BoneChat #120 - AI-Powered 2D to 3D X-Ray Tech: The Future of Orthopedic Surgical Planning

Okay, this is Bone Chat discussion number 120. We have Laurent Paul, who's the CEO, founder of 3DSIDE. This should be kind of interesting. Laurent and his team have created a technology that produces a 3-dimensional scan like a CAT scan, MRI using 2D plane X-rays. And his company has kind of two offerings. They provide software for patient planning and then they also deliver, I think, patient-specific cutting guides for various orthopedic procedures. I always think of it as Materialise, if everybody's familiar with Materialise, but simpler imaging. So, but I could be way off. I'm going to hand the mic over to Laurent, and we're all going to learn a lot right now, and I'm going to put his information in the chat. Take it away.

Thank you. So I'm going to share my screen. Um, sorry to full screen. Oops, wrong thing.

Oh, I gotta get caught up.

Yeah, we'll keep it this way if you don't mind. So thank you for the invite first, and I'm going to speak a little bit about myself. So I'm the co-founder of the company acting as a CEO right now, and my background is about orthopedic oncology. I developed a few solutions preoperative and intraoperative, and back 25 years ago I used to develop— I developed a navigation system. That was my main R&D topic. I led a PhD thesis that I defended back in 2009. And since then, I tried to, to push a little bit further, turning this project into a product that I managed to, to do. And, um, back in 2015, I created that company 3DSYDE with my co-founder. So I'm going to, um, go a bit further on speaking about the company and the products. So we have two families of products. One is the 3DCut family, private label. So we are pushing the product and bringing them on the market under our own name. We are in contact with customers, so surgeons and users, B2C then. And the second part of the family is the 3D Artflow family. That is a turnkey product. We are a bit behind the scenes serving implant manufacturers as a turnkey product. Then this latter family is a bit more recent. It all started in 2020 during COVID actually, and we started the activities, developments with a few, a few customers. We reached the market in 2022, and since then we, we grew the business. I will give a few words later on. So I'm going to speak a little bit about other products first, and I will slightly come to the AI developments and the software especially. But on products, so the 3DCut product is divided into subfamilies. One that is oncology, that is directly resulting from my R&D project. Um, as said, I initially started with a navigation system, coming from navigation to PSI at some point for the sake of let's say, ease of use and certification. And 3D printing being more and more available, it was definitely an easier solution than going through a navigation system development and certification. So long story short, since then I treated more than 800 cases. By the way, I'm the, the guy who designed and manufactured the very first cutting guide for bone tumor resection. So that's very brand new product at that time. And since then we reached the US market. We've got the fact that Intech is covering this product. Second subfamily within the 3Cut family, that is osteotomy. So it's directly derived from the oncology product. As you can see here on screen, we are treating any limb deformation with or without hardware. Our specialty is to be totally agnostic to any hardware. So we are not connected to any implant manufacturer except if they become first a distributor or a customer of the platform. We, the product is covered as well by the 510 clearance. We then reached the US market 2 years ago and we are getting more and more traction on that product. I will get back to that a little bit later. Second portion of the, the company, let's say, 3D Arthroplasty. So this is the, the family that is dedicated to implant manufacturers. So basically we are serving implant manufacturers with a software and when need be a cutting guide or a drilling guide when they need to, to guide the intraoperative assistance through cutting guides and drilling guides. So we have three products: ankle, knee, and shoulder. The story started with ankle. We jumped to shoulder when we met a French guy who was established in France— in the US. And we recently moved to the knee application. This is where we developed the specific AI feature. In terms of manufacturing, we have 3 printers in-house in Belgium and in Chicago. So we are serving both European and US territories. Again, in case of need, if there's a connection with Cutting Guides, we can run the design and the manufacturing. In terms of software, so the software is in the middle of everything at 3DSYDE and in the middle of the process, of course, we are dealing with images first, but everything is converting to 3D models to be interactively planned with surgeons. We are engaging discussions with sales rep as well. So the software is gathering any party that needs to be involved somehow, from by far, by close to the patient. Of course, the software is powered by AI. I say of course because everything is about AI nowadays, and this software is definitely driven by AI. Few features. The software itself is pretty common, I would say. So you won't find something very different than the competition. You mentioned Materialise at the beginning of the presentation. It's definitely similar to Materialise. You've got a few viewers, 2D, 3D, the ability to position implants or 3D models. The segmentation is built in within the software, full traceability, basic features, let's say. And on top of that, we added a few automations. This slide is summarizing a little bit The main features we developed recently: segmentation, landmarking, automatic sizing and positioning, and X-ray analysis. Everything is homemade, so that means we are building from A to Z all the algorithms, all the networks, including the data. So everything belongs to Freesight, from data to trained models, and we are even the only users of such AI features.

