here in Stockholm – are working on R&D, so more than half of our employees work on research and development. The business model: we are a software company, we run the classic license model: perpetual licenses, then you pay support, 12% annual support. And to give a sense of the amounts, it varies a lot, but on average, a new RayStation sale gives license revenue in kronor. But we also have a lot of revenue from our installed base: they buy more RayStation units. Each RayStation can be equipped with modules; today there are over 20 different modules that they add. These modules provide various advanced treatment techniques and other things so you can do more with the system. So existing customers buy more of these, and then support revenue. Our installed base is extremely important; it accounts for the majority of revenue at any given time, and new sales to new clinics are a smaller part.

A sense of who we compete against: there are not many players; mainly Varian and Elekta, which have systems corresponding to RayStation called Eclipse and Monaco. Philips Medical has an old system called Pinnacle, but it is still alive; by 2026 it will be taken out of service, so only two competitors remain. For our oncology information system, which I will describe soon as our next main product, we have the same competing companies and two competing systems called Aria and Mosaic. It might be interesting to see how our revenue has developed over time, going back to 2008. A couple of interesting things to note: the pandemic hit RaySearch quite hard. There was steady growth until 2019, and then for the first time ever, RaySearch's sales went down, and those were the only two years since the company was founded in 2000 that we posted a loss. But when the pandemic subsided, growth returned. The last bar in this chart is up to the most recently published figures, which is Q3 2023 and last twelve months for that bar. We release the Q4 2023 report in a couple of weeks. Details for Q3: we had a slightly reduced order intake of 3%, sales for the first nine months grew 25%, EBIT grew from 22 million to 70 million, and operating margin from 3.8% to 9.8%. We have said that within a three-year period, by the year 2026, we will have an operating margin of at least 20%, so we have some way to go, but we are heading in the right direction. Cash flow was strong at 145 million compared to before, and we had the largest ever backlog of almost 1 billion kronor. We have four products: RayStation, which I talked about; RayCare, which is an oncology information system; and RayIntelligence, which is AI-based for analysis and understanding population dynamics.

So here, common linacs treat with photon radiation, that is, electromagnetic radiation with high energy. We naturally handle electrons, and more exotic particles like protons, carbon ions, and neutrons. We also handle brachytherapy, where you insert radiation sources into the body. This is just to show how many different things RayStation handles – different treatment techniques. What you see in the top row: 3D-CRT planning, IMRT, VMAT, all done with a standard linac, but different techniques. But we also have pencil beam scanning, double scattering for protons, ion planning, helium planning, BNCT (slow neutrons combined with boron injected into the patient for an effect) – just to illustrate that it is very multifaceted. It illustrates that we support an incredible number of different machines. Because we are neutral, we have no own machine; we are a pure software company, so many actors – I should say almost all actors – come to us for help with dose planning, and new companies emerging in India, China, Japan come to us for collaboration. Here you see different functionality that exists in RayStation. You can note that machine learning is becoming increasingly important. We have two applications today where machine learning is very powerful: first, how to segment organs in the body, and second, when creating this dose plan.

There are different ways to adapt the treatment over time. The patient is treated for maybe six weeks, and things happen with the tumor and the whole anatomy. Then we can adapt the plan based on those changes. Adaptive therapy. Last week we announced we have over 1000 customers. The 1000th customer was a center in India. So we have 1001 RayStation customers. Here you see how they are distributed among our main regions. You can also note that we practically never lose customers. We got our first RayStation customer in 2009, and over all these years we have lost about 20 customers. So there is incredible stickiness, and our installed base is extremely valuable and important to us.

This picture shows a large customer we have in Austria, with carbon ion therapy, which is the most advanced you can do in radiotherapy. That ring in the middle is the synchrotron, the accelerator that accelerates carbon particles to about 70% of the speed of light. To understand the dimensions, it is 30 meters in diameter, and then it leads the beam into three different treatment rooms for delivery to the patient. You see one treatment room on the right where the beam comes out of that gold-colored square hole, and the patient lies on a robotic couch for treatment. The walls here are 7 meters thick concrete, so very massive installations. But here our products RayStation and RayCare command control the entire operation in this advanced clinic. And within particles, protons and carbon ions, which are the most advanced in radiotherapy, we have total market dominance with maybe 85% of the market. We have 116 installations in 21 countries, and the distribution is fairly even: 39, 41, 36 you see in the tables, which particles we handle and all the different machine manufacturers we work with. For example, with the Belgian company IBA, which is world-leading in protons, we have 55 joint installations.

We move over to RayCare, which is an SAP-like system, an operations management system that handles workflows, scheduling, patient records, billing, and everything needed to run a radiotherapy clinic. Some characteristics of RayCare: it has been built with very modern software, so it has a stable and scalable architecture. Workflows are what we call active workflows: they help users through a treatment process that may consist of maybe 78-100 tasks performed by different people at the clinic. The system helps – you click on each task and the system provides services to perform that task. It is very closely integrated with our dose planning system RayStation. We are also looking at resource optimization, so we optimize not only the treatment per patient but also consider what resources we have at the clinic: which machines, which staff, and what resources we have overall, and how we can use them to give the patient population the best care.

We have 24 RayCare customers in 11 countries, and you can see which ones here. It is early, but we have very good momentum now for RayCare. This is a small slide we show: Newsweek does a survey every year where they look at many different types of hospitals, but specifically in oncology hospitals. In the latest survey from 2023, it turns out that of the top 15 clinics in the world, 10 have RaySearch systems. The symbol you see there symbolizes RayStation. But also at 12 and 15, there is another symbol meaning they also have RayCare. So we have two-thirds of these top clinics. In terms of number of systems or clinics, today we have 1000 out of 8000. But among these most advanced, we have a very large share. Our goal is naturally to increase this – maybe not reach that proportion overall, but to get well above 10% market share. We are working on that. Finally, a few words about RayIntelligence and the idea behind it: while treatment is ongoing and managed via our systems RayStation and RayCare, patients are treated. We can capture data from the workflows I described earlier, so we can automatically absorb all data created around each patient. That data is sucked into the target of RayIntelligence, where we can process it, transform it, and understand it on a population-based level. And that is illustrated by the networks going back into the systems. It is not only machine learning – you can do it other ways too – but it leads to us being able to improve our algorithms, improve the efficiency at the clinic, help with decision support: which treatment is best for this patient, which combination of surgery and chemotherapy is ideal for this patient. And most importantly, try to achieve improved treatment outcomes. So that was the last slide. Thank you very much for your attention.

Yes, I mentioned them quickly earlier. One is that we support so many different machines; we support everything out there. That is a very strong competitive advantage: if you have a clinic with mixed machines, many different machines, RayStation is a single platform so you can have one system, all staff work with one system, you only need to train everyone on this system, and you can treat on all those different machines. That is one thing. Another is calculation speed: we calculate incredibly fast. We are constantly working on speeding up our algorithms. Dose-dependent calculations that take hours or at least many tens of minutes on competing systems take a few seconds in our system. Those are two important things. And then that it is so multifaceted with functionality; we have an incredibly large toolbox with different things you can do. So those are our competitive advantages, I would say.

Yes, it will come through revenue growth. We have set a very clear target of 20% operating margin, and I am not at all worried about achieving that by 2026.