Sobotta, CEO of Perimeter Medical Imaging AI presenting. Take it away.
All right. Thank you. Thanks, Bloom Burton & Co., for having us here today. Excited to give you some updates on Perimeter. We're at a critical junction in the company. 2023 is gonna be a big year for us, so looking forward to sharing that story. Before I dive in, the obligatory safe harbor, I will be making forward-looking statements in today's presentation, and I direct you to our CDR filings for a complete discussion of our risks and facing the company.
Perimeter Medical Imaging AI, we're a medical technology company, and we're driven to transform cancer surgery. We're doing that with the intersection of ultra-high resolution and real-time advanced imaging tools to address unmet medical needs. The primary unmet medical need that we're addressing today is something called intraoperative margin assessment.
In the case of a solid tumor cancer surgery, surgical oncology, the standard of care for a patient is the surgeon would go in, cut out the tumor, and they're looking for a cancer-free barrier of tissue around that tumor to ensure that they did not cut through the tumor and leave it behind. Today, unfortunately, they have to wait two to up over a week, two days to up over a week for pathology to get a sample of that tumor, process the specimen, and put it under the microscope.
What Perimeter does is it uses optical coherence tomography. We think of it as ultrasound, but we shoot light waves into the tissue instead of sound waves to get a, you know, ultra-high resolution down to 5 micron view of that specimen so that the surgeon can see that same margin, that same cancer-free barrier of tissue at the point of care in real time and take real-time action on the patient.
We're a commercial stage company right now. We're commercializing our you know, flagship product called the Perimeter S-Series. What I say about this stage of our revenue is we're in kind of the train the trainer phase of the rollout.
We're looking for early adopters, reference sites that can become the future benchmark for us as we look to bring the next technology, which is our Perimeter B-Series with artificial intelligence, to market. We're really playing in about 20%-30% of the market today. The AI, which is coming, will turn that 20%-30% into 100% overnight.
Really excited about that technology. It's designated as a Breakthrough Device by the FDA. We're in clinical validation right now, enrolling in a randomized controlled trial up to over 300 patients, and we expect to have top-line data on that later this year. The people behind Perimeter, we're a management team that's all medical device background. I, the CEO, joined from Smith & Nephew, but spent most of my career in Stryker deploying over $4 billion in capital over the years.
It just so happened that one of the spaces I was extremely high on was something to solve intraoperative margin assessment. I'd done a lot of diligence on the space and really liked the way Perimeter was solving that. We think beyond the breast cancer market that we're talking about today, this is really a platform technology that can be applied to any surgical oncology application.
The same issue exists in, think, your hepatopancreatobiliary, so your pancreas cancer, partial nephrectomies with kidney, liver cancer, prostate, colorectal, all of the above, have the same, you know, margin assessment challenge. Lastly, you know, we're commercializing in this early stage, finishing this clinical validation study with a really strong balance sheet.
We ended the year, just reported our Q4 numbers with just over $28 million on the balance sheet with about a $12 million annual burn. Significant runway in front of us to get through these value creation milestones. To frame the problem a little bit better, breast cancer, you know, many people are impacted by that. Unfortunately, one in eight women are diagnosed in their lifetime. That's over 2.2 million annually, with a little over 300,000 in the US where we're cleared to commercialize today. And about two-thirds of those women elect to have a breast-conserving surgery or a lumpectomy.
Now the thing about lumpectomy is if you don't achieve that negative margin that I talked about earlier, the clinical data would show that risk of cancer coming back in the same breast doubles. Instead of accepting that risk, the surgeon would call the patient, say, "Sorry, I didn't get it all out. We're gonna have to have you come back for another operation."
Unfortunately, that happens 20-25% of the time on average. It's a glaring opportunity for technology to come in and improve that outcome. In addition to the risk of cancer coming back, when the patients come back for a second operation, there's nearly a 50% complication rate in that second operation.