So, Laurent, let me ask a question about that. The segmentation, landmarking, sizing, how much of that is manual and how much of it is done with the algorithm automatically?

So we are, we came from a purely manual operation. I mean, in terms of planning on the arthroplasty side and the private label side. So we were doing almost everything manually. That means we were delineating manually, on every slice, old school, old-fashioned. But since then we developed these automations and right now we consider that these automations are mature enough to go to the surgeon without any manual addition. We still keep a manual supervision over that process because the Fact&Tech clearance is not yet covering the full automation. That's our next on the list. We are going to submit the shoulder application in literally 2 weeks from now. So, and this new feature will be onboarding the new 510 clearance. So in, let's say, a matter of few months, we'll have a fully automatic software solutions from DICOMs to implants being approved by the surgeon.

Got it. Thank you.

Feel free to interrupt at any time. Of course, this is not meant to be a lecture. The most recent development is this 2D to 3D thing that we call internally a converter. So it takes 2 X-rays as input, a face profile, and makes not 3D models, but it generates a CT scanner. You see an example of that here. So the CT scanner, the original CT scanner, the generated CT scanner, and the ground truth comparison. So I'm going to speak a bit about the accuracy and the very nice results we have so far, but I would like to highlight a few new possibilities and applications based on that new feature. So first, of course, it eliminates the needs for a CT scan, so it opens up new doors for new applications such as knee arthroplasty and osteotomy. Second, and nonetheless, it enables weight-bearing CT scan, weight-bearing and full limb, coming from an X-ray that is standing. So when it deals with biomechanics, of course, having a standing weight-bearing X-ray, CT scan X-ray, is very, very well— like a strong asset for, like, simulation of biomechanics compared to a CT scan that would be lying on his back. The third one is it significantly reduces the radiation dose for patients, and last, it's a significant decrease in terms of cost for hospitals. Availability is much better and it's much faster to acquire an X-ray. So finally, all parties could benefit from that new tool. In terms of accuracy, we did a first run, uh, a first training on tibia that was incomplete for the sake of iterate very quickly and determine what was the best network to train. So that's the result we have here. You see a few, let's say, default, like, drawbacks, uh, stripes, um, a certain accuracy that is 2mm, and soft tissues that were almost not present on this CT scan. So we decided to stop the development on the tibial side, considering that it was a very good first trial. We established which network was the best one, and we go— we went to the femur side. So we did the same training but complete this time. So 5 weeks, a bit more datasets as well, and we reached an accuracy of 1.5 millimeters that we consider a very promising feature. As you can see here on that example, there's no stripes anymore, almost no stripes, and we've enlarged the field of view as well. So by going stepwise, we like scratched a little bit of few drawbacks that we observed during the, the T-Bill training. So based on those results, we consider that we are confident enough at least that we can reach the next steps That will be the regulatory approval. So the regulatory approval will be the following. We are going to submit, like, let's say, first do another TBL training in the course of April. So as I speak, the model is currently being trained. We are going to wrap up the technical file for a submission somewhere in June. This tool will be connected to existing 510 without breaking the 510. That's the beauty of that solution. I'm going to speak a bit further about the company, if you don't mind. If you have any technical questions, I will be happy to take them after those two last slides. So readiness to scale. We established the company 11 years ago. We have a few US clearances already. We are running several businesses, Arthoplasty on one end, Private Label with homemade products on the other hand. We have more than 1,000 surgeons, more than 250 hospitals, and we have very significant traction in the US right now. It's growing quite fastly as I speak. We would like to expand our current capabilities by going to other, let's say, joint potential modalities. So of course we'll go to the hip joint. That's going to be the next on the list. We are speaking about shoulder as well as we've been told that there is a bit of interest having an X-ray to be converting into a CT scanner. So that's going to be potential new applications in the future. And we are going to apply this feature to osteotomies as we are running the business with our private label. The connection to other OR technologies, of course, it's not restricted to cutting guides. We can connect any planning to an existing assistancy such as robots, augmented reality, or navigation systems. And we envision a bit more applications in the future for, for instance, lumbar spine, or even going further, coming out from the orthopedic space. To do so, we are going to raise funds mostly to cover US expansion. So threefold, three pillars. First, going, reaching the market, leading the latest steps on this R&D project, going through the regulatory steps for all of those applications. There are some locally, local-based US activities by establishing a US sales team. We are going to have our own US manufacturing. So far it's a local vendor, and of course having a US support as the US business is growing quite fast, and there will be working capital requirements to be covered, of course, as we need to structure the company a bit further in the US. The conclusion: we are building something very significant. We have new technologies and products to reach the market in '26. This new X-ray to conversion, to CT conversion tool, is a game changer in our mind. And the, the US revenues are growing quite rapidly and they will overcome the, the European revenues. I'm going to end my presentation right here and I'm, I will be happy to, to answer any question you would have.

