Perimeter Medical Imaging AI, Inc. (PINK) Earnings Call Transcript & Summary

Greetings, and welcome to the Perimeter Medical Imaging AI Investor Presentation. [Operator Instructions] As a reminder, this conference is being recorded. At this time, I would like to hand the call over to Glen Akselrod of Bristol Investor Relations. Thank you. You may begin.

The purpose of today's presentation is to give our audience a better understanding of the business through a presentation and then questions with management. The presentation is going to be led by Adrian Mendes, CEO. He's also joined by Andrew Berkeley, Chief CIO; and Sara Brien, CFO. You'll see the presentation on the webcast. If you'd like a copy of this deck, simply email me at [email protected]. I'll be happy to send you one. When we do break for questions at the end of management's presentation, we encourage those questions. As a reminder, we're only taking questions through the web portal. If you're listening over the telephone, please access the web link that I sent earlier to ask a question. Remember, you could submit a question using the portal within the webinar at any time. I'll ask the question on the air for everyone to hear and Adrian, Andrew, Sara will answer. I'm not going to reference any names but simply read the questions back. As we have a fairly large audience today, if I can't get to your question online and has not yet been addressed during the call it can be, I'll come back to you by e-mail. I'm not going to read the forward-looking statements, but I do state that they apply, and I reference them on Page 2 of this PowerPoint. With that said, thanks once again for joining us for another that's fairly informal, and we do encourage questions to help you better understand the business and its growth path. And now I'll turn the call over to Adrian to start his part of the discussion and presentation.