You have things like increased pain management, poorer cosmetic outcomes, the never event of a surgical site infection, and on top of that, waiting for these negative margins delays any adjuvant therapy. The next step in the patient's journey is typically radiation, and a surgeon will not allow radiation to happen because they don't want to perform that second operation on a radiated tissue.
From a hard dollars and cents perspective, you've got an additional $16,000 in cost on average. You got one in four patients coming back for a $16,000 cost, averaging out to $4,000 per patient, of completely avoidable additional treatment costs.
That's really inefficiency in the system. What we do to help the surgeon be informed to make the right decision on those margins, like I said, the top half of this slide is what happens today without Perimeter's technology. Surgeon would go in, remove the tumor, they send that tumor to a pathology lab. Pathologist, everyone I talk to is at capacity.
Hospitals are saying they can't get enough pathologists, and they're waiting, you know, two days to up over a week to sample the specimen, process it, put it under H&E stain, and show that pink and purple slide you see there on the top half of our presentation. What we do is we produce that bottom half image in real time in the OR, so the surgeon can see it right away.
Instead of waiting a week to find out that you have negative margins and make the dreaded call to your patient and say, "Sorry, we operated and we didn't get it all," you can see it in real time and just go back and take extra tissue in the OR 'cause the patient's still open.
That's the same piece of tissue that you're looking at, one under the microscope, one with our OCT technology, and you can see very clearly that the surgeon would see that, you know, cancer-filled duct right there, and they'd say, "Oh, oh no, this is right at the perimeter of the specimen. I've gotta go back in and take more tissue," thus avoiding the $16,000 hard cost and all the complications that come with those second operations.
Our product offering here, we're a razor blade business model, so we place the capital unit that you see on the left-hand side of the slide here. That is where the image is generated, you know, and then what we would do is the surgeon would remove the tumor, put it in that second piece, our single-use consumable, and they'll drop that entire dish into our machine to generate the images.
Some of the complicated things and some of our, you know, great engineering that has gone into this is you're dealing with very, you know, shapeless, fatty pieces of tissue that are very difficult to immobilize and, as we've talked about, knowing the distance from the perimeter of the specimen is extremely important.
It's not like an X-ray machine where you can just pancake it with a plate and throw it in there. You have to maintain, you know, the shape and the distance from the edge. We've got a lot of engineering that went into that to create distributed pressure to ensure that we're not distorting any of the margins and the surgeons can make informed decisions. The next piece is really, you know, our onboarding for the surgeon. We've been developing this technology since, you know, since before 2013, and we've accumulated a mass of images in different tissue types, different cancer types within this, within breast cancer.
If a surgeon is going in to perform, you know, a partial mastectomy on a patient with very dense breasts and they know that it's ductal carcinoma in situ, they can look at what exactly that looks like in our Atlas image library, so they're prepared and ready to go when they walk into the OR. Lastly, in clinical validation today is our AI software.
Today it's very buzzy, we acknowledged pretty early on that surgeons, you know, they're not used to reading grayscale images like you just saw. AI would be a key enabler. We've been working on this for several years, amassing a library of fresh tissue images with cancerous. You know, classic medical imaging AI problem.
You have a lot of negatives, but not a lot of positives to train your algorithm. We've been doing it for many years now, and we were really excited with the results and launched our AI into a clinical validation study, which we'll talk about here. This was a three-part project, what we're calling the ATLAS AI Project, funded by the Cancer Prevention and Research Institute of Texas.
This is a government-funded body from the state of Texas, tax-based, so you can imagine the level of diligence that they went through to validate the IP, validate the business model and the technology. We were really the only product development company in our cohort to be awarded a grant, so really excited about that.
The first part of the project, like I just mentioned, was getting in pathology labs, getting in the OR, collecting images to train and validate our algorithm. Really pleased that we exited that part of the project with a 0.94 AUC. In the AI world, effectively think of this as a 94% accurate algorithm being able to distinguish suspicious parts of the image to non-suspicious parts of the image. That's, you know, best in class. This is, you know, AI applied to mammography is a little ahead of the game here for us, and kinda best in class there is 0.85 AUC.