Well, congratulations. I didn't know you were that far in the US. That's, that's terrific. What's the— we all started off— what's the difference between surgical planning in Europe versus the US and what's from a system point of view?

So first, first thing to mention, the process remains identical. Identical. I mean, as we developed the techniques here in Europe and we are kind of leveraging and benefiting from our past experience and patients that we accumulated over the years in Europe. So we maintain the same process, the same internal procedures at least. But I would say that the main difference comes from the culture in Europe. The surgeons are doing everything on their own. I mean, they are doing, for instance, for the oncology product, they are delineating the tumor on the MRI themselves. They are sending the DICOM images themselves on the platform. They are reviewing themselves. They don't need any call. They do it at any time, like 11 PM or something. In the US, we have a— there's a very big gap, very large gap in terms of culture, as we need a sales rep, we need someone to accompany the surgeon, we need to make some calls. So we learned that difference and we adapted our internal process in terms of communication. I would say mostly communication.

Got it. Oh, thank you. Andrea is up and then Kyle.

Yeah, I was wondering how you deal with your due diligence from a patent perspective. It's an incredibly crowded area, especially the imaging. And I know multiple companies have coverage of 2D to 3D claims, especially Stryker, one of the bigger— biggest ones, has a pretty big portfolio.

So there, indeed, it's very crowded. So first, speaking about arthroplasty, There are many patents covering guides especially. We are operating a freedom to operate. We are seeking for freedom to operate first before launching any product and trying to find workarounds as any company. From 2D to 3D perspective, as far as I know, there is no such a tool that is coming from X-ray to CT scan like an intermediary state. That's something. So generative AI basically. Going through a CT scan generation, with a specific application to knee and hip arthroplasty. But we probably need to go through a tour of the Markov fleet FTO.

Yeah, I would highly recommend to do that. I've done something very similar and there is a lot and anatomy agnostic, so not necessarily to, to a certain part of the skeletal system.

So if we face— would we face any challenge regarding IP? So that tool is meant to be internal, so it's going to be an internal tool. We will not bring it on the, on the market, so we, we would be the owner of that tool, putting it at, let's say, a disposal of companies that would be willing to engage with us in the business. So that would be a different, different way to approaching the business.

Yeah, I mean, it depends on where you're doing it. Let's say you, you do it in, in Belgium, you said. If that's where you perform the actual step of, of conversion, you would have to look at, I would say, an IP that covers Belgium, EP, Europe. Yeah, it is interesting if you bring the product that, you know, it has used that process and bring it into the US, how that is viewed in terms of IP coverage. I don't know. But Striker for sure has a big portfolio and they file not just in the United States. So I would take a closer look.

Or Playground is related. Anyway. So we're going to face that situation. We don't want to restrict it to Europe, definitely. So we'll take this point and we'll try to cover that point. Thank you.

Good question. Patrick.

Laurent, I have kind of a, I guess you could say, a two-part question that's kind of regulatory related. First is what, at least just talking about the US, but What within your portfolio requires FDA clearances, even if they're Class I, but especially Class II? And then what, like, what sort of doesn't? This, you know, what software counts and what software does or does not need regulatory clearances, I think, is an evolving and, you know, 10 different people might give you 10 different answers type of field. So that's the first part of the question. And then the second part relates to AI, correct me if I'm wrong, my understanding is that FDA is like very quickly evolving in how they view the use of AI in planning segmentation and all of that. And sort of like what was the rule last year probably won't be the rule this year and probably won't be the rule next year. So especially with homegrown software developed, I think you said in 2018 or '19, like how do you guys think about and approach the use of AI in your planning software as it pertains to FDA interactions?

As we have two kinds of software, one that is Fact&K exempt, that's a communication software that is running, like, powering the private label products. So we are pushing 3D content. The surgeon can interact and answer questions based on what he sees on screen, but there is no interaction with 3D models. He cannot move 3D models in the 3D space. So this static viewer is just a communication platform. So this is perceived as a Class 1 device by the FDA. So we have a letter that, that says that we are exempt around 510. Any software that is powering advanced features such as segmentation, landmarking, interactive movements, so dynamic planner per se, requires a 510 clearance. So we have both, the two families are powered by two different versions of the software, one that is exempt, one that is 510 exempt. So that's for the first part of your question. The second part is very interesting indeed. It's evolving very, very rapidly. So AI is itself evolving every day, every single day. You see many things that are popping up. So the FDA tries to follow the pace, but of course it goes a bit slower. But there's a very recent guideline that is driving us to, let's say, or at least providing a certain framework to work with. And for instance, we are acquiring data to respect certain guidelines from the FDA. For instance, we have to, to get a mix of patients, a minimum 80% of US patients, a certain ratio between gender, between race, between age, a certain ratio about pathology. So if you are targeting knee arthroplasty, you need a bunch of degenerative need in your, in your datasets.