Thanks, Glen, and thanks to everyone who joined. So it's great to be here. I think Andrew, Sara and I are happy to walk through the business with you. Both Sara and I are new to the organization as of earlier this year, about 6 months now. Andrew is one of the co-founders of Perimeter's that he's been here since the beginning. So we're looking forward to kind of sharing with you where we're at, what our story is all about. And then any questions you might have, of course, you've got access to all of us after we go through a few slides here. Okay. So let's hop into it. There's our forward-looking statements. Okay. So Perimeter, what's our -- why are we around? What are we up to, right? What's our purpose? So we envision a world where patients no longer experience the emotional and physical trauma. Have been called back for a second surgery due to cancer left behind, okay? This is a big problem. And what we're trying to do is we're trying to help improve this, helps solve this problem. That's what we're -- sort of that's our reason of being. Let's talk about cancer for a few moments. So I don't think that we necessarily need to spend a whole lot of time talking about what cancer pad. But cancer is -- it's a broad region, right? 1 and 4 or 4 and 10 people have a lifetime risk of it. And it's a tough one because everyone knows if you get cancer, one of the modes of treatment is to go to surgery to excise the tumor. The problem is that frequently, that surgery isn't fully successful on the push shot and then requires what they call a recession or a reoperation to get the rest of the cancer. So we've got some numbers down here at the bottom, which are rather large when you think about it. So almost 23% of all lumpectomies, breast cancer lumpectomies, require a reoperation. 11.5% of thyroid cancer, 21% of prostate cancers is a big problem in this very broad region. So I'm going to take a few slides to kind of delve into why this from exists and then talk about how the perimeter solution helps to address this. So the reason why these recession rates are so high really comes down to what they call margins, okay? So what the goal of the surgeon is when they go and try to address a cancer situation is to get the entire cancer out of the body and one of the tests that use for that is what they call negative or clean margin. And sort of picture that like cancerous tumor fully encompassed with healthy cells, at least some amount of healthy tissue around this. And if you can do that, then you can be fairly confident that what you've done is you've removed that entire tumor from the patient's body, and that's good. However, so that's good, and that's called a clean margin or a negative margin just to get some terminology onto the table. However, if you don't quite achieve that and you can see in that little cartoon in the spread in the center of this slide, you have a situation which is call a positive margin. And a positive margin means that the tumor is exposed at the edge of the excised tissue, the tissue that's removed from the body. And what that means is that more than likely you've left some cancer cells behind in the patients. And of course, that's a bad thing. That's what the center cells are there, they can continue to grow and spread and you really haven't taken all that cancer out. And that's the root cause of many of these recessions not having the required positive margin when the tumor is excised, okay? So that's the goal, right, getting a negative clean margin. The problem is it's very hard to tell from visual inspection, right? So when the surgeon takes it out, it's very, very challenging to take a look at that tissue and determine whether your margins are clean or not. So the surgeon does the best job that they can with preoperative imaging and other techniques. But what ends up happening is really when ends up being the standard of care if they send that specimen down to pathology. Pathology takes a few days, up to a week, sometimes more than a week to analyze it and then come back to the surgeon and tell them either your margins were clean or you're in negative or you had a positive margin and you have to call that patient back. Tough situation, right? Because that patient is now at home, maybe back in their job, back in their life, thinking that they've gone through the surgery, and they're in good shape and then they get it dried phone call. So that's the current standard of care. What our solution does, which I'll talk about in a few moments here, is it allows the surgeon to come -- to get the information in real-time` while the patient is still in the operating room within about 15 minutes to scan the specimen and to give the information of a surgeon for them to make clinical decisions before they send the patient home to help reduce the chances that there's cancer left behind and, therefore, reduce the chances that they would have to come back and reoperation. Okay. So there's a problem, right? The problem is its recession epidemic. For breast cancer, 1 in 4 patients will need a recession, which is huge. That's a huge defector, what I call defectors. And the reason this calm still exists today after all these years is there really is a technology gap. And so Perimeter is a technology company. We've got a couple of different modes of technology we bring to the party. And these things together are how we're addressing this technology gap. What we've got here is a picture of the device. That's the picture on the left-hand side available now. We've got 2 products listed here. On the left side is what we call our S-Series device on the right side is what we call our B-Series device. The S-Series is today, FDA cleared. It's an operating room, surgeons are using it. The B-series is not FDA-cleared yet. It's in clinical trials right now. And so that's our next-generation device. The base technology that both these devices use is something called optical coherence tomography. It's an imaging technology. If you look at the device picture on the left side, this what we call OCT, optical coherence tomography. The OCT module sits in the cabinet looks upwards and a little circular thing, you see sitting on the -- starting on the desktop is where the tissue goes and it images the tissue. And it looks for those margins that we talked about a couple of slides ago. And it provides that insight to the surgeon in real time without having to wait for pathology to get to the analysis a week later. So in real time, the search and look at that, what's key to this is really the resolution. So OCT technology has much better resolution, much higher resolution than x-ray and MRI 10x or 100x. There's a little diagram here, which depending on the size of your computer screen, as you kind of see that the image on the top, which is an OCT image, you can actually see much more detail than the ultrasound image, which is below. And that difference allows a surgeon to distinguish in different types of tissue. And if they can tell the difference in different types of tissue, they can tell the difference between a cancerous tissue and healthy tissue. And then you can just imagine how that lends itself to identify margins, okay? So that's the base OCT technology. That's what we have in market today. What we have in trial, in clinical trial, the B-Series takes that and as an AI layer on top of it. And so I'll talk about that a little bit more in a moment. Before we go there, I want to explain -- so what's very important for new technologies and really any workflow, but in the searches work flow in this situation is does it fit in? Or is this going to require the surgeon to learn an entirely new way of doing business. And the good thing is that it doesn't require them to learn an entirely new way of doing their procedures. So these 5 steps are kind of the workflow. The first 2 steps are standard today. They do this with or without our technology. So step 1, promote the tissue, step 2, exert to make sure that they actually got the biopsy click or the seeds that are the localization seeds out of the patients, so there's actually that's done. Step 3 is where we come in. So take that specimen, put it onto the device in images -- based on that information, the surgeon can make a decision of whether they believe that they don't have the margins they require. And if they don't, because the patients are there, they can take what's called shave. They can take a little bit more tissue in the area that the image is suggesting that the margins might not be where they need to be. Take that shave, scan that again. Hopefully, it's good. If not, take another shave, but if it is good, then you can close with a higher degree of confidence that, that patient is now cancer-free. And then, of course, so all of that stuff happens in the operating room. And then, of course, afterwards, postoperatively, step 5, where all of that you should get sent down to pathology where the exam is it to confirm what the margins are, how they should be. So this fits right into the service workflow Okay. Great. How are the surgeons actually reacting to it? I'm not going to read these quotes, I won't pause for one you can read on me talking, but you issued just 2 of many surgeons that we've got currently using the S-Series and having positive results for me. So we're getting positive feedback from our surgeons, which is good news, and this is all on the current S-Series device. This is the one that does not have the AI built into TS. Moving over to that, this is a slide that describes the B-Series, which is the thing in trial right now with AI. It takes the same basic hardware, adds an AI software labor on top of that, which what it's really doing is looking at the images that are created and then indicating sort of highlighting those areas, which the AI is picking up as being suspicious based off of all of the training data we've given the model. And it's allowing the surgeon then to quickly go to that place that the algorithms are saying, there's a suspicious thing here in this picture and inspect that. So rather than on the current as of surging us to kind of look through everything, here