Really, really pleased with where we're at there, you know, we really wanted to collect data until we were able to achieve a result like that to de-risk our pivotal clinical validation study that we're in now. The next stage was every surgeon who is participating in our clinical validation study went through a reader study.
We put them through, effectively, a robust training program and a bench-top evaluation to ensure that they could successfully read our images and successfully achieve the reoperation reductions, based on that data. You know, really pleased with that result as well. We started launching physicians into the actual study. We had eight planned sites. We've actually gone back to the FDA and asked for a couple additional sites, which they've approved.
We have about 15 surgeons enrolling patients in the study right now, and we're about halfway through those patients, looking to get to 330. With now kinda all of our sites up and running, our enrollment rate has really picked up lately. Looking to have top-line data on that, which is really a, you know, a key event for us later this year.
The way I, you know, the way I frame it is, you know, this is instantly expanding our hunting ground from the 20%-30% today of early adopters to 100% of those, and it flattens the learning curve. You know, this is when we would anticipate kind of our hockey stick commercial ramp-up from there. On the microeconomic side for our customers, you know, I walked through this a little bit earlier.
You've got one in four patients coming back for a $16,000 reoperation cost. This is averaging out to $4,000 per patient. What we've been going to these early customers with is saying, listen, you know, we don't have the AI. We don't have the data from that randomized controlled trial. You can pretty safely bank on cutting out half of your reoperation.
Your 25% goes down to 20% or goes down to 12.5%. We take that $4,000 cost, cut it down to $2,000. Thus, from a cost avoidance perspective, creating $2,000 worth of value. We just want to split that with you 50/50. You pay us $1,000 per procedure to use our technology. You net the $1,000 in savings.
What I can tell you is, you know, anecdotally, these commercial users are blowing those results out of the water. We've really seen a much, much better efficacy than I would have anticipated. They're going from Well, some started early in the, you know, in the early double digit, 10%-15% reoperation. They're going from their benchmark reoperation rate down to early single digits.
You know, 3%-5% reoperation rates with our technology before This is just with our flagship launch before the AI is on the market. You know, really netting very positive savings and efficiencies and happier patients and customers along the way. As we think about building that up to our, you know, scaling the addressable market.
Today, if you look at the kind of the bottoms up view, there's about an average of 150 breast lumpectomies per hospital per year. As we build those up, we, you know, we see we're kind of playing in the early 500 hospitals right now. Again, that 20%-30% top-down view is this early adopter market's around $50 million-$60 million revenue for us today. Again, we turn that 20% into 100% with the AI later this year.
On that note, this is kind of our go-to-market strategy. Today, like I said, we're in the early adopters trying to get reference sites, and we're up to an install base of seven units today across the U.S.
What I, you know, what I'm really encouraged by with those early users is we have a little bit of a microcosm of what we're trying to accomplish with the S-Series. We had our first user in the Dallas market. Now we're up to three users across the Dallas market because it's just expanding after that pioneer surgeon picks it up. Same thing in Southern California. We just announced yesterday that we have a new installation of the first in Utah.
What you're seeing from this, the demographics across these users is a number of, you know, large integrated delivery networks. We're in the largest for-profit hospital system across four installs there, all the way down to physician-owned ambulatory surgery centers, where the site of a lot of procedures are moving to the ambulatory surgery environment to save the cost that we were talking about earlier.
We're really establishing the, you know, validating the commercial model for expansion once we have the B-Series on the market. B-Series, we're in the clinical validation study today. Really excited about the physician engagement we have there. You can see the variety of sites that were participating with us, you know, in the, you know, from, again, rural community hospitals to the top five cancer centers in the U.S. there.
you know, we just had our Society of Surgical Oncology conference where most of our investigators participated or presented. A ton of buzz about the technology. People are really excited about picking this up and solving margin assessment with it. Once we have that clinical validation study done, we'll then move on to the AI approval and really begin to ramp up our commercial efforts there, leveraging the reference sites we're building today with the S-Series.