This is to train the models, right?

To train the models. Yeah, correct.

Yeah.

And train and validate and test the model. So all of those steps require a mix that are— that should follow certain guidelines. And of course, we are compliant with those guidelines. In terms of evolution, there's a very nice feature that the FDA has issued very recently. That's, it's called PCCP. And you can anticipate a future. Let's say you are training a network today, but you are going to train a new network in the next 2 months. You are envisioning a new AI coming out. So you are anticipating this and saying to the FDA, I'm going to develop this feature, I'm going to describe it. I don't have accuracy data yet, but once I will have them, I will make them available for your review if need be. So this is like an anticipation on your future development. And if you are following that framework very strictly, you can somehow clear the feature that is not yet developed. So that's a very nice feature that is definitely AI-oriented, I would say. So we're going to leverage that aspect. But on segmentation, we already have a segmentation algorithm that is AI-based and cleared by the FDA for Anchor. And show their applications.

Is it correct? My understanding is that it's basically a one-way street and that like the requirements only keep getting tougher. Is that— is it— do you think that's true or is that—

it's true. Yeah. Like 2 years ago, there was almost nothing, no framework. You could develop AI and show a certain level of accuracy. Now it is— you can't do without respecting those Guidance.

Yeah.

Yeah.

I think it's still a draft guidance, but it's, it's already—

Well, but yeah, but good. You know, you're, you're at major risks that not follow it.

Yeah. Yeah, for sure.

Yeah.

Okay.

Thanks.

All right. Great questions, Patrick. You went right into the software and AI. All right. Dr. Slaughter is up.

Yes. Thanks Tiger Buford for taking my question. Yes. Dan Slaughter, Orthotrauma and Adult Recon in Atlanta, Georgia. So I'm very curious about the X-ray to CT conversion. When I think of CT, I think of, sequential images 1 or 2 millimeters apart. You're able to achieve that with an X-ray?

Yeah, that's correct.

That's amazing.

To get back to the slide, so we are generating a CT scanner of, uh, so multi-slices. What you see on screen here is, uh, the transversal reconstruction, frontal and sagittal.

But not sequential cuts.

Sorry, could you repeat again?

Not sequential cuts like, you know, anterior to posterior sequential images.

You can't show slices, you can't show slices like a CAT scan, can you?

Yeah, yeah, we can. So you can browse through the entire volume as if it was an actual CT scan. We can dedicate a specific thickness to those slices. So we are totally mastering the quality of the CT scan, I would say the resolution. And in terms of accuracy, we are comparing it to the ground truth CT scan. So we had a pair of X-rays and CT scan and we made a comparison of the generative against the ground truth. And that's what we get, 1.5 millimeters.

That's amazing. Congratulations on figuring that out.

Dr. Schlatter, what applications do you see of 2D, you know, plain X-rays into 3D image? And what does that do for your practice?

Well, the 3D information is critical for treating fractures, especially a couple slides before this, there was a depressed area in the tibial plateau. If you can go back one or two slides. To 12. Yeah. So the original CT scan, the middle coronal slice, so you can see the depression in the plateau. I can't tell if that's medial or lateral, but then you lose that in the X-ray generated below that. That's what generated my question. So if there's some pixel averaging or something like that, then you know, it'll— so it, it's, it's great for preoperative information to know where the pathology is.

So, um, it's not really a question, but I'm going to answer to comment on that. So you're right, indeed, uh, the, the contour is not exactly the same on this specific plateau, but, uh, for the— like, our application will be osteotomies, so HTOs or DFOs, or knee arthroplasty. So the The surface of that plateau is not necessarily our highest objective or utmost importance. When it deals with PSI, we need to have accurate surfaces to land the piece of plastic on the, on the piece of bone. So we are going to rely on certain landmarks, and on that aspect, the plateau is not helpful for us. So indeed, there are a few differences here and there, but we are going to focus on certain aspects of the tibia to make sure that the cutting guides are landing at the right spot. Nonetheless, we have to be close enough to make sure that we do not create not realistic CT scanner, of course.

You know, the reduction in radiation alone is, is it would be incredible. I mean, that would be a huge jump. I mean, because a lot of times, you know, my partners don't get a post-operative CT scan or one, you know, down the road because they're worried about radiation. So if we could avoid that somehow with your technology, that would be a game changer.

