Hello, and welcome to the NVIDIA AI Podcast.
I'm your host, Noah Kravitz.
The theme that AI should and does augment our human capabilities, give us superpowers, if you will has been central to the podcast for some time now.
People as conductors of AI tools, be they on-screen tools or tools manifested in the physical world, is a metaphor that's been used more than a few times around here.
Today, we're going to dive into that idea quite literally with Moon Surgical and Maestro.
I'm going to let our guest talk about it, but it's an amazing surgical robotics technology.
That's more than that.
It's a platform.
It's a reimagining of minimally invasive surgery in the modern world.
Anne Ostrois is CEO of Moon Surgical.
She's here with us to talk about the company, how it got started, her own background in medicine and kind of what led us to the point where Anne and her teams at Moon are really rethinking and making real a new approach to surgery.
So, Anne, thank you so much for joining the podcast and welcome.
Thanks, Noah.
Thanks for the invitation.
So I'd just love to turn it over to you to tell us about Moon, how the company got started and the vision to humanize robotic surgery.
Yeah, with pleasure.
So, you know, Moon got started really in a way many years ago, right?
The company... really got going in 2020 or at the end of 2020.
But it was based on work that had been really pioneered by this surgeon, Professor Gaillet, and a robotics lab based in Paris at Sorbonne University where they had been looking at this concept of really augmenting the surgeon, but leveraging the surgeon and what their capacities are today, right?
And really turning them into super surgeons, essentially, for years, right?
They had experience, of course, from other robotic platforms and surgical approaches, but had this frustration or concepts that hey, you know it's great to be completely changing the way surgery is done, but wouldn't it be greater or easier to implement if you could essentially just turn the surgeon into this a lot more powerful surgeon and leverage on?
You know the standard of care that has been developed in surgery over many decades.
And if he did that, wouldn't it essentially broaden access, right?
Because it would be easier to teach, easier to learn, easier to deploy.
And so they had this goal in mind and had been toying with technologies and approaches to really get it to life, and had a demonstrator in the lab, which is what I saw in early 2020.
That got me to think about okay, how do we apply this in an operating room?
What are the current pain points that hospitals are facing in terms of staff shortages, in terms of, Just you know, not only keeping and increasing efficiencies within the operating room, but also empowering surgeons to do things that are a lot more tailored to how they do things, to who the specific patient they have in front of them is.
And so we really try to write the specifications for our platform so that it would really meet all of these needs and be equipped to grow into that over time.
Right.
How specific, how individualized are these procedures?
If two random people go into the operating room for the same, and I'll let you, if you want, give an example of a procedure.
You know, how similar, how different are the experiences?
That's a great question.
So, I mean, overall, they're fairly similar in a way, right?
You know, surgical techniques have been described and are being taught in a way that it's fairly structured.
You have a little bit of geographical specificity as to how they do things, but I would say minimal.
However, every single surgeon does things differently.
They have their own surgical preferences, as you call them, to the point that in operating rooms there are physical cards, like cardboards, where those surgeon preferences are written down.
And this would be, you know what instrumentation they typically use, how they like things set up at the beginning of a procedure and then what are the specific steps or things that they might need throughout the procedure, how the surrounding staff you know should be supporting them, where they should be positioned, et cetera.
So really, you know, surgeon preferences.
All the preferences, yeah. are unique.
And then each patient is unique, right?
Their anatomy is unique.
And, you know, the way the surgery unfolds is to some degree unique.
So you know, the sequence might be similar or supposedly similar, but the execution is incredibly variable.
I mean, this is well known, right?
We all know, you know patients who are supposedly going in into a very benign surgery and then things happen right.
Because a A part of it, to some degree, is unpredictable.
A part of it is just sheer variability.
So I would say fairly variable after all.
Right, right.
And so, I mean, how does Maestro work in practice?
And you have to go into great detail.
But you know, I've seen clips of it and the robotic arms and the whole thing.
But how does it work?
And then what I want to get into or have you talk about, if you will, is you know how?
Then it's all.
The system is redefining how surgery minimally invasive surgery happens and what's possible.
But maybe if you would start with kind of the high level of how it works.