As we think beyond the surgical breast market, we see a $3.7 billion TAM from that single-use consumable I talked about. We really have two opportunities for expansion there. The first is into other tissue types, what I, you know, kinda call horizontal expansion.
You can see the blue part of this pie chart where the size of the market for all those other applications. You know, in the case of breast cancer surgery, we talked about the economic impact in addition to the clinical impact. You know, in the case of prostate cancer, those patients are going for additional radiation. If there's positive margins, radiation's very expensive. In the case of pancreatic cancer, a positive margin cuts your 12-month survivorship by 66%.
We're actually saving lives by achieving negative margins. The surgeons don't wanna go back in and reoperate on the pancreas. A ton of value to be added kind of across all those different tissue types. Also, we can play across the continuum of care. The same situation exists when you go for a biopsy.
The, you know, interventional radiologist will take a biopsy of the specimen. They have to send that to pathology. A lot of studies showing the cortisol impact of patients waiting and just the, you know, the lack of being able to expedite treatment, waiting for pathology to evaluate your biopsy results. Bringing that to the point of care is pretty exciting.
Post-surgery, what we talked about from a pathology perspective is they have to take a sample of the specimen to put it under the microscope. When they're looking for positive margins, they're actually only looking at about 16% of the margin when they're evaluating it. That's been some of the most robust feedback from our early commercial users is, you know, "I didn't realize just how much my pathologist was missing when they were looking at it.
I'm finding a lot more disease because I'm looking at 100% of the margin at the real time." you know, our surgeon customers want to retain control of that decision-making, so they're happy to embrace that as well. you know, a lot of value to be added there across the continuum of care, as I said. With that, I'll open it up for questions. Thank you. Okay.
Yeah.
Oh.
How it's going with the AI study right now?
Yeah. I'd say, you know, like, most studies right now, we're behind where we wanna be. Definitely, you know, we wanna acknowledge that. I think it's been a, you know, a constant piece of feedback from our investigators, is just the difficulty in staffing, and the research staff in particular. You know, we've had instances of pathology labs closing, delaying patients. Instances of study coordinators turning over pretty regularly.
Now that we have, you know, nine sites up and running, all those surgeons effectively in the randomized portion of the study, really encouraged by the enrollment rates we're seeing. You know, we've got another couple of sites we can add that we're actively doing diligence on to get through the contracting process. Looking to mitigate any, you know, additional risk that could pop up from enrollment rates.
Again, really excited about the feedback from the investigators. you know, they're encouraged. They're recruiting their friends to look at the technology, you know, we're feeling really good about the timeline to get top-line data later this year. That, you know, for us is probably the biggest milestone in the next, you know, several years for us.
Oh, thank you for sharing. Talking about neo markets and the, you know, geographies.
Mm-hmm.
Can you disclose, share where you're just already licensing or setting the-?
Yep
... technology?
For other tissue types, we've got a lot of really strong data on correlation with pathology. We actually just earlier this year published a paper at the College of American Pathologists showing the utility in head and neck surgery. And you know, I can tell you every evaluation we go to, you know, eventually it makes its way to another surgeon and they say, "Oh, The breast surgeon's using that? I wanna try that in endocrine for thyroid."
You know, urologists wanna try it for bladder cancer. The demand is definitely there. We, you know, wanna focus our resources on getting the breast market, you know, established and validated, we'll look to expand. We'll let the market pull us into those other tissue types.
From a geographic perspective, you know, we wanna get the AI clearance, the country of origin clearance on the AI, and then that'll be the one that we take to the international markets.
Can I ask the question about COVID?
Mm-hmm.
The investment community seems to have forgotten about COVID. How much did it impact your progress? I mean, you really were launching right in the heart of it, and I'm assuming you have an idea where you would be by now if it hadn't been for that.
Yeah. There was definitely an impact, I'd say, on both fronts. You know, well, in some ways, there were some positive outcomes too, so I'll touch on that. You know, I think launching, getting access to hospitals when you're launching a completely novel product was extremely challenging. You know, that at least set us back a good six, eight months at the very least.