You mentioned something that was not on my slide, but indeed the intra, uh, power of that would be very interesting if you want to check anything during the, the, the operation, the, the surgery. And the post-op view could be very helpful as well if you want to like reconstruct a CT scan based on a post-op acquisition. You could get it from an X-ray. So in terms of prompts, it could be a very strong and powerful tool to assess the post-op situation. So far we have not trained the model to work with hardware or metal, but that's on our roadmap as well.

Next year.

Next year.

Good question. Kyle's up next.

Yeah, so a couple, couple comments for you. First is when Depew Synthese was doing TrueMatch going back more than a decade ago, what they were trying to do for their 2D to 3D conversion is they wanted to get to a threshold of like 20,000 CTs in the library, and they simply would take new x-rays and try and do a best match with whatever was in their library, and that was their, their pseudo way of getting to a conversion. I assume you're doing things a little bit differently, but I thought that was a fascinating way to do it. And it'd be better to go the route of having it generated for you instead of having to get 20,000 CTs of actual patients on every different indication, probably a problem there. Second thing is I saw you mentioned bunion as far as some of the other things you're doing. I'd encourage you to look all over the place in the foot and ankle space. There's at least 5 different applications in there, especially since when you showed the shoulder, you're doing not only cut guide for humeral resection, you're doing pin placement for the glenoid. If you're doing both of those, there's a long list of things in the foot and ankle space that are gonna be an opportunity there. And if you find yourself looking for doctors, enterprising surgeons that work in that space and you'd like some introductions, just let me know. I know a lot.

Thanks, thanks, Kyle. So we didn't talk about the business model, Lawrence. So I think you sell guides and then you also sell software. So you can explain kind of how, how you make money?

Um, so two families, uh, the private label products, um, those ones are revenue generating, especially in the US. Um, the oncology space is being what it is. Um, we have a very solid traction given our track record, more than 800 patients. So we are generating revenues from the private label products, especially on the oncology side and on the osteotomy side as well, as we are distributed by a few companies that are— one of the companies that I cannot mention here is a very large company. We are going to enter in the business with them. It was kind of public at the Academy, but I'm still a bit reluctant to communicate over it. You, you will see it in the news very, very soon, most probably. So this gonna be a new driver for revenue for us in the coming months and years. So that's on one end, products. The software comes for free, so no one pays for, for the software. The surgeon uses it for free, so there's no revenue from the software on that side. On the arthroplasty side of things, we are engaging discussions with customers and we are developing, tailoring a solution for for their implants, their key opinion leaders, their needs. So we are generating revenues from development first, and then when we reach the operational phase, we are licensing the software. So we are getting revenues from the software usage to cover maintenance, support, and all of those fun activities. And manufacturing cutting guides, like producing virtual data as well. So this service comes as a fee per patient.

Of course. Got it. Wow. What— so what's preventing you from just giving everybody the software for free and open, open source? And then when they want to buy a guide, they have to— obviously when they need a product, they have to go through you.

Hmm.

I don't know.

Just— yeah, no, interesting, interesting comment. Open source has definitely been very popular back 20 years ago. It's still very popular. There are some business models, one that is very popular in Belgium, for instance. They are, they are called the Unicorn of Belgium. As I speak, as a friend of mine, by the way, and it's running a business model that is purely based on open source software, and the support and maintenance are coming against certain fees. So Yeah, not so crazy. It's not a crazy idea, definitely.

Yeah, because I'm thinking, I— my only, my only benchmark is Materialise, and Materialise basically creates guides. They sell guides for complicated osteotomies or oncology or, you know, crazy trauma. But they don't sell the software either. The software is kind of a pass-through.

What they sell is a license, and the license fee is pretty expensive. We know a bit about the background things when it deals with software and PSI, and we know that the development cost is pretty high and the license is quite expensive. And running the business ourselves, we definitely understand why. The maintenance and the support is definitely a big task. So we need to overcome those hidden costs somehow that come on top of the production per se. So that's definitely a business that is very hard to be totally open source unless you are going to offer some side services such as offering support and maintenance against a fee.

Interesting. Mark, oh, you're on mute.

I like the idea of the open source idea. It allows you to really, really penetrate the market fast and then develop your revenue model a different way. Uh, so that was a good thought. Lorin, I see that we have almost 100 mutual connections on LinkedIn. And I remember that we, in about early 2000s, when we were developing early digital imaging software, the open source that we borrowed from Switzerland was Osiris. It was very popular, very used by a lot of Mac users. You've got a really smart-looking board too, look, smart-looking crew. This has really been an interesting invitation, Tiger. Laura, I'm really happy to hear. I missed the first 20 minutes, but I was trying to catch up pretty quick.

Offline if you need to.

Where's your base in the opera— where's your base of business in the United States so far?

Um, so we are— we have been incorporated in Delaware. Um, we are running the, the manufacturing in Chicago through a vendor, and we have a local footprint in Denver, uh, the army of one sales guy that is located here in Colorado.