it with the AI, it really has a surgeon to spend time looking at the things that are most suspicious. So it improves the workflow, it allows us to expand the user base because you don't have to be -- it's less effortful as a surgeon to use it if you've got an AI system with you. And it speeds up the time in the operating room. So this is what's going through FDA right now and -- or through clinical trial right now. And here are the sites that the trial is being run at. So we're really working very closely with some of the top cancer centers in the country, Mayo, MD Anderson, Moffitt, kind of a good swath of hospitals across the country that are really specialized in cancer research and cancer care. Okay. So, that's the quick summary on the problem statement, re-excisions the [indiscernible], margins, the solution, real-time imaging of the tissue to understand the margins and then some validation of how things go in the field and a little bit of a road map for how we're adding AI on top of the products and running that through a clinical trial right now. Okay. Flipping over to the business strategy. How do we create value with this product? Like kind of intuitive, but let's talk about it a little bit more concretely. There's 4 constituents, I think that we create value for, right? The top 2 are kind of obvious, a patient that doesn't need to come back into the operating room and kind of the first time they have the surgery, their success immense value from just an emotional standpoint, from a time of work standpoint from a -- from all the patient care type standpoint, from pain from risk of complications for having to go through a second surgery, just massive benefit decrease to the patient from only having go through one surgery versus multiple. Benefit accrues also to the surgeon. They can provide better patient care. It's less dramatic for that surgeon to have to call up the patient and tell them. I promised you I was going to do this for you, and I did it. And unfortunately, I have to do it again, right? That's a horrible phone call that the surgeons have to make. So there's value that crews there. Then the bottom 2 is the facility and the payer. So the facility, obviously, would like a reputation of being able to do have a high success rate on the first time is cancer surgery. There's all kinds of market benefits from that. There's OR efficiency benefits from that. So a lot of value accrues to the facility. And then the payer also because if you have to bring a patient in for a re-excision, that costs money. And it costs at $16,000 for every re-excision you need to bring in. And if you can eliminate that $16,000, that's just a straight-up savings for whoever the payer is. So there's these 4 constituents. They all oversee benefit from being able to reduce re-excision. And let me go from there. Let me like segue from there into the actual economics behind this. Okay. So this graph is focused on breast cancer lumpectomy, specifically, and U.S. only, okay? So these numbers you see here are in that rate. So there's about 230-ish lumpectomies a year in the U.S. That's the pie. That's the full pie that you see on the left in little box there. And there's data that shows that there's about a 23% re-efficient rate, right? So 23% of [indiscernible] is 53,000 surgeries that happen that don't need to happen, but they have to happen because the margins weren't clean when that first surgery was completed. And if you take those 53,000 reoperations and you multiply it by the $16,000 cost per reoperation, you get that 857. So $857 million per year spent just on fixing something that unfortunately wasn't done correct. Correct for the first time, $850 million, okay? So that's sort of the total -- what I call waste in the system around this particular area in the U.S. alone every year, just for breast cancer, okay? It doesn't talk about any other case, just breast. What we're target, okay, now move your eyeballs over to the right side. What we're targeting is -- and we think we can achieve this, and we've got some reasons to believe from some things that we've seen that we can achieve a sub 5%, 3% to very reasonable re-excision rate based off of using our technology, surgeons can get to that, we believe. With that and then you just move the math on that, right? With that, we can move the re-excision rate if we were everywhere and certainly just for everything, right? Move it from 52,000 re-excisions a year down to a little bit less than 7,000. And so if you just do the math on that, the difference between that almost $750 million in health care savings if we could take our re-excision rate from 23% down to 3%, which is a very reasonable goal with the tech from what we've seen out there, sort of as we've been using this. So massive amounts of saves from health care. And if I flip to the next slide, it kind of shows -- it kind of puts that number then in context of the whole problem. So one quick thing on the right side, the U.S. only represents 10% of global cancer diagnoses and only 13% of global breast cancer, right? So there's a massive market beyond the U.S., and of course, it won't apply to. The dollar figures will change because U.S. health care is more expensive, but at only 13% and 10%, there's great areas to expand outside the U.S. And then we're just talking about breast right now, right? But as you can see in this pie chart, so breast cancer is that dark blue pie slice sort of near the top. The one right beside is colorectal, in gray, the pink sort of color is lung, the I guess color is prostate. All of those are approximately the same size, head and neck is below that. So about the same amount of dollars is spent in those other cancer types as in breast. So there's a way for us to take this tech and march it through the other types of tissues to increase our SAM and our TAM by expanding beyond just breast. So it's a massive market, and we're just starting in it with breast. Okay. This is a recognized problem. The fact that there's a very, very high re-excision rate cancer. The fact that this isn't just breast cancer but multiple other cancer types or tissue types is a problem that has been recognized not just with us, but most recently, for the U.S. government, okay? So Biden has a cancer moonshot fund that they funded through that they're providing and calling this is money that's already been appropriated by Congress $330 million is a close here from Joe Biden. $330 million focused or sort of like sort of put towards ARPA-H. ARPA-H is a department within the NIH. And they've got a program that's focused on what they call Precision Surgical Interventions or PSIs funds there. What they're doing is they are pushing this money to industry to try to help solve the problem that I just said in the last 10 minutes or 15 minutes' topic bets. They're running a process right now. We're in the middle of the process. We requested abstract in August. We submitted an abstract. We got approved for a -- to go to the next level of so abstract was accepted for a full proposal. We'll be submitting our full proposal, November 16 is when that's due, and we expect to hear back from that in January, February Q1 of next year sometime. There's significant money behind this. We've been working on this for 10 years. We've got a great solution that actually addresses this problem directly. So we're pretty excited about this. We're excited by the fact that there is a widespread acknowledgment of this significant problem. We're excited about the fact that we have technology that addresses this firsthand, and we're excited by the fact that there is actual funding coming in, non-dilutive funding coming in to help solve this. So we'll keep everyone updated as we progress through this process. But this is good news for us. Moving on to how we monetize. So what we do right now is we have a managed solution type of approach where we place a device into a facility for free. And then we charge for the consumables. That middle ground thing that you see in the picture there, that's consumable. That's where the tissue goes into and then that whole thing gets put on top of the machine. And of course, you're going to want to use one of those for patients, so it gets turn afterwards. So the model we're trying to build the company around is a recurring revenue model where we place the machine and then every procedure, we make some money off sanity procedure. The device is kind of a platform device, right, like which means although we've been focused on breast right now, it does expand beyond that with -- it should be the same hardware, same thing sitting in the operating room can expand to other tissue types as we continue to step through that process. Okay. Our go-to-market strategy right now, we're feeding the market with the S-Series. That's the product that is FDA-cleared right now. That's out there in the world. That's what we're doing right now. In parallel, we've got our AI product in clinical trials. Once that goes through, we look to step 3, where we actually launch the B-Series, the product with AI. So we get to launch in that broadens our user base in the U.S. And then beyond that, we expand the market, right, in 2 directions. One is expand by indication, which means start to target other cancer types. And then we can also expand by geographies, so move outside the U.S. and tackle the rest of the markets. So that's sort of our go-to-market over the next few years. We're in pretty good shape right now from a cash-on-the-balance-sheet standpoint. We've got over $20 million in the bank right now. That will take us into 2025. So we're in a pretty secure shape. That will take us through this trial and everything. So I think we've got the cash right now to be able to execute this strategy. So just to wrap up, not only are we envisioning a world where there are no re-excisions and all of the trauma is eliminated from that aspect of the whole or zero cancer, not just what we're envisioning what creating it. We're seeing it happening based upon our technology. And that's our story. And so with that, I think we head over to Q&A, and I'll hand it over to Glen.