Same thing on the study. Patients weren't going to get mammographies, so the, you know, the front side. Our supply of patients was getting choked off. What that actually. The positive of that is it kinda shifted the path of care for a lot of those patients. If, you know, ORs were at capacity, so their patient's ability to get into the OR was a lot, you know, slower.
Surgeons started doing a lot more neoadjuvant chemotherapy to try to contain, you know, the progression of the disease. The fact that we're collecting our data, and we collected our algorithm data on a greater proportion of neoadjuvant patients, we think that sets us apart going forward because we have, you know, that unique data set going forward.
That, you know, that was probably the in the middle of COVID impacts. What we're still feeling the lag effects of is just the staffing issues in hospitals. You know, if you listen to the publicly traded hospitals, they're out there, you know, talking about spending three times as much on contract laborers to, you know, contract nurses to have their OR at capacity and, you know, finally starting to wane.
I think that's why you saw a really big Q4 commercial quarter from us because, you know, the surgeon can feel confident that when they walk in the OR, they're gonna have a good team of nurses, so they don't mind integrating technology like ours into their workflow there. Still a little bit in certain specialties on the clinical side that's still coming up a bit. You know, we've got radiologists and pathologists at capacity, I think. You know, that's still taking some time to work itself through the system.
Any other questions? Oh, one in the back.
Yeah. The question was asking if pathology is still doing the post-surgery evaluation on those samples. Yes, that is the case. You know, I wouldn't anticipate that changing in anytime soon. The pathologist isn't only evaluating for margins. They're also doing, you know, other Oncotype DX stratifying and things like that. You know, it is a part of dictating the path of care for those patients, so I'd say we're several years until, you know, substituting that completely.
You know, that said, pathologists are embracing our technology. You know, they're usually our secondary champion in the process because they say, "All right, you're taking this snapshot, effectively a microscopic snapshot in the, in the OR.
When I'm gonna block that specimen, I wanna look at your image so I know where to block it," which is really, you know, helping the throughput, helping their accuracy, et cetera. Yeah. One of the things maybe just, you know, as we're getting slow on questions, one of the things that I think, you know, forward-looking like that we're really excited about is there's a, you know, particularly on the back of AI, there's momentum to, you know, less is more from a treatment perspective.
Particularly in these early-stage ductal carcinoma in situ patients, they might not go on to radiation if they have the right, you know, oncology assessment. It makes margin assessment that much more critical.
You know, we're building a unique data set of images of DCIS correlated with those Oncotype, all the Oncotype data, so that we can, you know, build a bigger algorithm, a bigger model around, all right, can we avoid all the costs, all the patient, you know, issues with treatment and radiation, if we can assure that we get our margins right the first time? There's a lot of, you know, blue sky opportunity and just shifting path of care for patients as well.
Can you give us an idea, each installed device, how many single-use can be used per month or per quarter?
I'd say today in breast cancer, you know, the average volume institution's gonna be around 150 a year. Think, you know, three a week effectively, which, you know, we pretty quickly get to 100% utilization from our surgeons. You know, I'd say one of the more challenging reoperation problem for surgeons is DCIS. You know, it's earlier stage. It's not as invasive as invasive ductal carcinoma, and they can kind of palpate, and they can see invasive ductal carcinoma. That doesn't create a lot of reoperation.
We've had a few surgeons that said, "Well, I'm gonna start with my DCIS biopsy, confirmed DCIS patients." It doesn't take too many operations for them to have a DCIS-positive margin in a biopsy-confirmed IDC case, they pretty quickly say, "All right. No, you know, I'm not gonna chance it anymore. I'm gonna use it on 100% of patients," which is, you know, really exciting.
The great part about expanding tissue types for us is typically all your breast procedures in a hospital are done on one or two days. If we, you know, have a day of breast procedures, a day of, you know, head and neck procedures, a day of hepatopancreaticobiliary procedures, then we really start to drive utilization on our device, on every installed device for sure.
Well, thank you, Jeremy.
Welcome.
Concludes today's presentation.
Thanks, everybody.