Ah, they're all over.

Thanks. Yeah, Mark, Mark was one of the guys at eTrauma in their early days, one of the first digital imaging companies. Um, if you ever heard of eTrauma before.

Yes, I've heard of it.

Okay. Why, why don't you tell that story, Mark? 'Cause a lot of people don't know.

You guys were, well, we were early. We, um, We met a young orthopedic surgeon in 1998 who had rented his own table at a meeting in California by himself with no clue how to start a company or anything else. And he was talking about digitizing x-rays so that when you're on trauma call, you would see the x-ray before you go in and you knew whether to call the OR, operate tonight, operate in the morning. Put it off, you know, call the Synthes team, whatever you had to do. It was a really smart idea. So we took that idea and it was the easiest thing to sell in my life, but it was the quickest way to bankruptcy because all the surgeons wanted it, but none of the hospitals would buy it because we learned the hard way that radiology in the hospital wants to see the image first 'cause that's how they make money. So they wouldn't let the hospital buy it. So we pivoted to doing the PACS system in medical practices, which nobody had done. And that's how we brought eTrauma to the medical practices and not the hospitals. We let radiology keep the hospitals and we took it to the practices and our, basic line was, you know, we speak orthopedics, not radiology. You should do business with us. It worked out real well. So, but the, the, uh, what you've done here with 2D to 3D, that's a game changer. It sounds like you are betting the farm on the patient-specific medical device, right?

Yes. Yeah, we, we are So we have been kind of pioneers in that field, especially in oncology and histology. So that's why we're going to leverage our experience and bringing that tool to existing applications will be a game changer. And as I said, the beauty of that is we have FastNCA clearances that will not suffer or that will not be impacted by this converter, as we are going to have a dedicated FastNCA conversion tool that will then generate CG scanner that will be input to existing Fast K currencies. So that's the smart way of approaching the business, in my opinion.

Wow. But you got the right guys on this channel with a couple of these guys really know a lot about that. Thanks for your time today.

I'm trying to wrap my head around, is this a product company or a software company or both? I'm trying to figure it out.

We are a bit schizophrenic as well. So one day we are a bit of a product, and someday we are a bit of a software company. We have— we are a mix of— and nowadays no one can speak about hardware without having a very strong software behind the scenes. So anyway, we are running everything behind the scenes, but right now we are bringing all of those tools to the orthopedic world, and we are making them available for customers, customers of us. And the AI tools are definitely very, very cool toys. So right now everything is focused on AI internally because it improves efficiency, so it improves the margins. So it's— we are going faster, we're going further. It's online, so it runs over the cloud. So that's definitely the new business model, the new trend. So the piece of plastics we are producing are definitely helpful to execute what have— what has been planned. And I would say it's mandatory to have such a— to get such an assistance. But the plastics is not our life. We are dealing with software to produce 3D models, plannings, virtual data to power an intraoperative assistance. That's the way we— I would define 3Sight the best. Wow.

Wow. It's— yeah, it seems like you're gonna have to pick one of those channels to figure out— you'll figure out where the value is. Um, seems like you could scale way faster if you were licensing or selling software tools as opposed to making products? Because, you know, printing a cut guide and shipping it across the world can't be really efficient. I'm just thinking out loud.

It's terribly true, indeed. So that's why we focus on very high added value on oncology and osteotomy. And in terms of osteotomy, not necessarily the easiest ones. We deal with extreme osteotomies, multiplanar corrections, multiple angles, elongation. So the most extreme ones, I would say. So this is where the added value of having internal private label cutting guides brings revenue first. But second, on the arthroplasty side of things, definitely I agree. That's why I would define Freeside as being a producer for virtual planning to be connected to an intraoperative assistance and connection to robotics, augmented reality, or navigation system is definitely our future. We don't want to sell pieces of plastics all over the place. It's time-consuming. It's full of powder. So we don't like it that much, but we do it for customers who need such cutting guides.

Interesting. So who's your competition? What's the, what's the standard of care that you're trying to change in the US? You know, and there's companies like Restored obviously that are selling implants in the US based on— there's a lot of patient-specific everything in the US.

But yeah, but the fun stuff, so Dan Libro, you mentioned Dan Libro, he's still in front of the screen here. He's a good friend of mine working for Restor3D, and we are not competitors, but we are, uh, good friends, and we are collaborating on certain patients. Uh, there's a very nice synergy as we do not manufacture any implant, nothing that stays in the human body. That's a good point. We cannot be perceived as competitors, so we, we stay on the planning side of things, uh, potentially cutting guides, but that's the maximum we do. We, we never go to the implant side. So somehow the entire landscape could be one of our customers. So there's no barrier.

Got it. Are you familiar with TechFit?