[Operator Instructions] So first question for you, Adrian, is what is the learning curve for a surgeon to implement the OCT imaging? I find specialized doctors to be some of the hardest humans to convince to learn something new. How are you exposing new surgeons or your devices used as teaching hospitals, this seems the key to becoming an industry standard.

Yes. Okay. So what we've -- so yes, it's a great question. Surgeons are some of the hardest humans to change their patterns of behavior. And so what we found -- okay, so there's 2 things. One is you have to find the right surgeons who are willing to use -- like bring new tools into their practice. You have to target them. The stage where a company, that's super key from a sales process standpoint, targeting the right surgeons. Once you identify and target those surgeons, what we found is it takes 2 -- there's 2 elements of the training. One is the ability to read the images. So you saw the screen with the OCT image. So we've got a whole training curriculum that allows the surgeon to look at many of those images and then start to train their eyes to the pattern recognition in their brains to identify like what the different tissue types are. So we have a curriculum around that. In addition to that, we find that it takes about 5, maybe 10 cases where we have a clinician in the operator with the surgeon. And around that time around 10 is when they get comfortable. And so we're able to -- they're able to at that point in time after about 5 or 10 cases able to then be on their own. We have been successful in this manner with surgeons. Our RCT, the clinical trial we're going through right now. That's been very successful. In those trial cases, we're not allowed to be in that room, and this is the exact position that we've used. So I think you're right in what -- the premise of the question is absolutely correct. The key is find the right surgeons and then having a train or curriculum, which we do have and we've exercised.

Next question is what are the FDA requirements for sensitivity, specificity of the AI?

Okay. I'm going to hand that question to Andrew. Andrew, are you able to get into that in a little bit more detail.

Sure. If we haven't outlined a specific sensitivity and specificity, but in our clinical trial, the primary endpoints and the secondary endpoints leads to demonstrating that using the technology can reduce positive margins and then subsequently reoperation versus what the standard of care is today. So we -- I can tell you that the off-line level that our AI works in the mid-90s for sensitivity and specificity at the patch level. That doesn't necessarily translate into the operating room. But the FDA are not asking us for a specific sensitivity at specific numbers. They are asking us to demonstrate that the technology here and improvements under standard of care and then it also demonstrates safety. So safety and effect of the work we're looking at, not specifically sensitivity and specificity.

We've got some other questions in the queue that relate to this topic. So I'll just ask now, I think part of it, you've added it already. But is there a timeline around your FDA process?

So yes, the FDA process, it doesn't have a time associated with it. It has a number of patients we need to enroll into the study. We're expecting that to wrap up towards the end of 2024. We have the opportunity to have what they call an interim analysis; we get halfway through to take a look at the results. And if we've met the primary endpoints, then potentially stop the trial early. And we expect to hit that in the second quarter of next year sometime. So those are sort of our 2 dates that we're moving towards for the trial.

I'm going to ask another question and then going to come back to the FDA. What might be the next logical cancer indication for expansion? And how long might it take to be where you're in breast cancer today?

Andrew, I'll give that to you.

Yes, I think there's a couple where we also getting a device into the hands of a thoracic surgeon to start looking at retroaction for lung cancer. Also, we have a device right now at New York Presbyterian where we're looking at head and neck cancers. We've also done some work about Colombia and in the past. And then we're also looking at liver and pancreas. I think over the next 6 months, we will be able to identify which one of those cancer types is the next, the most appropriate one for us to go after. In terms of how do we get from there to press the level where we are pressed -- when you think about the cancer types, a lot of those will be -- I would reckon they be kind of easier to image on identifying where causes cancer is pretty difficult because of the kind of heterogeneity of the tissue and how breast cancer can grow. When you have a swollen organ under the lesion in there, I get the tumor in there, you have very well-defined structures. So I'm assuming it will be a shorter path to get our second tissue types to having an indication than what we've experienced with press. And also, we will be able to use all of our AI models through using transfer learning. All of the technology behind the AI models will be the same. We will just be putting in new data volumes for training and applying that. So we'll be able to get the AI component out of it much faster as well.

And I'm going to come back to the FDA portion and it is are you going to have to take both the FDA and the B-Series through another FDA process for all other indications?

Everyone indication, yes, we will have to go back to the FDA approved that we can be cost and festive use of the technology on all patient groups. However, we will be able to use the existing technology in the S-Series and then the B-Series out the size of predicate. So the level of clinical validation that, i.e., the clinical trial, what we would have to do would probably have to be to be smaller than what the initial press indications or the original price regulatory approach to.

Switching to the technology. A couple of questions here that I'll combine into one. What is your IP around the technology? And then the second part is how defensible is that IP?

So I'll start and then Andrew can go to more detail. We've got 6 patents right now. and a couple more in the pipeline, mostly covering the OCT portion of the technology in various aspects. And then -- okay, so that's part of the IP that was done in the company. The other part of the IP that we have in the company is on the AI side of the house. And that's, of course, you don't pass it end up being trade secrets. And so that sort of describes the IP portfolio. Andrew, do you want to go to the next level of detail on any of that?

Yes. So technology we use is OCT. OCT has been used in ophthalmology for the last 20 years. It's about $30 million something over year with OCT. But the field of imaging with OCT is generally very small on the order of 1 centimeter by centimeter, like the back of your eye or for cardiovascular imaging planning inside investment. What we've done is we take an OCT we've applied some translation to the imaging probe, so that allows us to image much larger complex tissue investment and a really fast amount of time. So from an imaging perspective, we are the creators of lives [indiscernible] OCT. We can stand up to a 10 by 10 centimeter area, which is 100x bigger than the footprint of traditional OCT. We have IP around all of the stitching algorithms that take all of the images on together to create a new form airport. We also have -- and its quite significant IP around tissue handling. When you're imaging a really high resolution, you need to have a very steady still flat surface for the tissue settlement. So we created a tissue management system that whole assessment using by very steady, and allows us to do rapid scanning, but that's also part of our IP. We've also looked at combining OCT with specimen next way, which part the workflow kind of like protection against some of the special x-ray companies. So we have IP around combining those. And then we also have IP in the AI space for the work that we're doing with our cancer suspicious feature finding technology, which is not specific to breast that IP will be related as regard into other indications.

Next question. Will the launch of AI technology require any changes to existing, I guess, S-Series hardware? Or can existing machines be given a software update?

Yes. The short answer is -- we're about to say the same thing, is that the hardware is the same. We will have to do -- it's mostly software. So we'll have to bring the machines back because they are going to be different FDA part numbers, right? You have to distinguish them very clearly. So we do a software upgrade, but all the hardware is the same. The supply chain is the same, the manufacturing of the hardware is the same. So we get to -- it's a big leverage that we can to use across both the S-Series and the B-Series.

On Slide 7, for the OCT view, what is cancer and what is non-cancer, and how easy is it for surgeons to differentiate?

On Slide 7. This slide?

I think that's Slide 5.