Um, yes, in Florida. Yeah, I know, I know that name. So Colombian company, if I'm not wrong. They got acquired recently.

They were acquired by Acumed. Yeah, and Mauricio Toro was a BoneChat member. Yeah. And, but they, theirs was a little different because they allowed the surgeon to manipulate the model and basically order, order an implant. And they're mostly CMF, mostly CMF procedures. Anything you want to ask us, Laurent, is just advice in the US on about anything?

I don't want to— I won't pretend that I know everything about US. Of course, we're still learning a lot every, every day, but we have US, let's say, trained guys, and we are growing quite interestingly. But I would say that's the last slide I kept. We are building something significant, but we need money. So my only request here to the attendees would be if you have a very good address to provide me with, I would be more than happy to reach out. We are strongly encouraging partnerships. So we'll first do a quick turn with a quick round with our partners and customers, but we'll probably explore new opportunities, new ways.

Yeah, everybody's, everybody's raising money and it's always hard. It's like, well, uh, Stephen, I am putting something in the chat for you.

It is an investment platform our team built with the, the largest healthcare investor intelligence platform out there, as well as you could search for individual angel investors anywhere globally and get their contact info. I also put a code in there so you can get a month for free on it, and anyone else can in here can use it.

So there you go. Much appreciated.

Yep. It's got well over 1,000 firms, 60+ family offices.

You kind of—

you name it, it's in there. And then you can also run due diligence on all of these companies as well. Imagine it's like PitchBook but for healthcare.

So, yep.

Excellent. That's great, Stephen.

Thank you.

Great resource. So Mark Maloney is asking about your annual revenue. I think he's just trying to figure out how big you are right now, roughly.

But we are— so we are growing, but we are still kind of very small. We reached the $2 million bar and we envision $3 million next year.

Okay, that's 50% growth. That's, that's really healthy. What's the biggest, what's the biggest pushback in the US when you try to sell the planning?

Um, you mean, so again, schizophrenia, again, uh, arthroplasty product, you mean?

Well, let's talk about both.

Let's talk about arthroplasty and on, on arthroplasty, uh, so I would say that the place is a bit throughout it. So major companies already have in-house platform or acquired a platform to make it in-house, to internalize it. So that's our first pushback. But the 2D to 3D conversion could unlock new doors, so we'll knock at those doors once again. So that, that's one. On the, the private label products, um, The— we have very good tractions, and most of the time when we get back from scientific meeting, we have plenty of potential leads. But it's the being a bit far away coming from Europe. I would say that the distance makes a difference. So that's one. And we have to establish a local distribution network with local sales rep. That's not an easy task. And having those sales rep highly motivated selling your product is another hurdle, I would say. So to me, the, the distribution network is definitely the, the most significant hurdle to—

Yeah, sales, building sales is really hard, and there's so much noise. It's hard to find the right 1099s that have access to the right customers. Yeah, I mean, I think I'm with Mark. I think the 2-day to 3-day thing is just a game changer. Just, just trying to find the word. I mean, geez, I didn't think about it. Preoperatively, then in the middle of surgery, you can get a 3D view. That's incredible. And then post-op, if there's no hardware, there's no metal hardware in place, I mean, that's huge.

It's just that if some of these big companies like, like Stryker, like AlphaTech, who has EOS, You know, if they're already in the hospital and they're already offering something like that, that will be hard then to, to get a foothold in as a smaller company.

Good point. Contracts— you may run into contracts real fast.

But what about surgery centers? That might be more accessible and I don't know. Yeah, I kind of see this as a great SaaS model, and to make it universally accessible, just bring it down, bring it into the trenches.

That's a really good point. Have you— are you selling into, uh, surgery centers and hospitals in the US? And is there— is that surgery center easier sell?

Um, not yet. We are reaching local sales through surgeons, as we are delegating our services and products to surgeons. But, uh, indeed, that's a new opportunity we could explore. Yeah, thank you for that.

Explain, Mark, explain to Laurent why the surgery centers are just— why this product service may be easier sell than a hospital.

Well, that's where most of the, uh, elective trauma goes. That's where a lot of the elective surgery goes. That's the trend for total joints and spine to go in as much as possible to the surgery center, which are often either physician-owned or private equity-owned. They're also more accessible from a vendor standpoint because, at least yet so far, you don't have to climb the, the Matterhorn of value analysis committees to get your product in, which is a serious impediment and a big barrier to innovation. And I just, as we're sitting here talking, I'm thinking about a lot of times when people think about, you know, imaging and patient-specific medical devices, they, you know, they think about hips and knees I think about spine, and they kind of ignore the, the small bone guys. But a mistake on a small bone is actually bigger than a mistake on a big bone, you know. I mean, ask— Tiger knows the tree story so well, and a lot of the bunion stories. I mean, that was a high-volume procedure that until they could do it accurately, orthopedic surgeons didn't even touch it. They left it to the podiatrist. But now it's very common for orthopedic foot and ankle surgeons to do that because they can do a good job. And I think the same thing, I think somebody mentioned shoulder earlier, that might have been Kyle. But anyway, I'm just, uh, if you could democratize the access to what you do, uh, and bring it down to a very accessible level in the surgery centers, That'd be kind of— you could beat the cartel at their own game, like the people that Andrew was talking about. It sounds like it'd be a— it'd be fun, that's for sure. If it's not fun, why do it?