Yes, the slide, what is cancer and what is non-cancer?

Yes. This is actually an image comparing resolution between TK and ultrasound. This is a cross-section of a human finger now. You could see on the OCT on the very left, you could see the top of the mail at the bottom and then the A route. And then you can see on the right side, that's the layers of skin and tiny growth vessels. And when you look at Ultrasound, you don't see any of that architecture. You can see deeper, but you can see fallout on the bottom you can one microscopic features that you can see in OTC. So when it comes to visualizing cancer, you don't actually have an image in here, but...

We do answer. If we go over here I think this is not, right?

Yes, we actually have a very early stage for breast example, whether it's to help things by the breast silk. Some normal DUC will not be expanded. They would to be very flat and singer, we can see that to see the sound of expanded I mean even a tiny little micro classification inside of this. And when you're panning through the volumes, you could track this and you can see how has it developed? How it evolves? So it's more or less an exercise in, first of all, identifying something that looks abnormal. And I'm just applying some pattern ratings techniques to confirm that exclusive areas. Contract looks very different to most normal tissue. If you want to organize the sales or a different change composes a density pack together where normal tissue features have much more of a structure to them. And that's what helps us differentiate to one of the others.

You've talked about surgeons, question is, is the radiology department in the hospital involved in reading any of these tests?

The entrepreneur is all. Radiology department is not involved in the surgery. Sometimes there's a pathologist who will help the surgeon post the vision. They will cut up the session and look at it. So we are actually looking at maybe in a particular sites using the pathologists on their skills to be image reader. But from a radiology perspective, the answer is no. But we say at we do have a device at Columbia University, New York firms of comment radiologists in the growth space who is looking at using our technology for almost immediate diagnosis at the biopsy stage. So what happens today is staring a biopsy tissue comes out. It goes to the lab. It takes a week, 2 weeks for typical diagnosis on whether or not the patient has cancer the basin that all is a very stressful time for 2 weeks. And what the premise of the technology would be is we would be able to have a really high complement of patients either those that have cancer or those had cancer, while they're getting their biopsy. And that would fast-track the treatment path for the patients who do have cancer and would cause the patients who have a nice-featured in a much more relaxed scenario also they got their final diagnosis.

Super actually sticking on the same theme, will the histopathologist see his or her workflow decreased?

Yes, that's a really interesting one. Throughout the whole development of this technology, we've worked more probably with pathology, especially on the AI side of things to -- because we're creating images almost practically what the pathologist sees on the microscope. Where we've seen the value for the pathologist is the information that's captured in the operating room to identify positive margins, kind of the export to pathologies. So I think began their processing of the tissue they can call up the image securing traction in the operating rules and almost used our technology like a Google Maps. They will be able to see where the highlighted areas of that are suspicious are, and then they will be able to see below the surface and actually have an indication of what they expect to see when they create the histology place. So right now, they are not brand-only sampling, but they make sample all over the tissue surface. With our technology and water reposition, it potentially would make them more accurate as to where they want to sample and then make them more efficient because they can spend less time processing tissue that has low disease or adoption areas and focus on areas our AI are really [indiscernible].

Next question. In your trial/testing, have you had any surgeons reject the use of this protocol? And if so, why? And further, were able to change their opinions.

We happened -- so in the trial, the answer is no because by virtue of the trial, they sign up to the protocol and then they perform according to the protocol. So there hasn't been that situation. Now I think the question -- underneath the question is, are there surgeons out there that don't want to take the time, don't want to take the efforts to bring this in? And the answer is yes, of course, there are those surgeons, I think maybe a previous question sort of alluded to this. And this -- and so I think there's a couple of reasons for that as we look through this. One is those surgeons that don't believe that their reaction rates are as bad perhaps as the average. And this is -- this is a little bit like the joke where everyone is a better-than-average driver. People often underestimate or overestimate I guess their performance. And so when we've worked with some surgeons to start to track or go back and take a look, frequently, they're a little bit surprised at how high the existing rates actually power. So that's one of the barriers we have to adoption that we have to work through. And then the other is identifying those surgeons that are actually open to learning new tools and learning new techniques. And at the stage of sort of go-to-market we're at right now. That's a very important filter for us, where we need to focus on the early adopters, the ones that are willing to take on the effort of learning it and then get the results from that. And there's a specific surgeon profile that fits that or teasers that sort of fit that profile. And so we focused our sales efforts on those types of surgeons.

I know you've covered this in your formal remarks, but maybe a little bit more time spent on this. Can you review your commercialization strategy in terms of cost and current placements, target placements and the sales force trainers in place and that you require?

Sure. So right now -- okay, so from a -- sort of how do we present this to our customers is the most common model we have right now is where we replace the unit for free. And it's like a managed service basically. And then we charge at the rate of $900 generally per consumable. So this is what we've been experimenting with so far. I think there are opportunities for us to evolve that as well. So not only have exactly what I just said, but ways some hospitals actually don't like placements where they don't actually buy the capital for different financial reasons. So I think there's opportunities to explore actually selling the product and perhaps having a lower consumable price, perhaps having a higher consumable price. And so this is an area for improvement for the company, just us experimenting with go-to-market in a way that is -- that allows us to really slow the boundaries of what's possible on the commercial side, okay? So that's #1. #2 is we do right now in the early stages of placing products with our customers. It does require clinician support. So our sales force looks like what we call sellers or MDMs, market development advantages, which are really the salespeople, the relationship sales people. And then the clinical support people, what we call clinical application specialists who are actually in the OR rooms with the surgeons, helping them to understand how to use a product, helping them with using the products and making them comfortable with it. One of the challenges we have had, that we're working through right now is the speed by which we get our surgeons to that truly being independent stage and feeling confidence. So I described earlier in the prepared remarks, the fact that for the clinical trial, it takes maybe 5 to 10, what we call roll in cases where our clinical specialists are in the operating room with the surgeon and by the end of that, they're comfortable. Depending the motivations for a clinical trial are sometimes different than the model-based commercial sites. And so what we're working through is how do we get surgeons to that independent stage quicker. The quicker we can get them there; the cost structure gets better because now we don't have to place a Perimeter employee in the operating room for as many procedures as we used to. We get to reuse that resource to get the next customer on board. And so this is a key part of how we build out our commercial team. Right now, probably for the next 6 months to 6 to 9 months, we're really focusing the commercial team on getting these high-value users, high what do you call, early adopter users on board advocates for us, those surgeons that can then talk to their peers and help promote our product with their peers and trying to make sure that every customer we get on board over the next little while is a highly leveraged customer that almost becomes sort of an extension of our sales force and can advocate for us, advocate for the product, advocate for the use of the product. And then also, frankly, give usability feedback back into our innovation and engineering team so that we can build out our pipeline of product development of how we make this product easier and easier as you sort of step through the entire market base.

Next question. Perhaps it's too cynical, but aren't doctors/facilities going to be hurt by the reduction of $750 million in revenues? Are all of their benefits reputational?

Yes, cynical, but there is some truth to it. On one hand, in myopic or a short-term phase, yes, if you take business away and that can hurt the top line of a customer -- sorry, at the hospital. What are -- but from our position, bringing efficiencies is always a good thing in the question is, how do you manifest that in the real world such that you get through the short-termism or the syncs the question just indicated. And there's a couple of ways to do that. One is to the payers, right? So working with those hospitals that are integrated health networks where the self-insured, right? So that's like an obvious benefit to those people if they don't have to pay an extra $16,000 for every re-excision. It hits the bottom line. And it doesn't affect their top line. So that's a big deal. The possibility of bringing this technology in and then advocating with patients -- with surgeons that there is this hospital that has this technology that reduces your risk of having to go through a re-excision and drives more business, so to speak, to that hospital. There's opportunities there that we are exploring that we continue to explore. So on one hand, it does, and it's something that we do have to work through and we are working through. But on the other hand, the benefit of the overall system is such that that's more of a tactical concern than a strategic concern. Strategically, we need to walk through it. And we are.

Next question is, what type of prostate cancer surgery can lead to a positive margins? Is it radical prostatectomy?

Andrew, do you -- can you take that question?

Sorry, can you just repeat that for me, please?

Yes. I guess I deleted the question, but just talk about what type of prostate cancer could potentially use your OCT technology.

Yes. So I think there are different stages of prostate cancer and dependent on the stage of cancer that can impact the cycle for how the surgery has performed. If the cancer extends past the capsule, then there's additional passionate taken around the neurovascular bundle one of the high extensivity of taking issue there. Generally, they don't go back for second surgeries when they have a positive margin, but they will administer additional radiation therapy on the deposit margin, which is actually quite a significant cost. So in the long term, what we would look at for prostate is can we establish for some prostate cancer surgeries if there was a positive margin and canisters and go back to actually take more tissue that's not going to be not on to retain.

Next question. Can you talk a little bit about the competitive landscape? Are there competing products? If not current? Are there any in development? And how far are they from the market?

Yes. There's no directly competing product like in the way that we're approaching the problem. There's other types of solutions that try to address it. And Andrew, yes, I think you're probably best positioned to be able to go into details about those.