I love that. I think surgery centers is the way to go. Yeah, give it a— give— I don't know, give every surgery center, orthopedic surgery center in the US, a You know, free software, you know, heck, 1% of them use it and you're off to the races.

Yeah.

They don't have the compliance stuff. They don't have all, you know, they just don't have all the baggage that the hospitals have. VAC committees. You can, yeah, they, they're more open to, they're open to any technology that helps them make more money and use their operating rooms more efficiently and bring in more patients. Yeah, I, I agree. Any, any, yeah. Any other comments, questions? Eric's been quiet in a distance.

Yep.

Nothing for me.

Okay. Can this work in spine? Have you done any spine?

We tried, we did a bit of spine, but it was a bit exotic applications for tumor resection or metastasis, but nothing really serious. I mean, not driving any pedicle screw, but we received a few requests in the past, but we didn't necessarily want to address the spine market at that time. But, uh, that's definitely a new, a new area that we'll explore in the future.

Yeah, interesting. I'm trying to think of procedures in surgery centers where they don't want to send the patient out for a CT scan or MRI, um, but if they could do plain films to 3D, they would— it basically gives them a new patient.

Yeah, most of the stuff that I've seen are for hip, in particular for hip. But I think, yeah, it can theoretically can be used for spine, right? And the fact that it's— you can do it weight-bearing. I think you mentioned that, Laurent, like that, that is a big thing that is starting to take off now is to do at least pre-op weight-bearing to see the actual skeletal angles and everything. That, that, that's huge. You can't really do that with CT.

Yeah, you can make a list, Tiger, of the bony procedures they do in surgery centers, you know, like distal radius, elbow, Uh, foot and ankle, obviously. Probably most of the knee and hip would be soft tissue, it seemed.

Distal femoral osteotomies, high tibial, you know, HTOs.

HTOs, yeah.

Yeah. I've had one. They're not fun. But, um, the DFO, but there's a lot of good, like, I'm thinking you got a base in Colorado, Steamboat Springs. They do a decent amount of them out there using the BodyCat system. And then the Vail Health System has their own. Um, they do a good amount of them there. As well. They use Bodycad, and that's where I had mine done. But they have a surgery center, the 3P. More willing to look into that.

Well, the Bodycad is a patient-specific item that made the HTO easy, and that was the key. The key to the HTO is it's a hard procedure to do unless you have a 3-dimensional genius doctor, which nobody really does have that. Oh well, very few. There are a few, but the With the pre-op CT and everything, they made that sort of more doable. Mm-hmm. If you could make the, if you could make the HTO or the distal femoral osteotomy easy to do, but for a community orthopedic surgeon. Oh yeah. It would revolutionize the, the market because that's a great procedure for many, many patients prior to a total knee replacement. It's just, it's really hard to do. It's hard to do.

Yeah, especially that it's very popular in Europe, and I think that the popularity is growing in the US as well. That's my perception of things, at least. And indeed, we are doing HTOs, DFOs, driving plates and screws. We are, as I said, we are associated with a partner that I cannot mention here, but we are going to drive, to guide their plates and screws, and that's our The osteotomy thing is in front of us. So that's, that's a market we're going to take up.

Yeah. Yeah. There's also, also a trend towards joint preservation in the US. People do not, people do not want a total knee or total hip. They want to, and, and so that's, that's forward momentum for osteotomies.

Well, Tiger, you probably can clarify that and say it's, that's a not yet circumstance. They know they're gonna need it at some point. So you're talking about an earlier intervention option, and there's, there's a lot, there's a lot of opportunity there for sure.

Interesting. Well, this has been great, Lauren. Thanks for sharing. You really got something there. You should heed Andrea's notes, probably, and you should, when you're raising money, they're probably gonna want to see an FTO.

Write-up.

So you probably need to do the FTO exercise. Most investors are going to want to know your IP is clear in the US. Well, thanks. This has been great. Thanks for coming on, and we will be watching. If you need any help, you know, you— I can give you all their contacts if you want to talk to anybody. Later on. So thanks, Kara, and good luck.

Good luck.

Good luck.

Thanks.