The cost of one to in terms of where they are in their development would be -- there's a couple of fluorescence companies where they inject an agent into the patients, the number of arrays before the surgery [indiscernible] to the cancer cells. The remove the specimen -- if there is console the patients, then they illuminate the cavity with special light and it's supposed to allow some cancer cells to fluoresce so that you can see them. Has potential for large invasive tumors, but when there's microscopic disease, particularly when insight disease, where there's not a lot of sales and they're already inside of the DUC, for example, it's very hard to get a forest signal from that type of situation. And then what would happen is you need the turn of the sensitivity of the camera to see that very low level fluorescence signal and then you can have a lot of false positives. So fluorescence I think, will have a place in transfer surgery, but I don't think it's going to solve the problem in breast cancer, which is our primary approach. And there's also some technical issues. Like a lot of these surgeries are those as an outpatient. If you have to inject these agencies to a patient 8 hours before surgery. I mean they have to come in the day before, which is completely different in the U.S. It's a completely different payment system. There's different poles in inpatients, the posit patients. It's much more expensive surgery rerate within the hospital system as a post being in an outpatient and with very career. So there's many different factors the workflow and the actual clinical efficacy in -- but that would be even the kind of most competitive technology or sector.

And sticking, I guess, on this team, here's the question. I read that Microsoft is getting into AI for breast cancer. Is this competition or a potential partner?

This is a potential partner, I suppose to competition. So what Microsoft intent is twofold. One is they're trying to sell cloud services, right? So they would love to get everyone's data on to their cloud or so. That's number one. Number 2 is they're trying to sell AI services and obviously, as we discussed, AI has a strong place in cancer detection. So this is how they're looking at it. They're not getting into the imaging business. They're not getting into sort of like placing things in hospitals or being that close in the direct line with the surgeons. So really, what they're trying to provide is compute services, cloud services and AI services to the companies like us. So there's a strong potential for a partnership there which is helpful for a company like us.

Next question. Can you explain more about the ARPA-H. What was the $15 million to $20 million in the slide referring to?

Yes. So the $15 million to $20 million in the slide was referring to what -- it's an estimate right now of what we think could be the order of magnitude of what we might get awarded if we are successful in our proposal. So there's, I think, $240 million is what we've been indicating has been allocated to the PSI fund. And then we're estimating somewhere between 10 and 15 in proposals being accepted for funding. And so that's sort of like the estimate. If we have to like -- there's a $330 million total fund, but our estimate is something on the order of $15 million to $20 million might be what if we are successful, could be a piece of it.

Next question. What is your current availability of inventory of the S-Series units?

We have, I think, in our inventory something like 30 of maybe 30 devices in inventory, both at our -- both in our own facility and in our manufacturing facility. So we're fairly well established to be able to ramp the business without having to put more capital, more capital expenditures in to ramp. So I'm not concerned about supply chain issues at this point. I'm not concerned about needing to find financing to be able to buy a bunch of new machines to grow the business. So I think we're in pretty decent shape on that front.

Okay. Great. And I see we're coming to the top of the hour. So we've got quite a few more questions in the queue. We'll get back to those individuals that didn't get an answer by e-mail. I think we've covered most of it. But here's a good topic that we haven't covered yet. Can you clarify the cash runway? And do you have cash to end 2025 or over the beginning of 2025?

Yes, we've got cash into the beginning of 2025 with the 2020 a little bit over $20 million we have right now in the bank or we had it the last quarter grows in the bank. And looking at our projected cash and it's kind of in line with historicals. So yes, that takes us into the beginning of 2025, which takes us over the boundaries of a few key items for us, right? It takes us over the threshold of when the ARPA-H decision comes in. It takes us over the threshold of us clearing the OCT trial for the B-Series. So it gets us a little bit of a ways into the future.

Okay. Perfect. Thanks, Adrian. What I'm going to do is ask you now, if you have any closing remarks. And again, if your question has not been answered, I'll be sure to get back to you via e-mail since we're at the top of the hour. Some closing remarks from you and then we'll end the call.

Yes. No, thanks. First of all, I appreciate everyone's time coming and joining and listening. I think fundamentally, my overall summary view of the business is -- the technology is unique. The technology gets the job done that we need to get done. The small portion of the market that we're attacking right now has a huge economic sort of benefit and what we have between OCT and AI can address that. The challenges we have in front of us right now is learning how to -- not learning, but continuing to evolve how we bring this product to market, ramping up the sales process, getting more and more surgeon advocates talking about this product. and really ramping up the commercial side of the business. I think that's a big job for all of us here inside the company over the next year. And then on the trial side is making -- is ensuring that we have given everything we can to have a successful trial on the AI product. But that's the work we've got cut out for us over the next 12 months. The fundamentals of the company is strong. The cash is us been strong. The team is strong. So I'm encouraged. A lot of work to do, but I think we're heading in the right direction there.

Perfect. Thank you for that. Thank you, Adrian. Thank you, Sara. Thank you, Andrew. And thanks to our audience, and this concludes this presentation.

This does conclude today's teleconference. We appreciate your participation. You may now disconnect.