Yeah, so the maestro system is really about empowering and augmenting the surgeon with two additional arms, right.
The surgeon typically has two arms, so they hold and maneuver two instruments right, which are called the active instruments in the surgery.
So these are the instruments that the surgeon would use to cut dissect, you know, take things out in, et cetera.
And then typically, you would have two additional instruments used in these surgeries that are absolutely critical because they really deliver basic functions to the surgeon, which are the vision right.
The ability of the surgeon to see inside the abdomen, which is delivered through a camera that's inserted inside the body.
And then the second function is access.
You know, tissue exposure, the ability to present the target tissue to the surgeon in the right way at any given time, right.
Because you get into the abdomen, but you're not usually not operating there, right?
You need to put things aside.
Right, right, right.
Is it odd that, as I'm listening to you, my own appendectomy scar seems to be tingling a little bit, or is that a pretty normal response?
Okay, good, good.
Sorry to interrupt.
I'm breathing fine.
Okay.
Yeah.
So, you know, these two critical functions are kind of inherent to surgery.
You can imagine that they are very surgeon, specific in terms of what their preference might be.
You know what they want to see, at which distance, how dynamic they want this to be.
And similarly, how they want tissue presented to them during the surgery.
I mean, this is incredibly important and incredibly individual in terms of how it manifests.
And so what we're doing with Maestro is we're giving the control to the surgeon over these things the vision and the tissue, exposure and ability to access tissue.
These would typically be managed by someone who is a surgical assistant first assist but that way of doing things is inherently flawed, right?
I mean, you're relying on someone who's standing somewhere else uh, around you know the operating table to position things and anticipate things in the right way for you, which which like, is really going to be imperfect, no matter what.
Inherently right.
Yeah, shut you really well and have to get it.
And even though you know you might have a fantastic versus that you've worked with forever, this person is better off doing higher value tasks, right?
Bringing the right thing, anticipating the next move, prepping the next patient, etc.
So the surgeons absolutely love the fact that you know they're in control of all these different elements, because no one is ever going to better assist them than themselves.
More efficient, more consistent, and more confident after all.
That's amazing.
So robotics has obviously been around prior to 2020 when Moon was founded.
And the idea of robotics and surgery has been around for some time as well.
When did AI, machine learning, deep learning, computer vision, everything we talk about when we talk about AI When did that sort of enter your vision as not just a thing maybe to explore, but a no, this like this, needs to be part of it.
Robotic surgery, AI, you know, it's all part of the vision.
Okay, so this is a great question.
And I think that AI and what we're doing today with you know, data and artificial intelligence was there from the beginning.
Because, if you think about it, this notion of surgeon preferences is extremely specific to a surgeon but can absolutely be taught to a system, right?
Right.
It is about how a surgeon likes their tools and cameras positioned over time and how they might dynamically be managed throughout the procedure.
And the more a surgeon uses the system, the more you can learn from that.
And it's going to continuously perform better and better.
And then the second aspect behind that is okay, you're equipping the surgeon with two additional arms and hands, but how are these two additional arms going to be controlled and moved around, right?
Because the surgeon, if they use two of their hands to maneuver four arms, it means that they have to let go and grab something and move it, et cetera.
And this is really where the concept of physical AI comes came into play before it even existed probably, which was hey, how about we learn and leverage this data for learning to actuate our maestro arms and control them?
Which sounded like a bold idea, right?
Hey, we're in a regulated environment, we're in an operating room.
Is it really feasible to do this without risks, et cetera?
But, if you think about it, the only way to deliver those efficiencies and to empower the surgeon right.
Otherwise they're always going to be limited by efficiency.
You know their own, you know body and capabilities.
Right.
No, absolutely.
As you were describing first the cardboard sheet with the preferences on it, and I thought oh, that's like a preference pane on a software application, right.
And that, you know, translates...
But then you know, kind of more interestingly and importantly, that kind of dynamic understanding and applying that to the idea of well, now you have forearms, but this surgeon for their whole life has had, you know, as you said, likely to but how many arms they have.
And so I'm thinking me as the patient.
I don't know if I want, like the surgeon, to have the burden of trying to think about how to manipulate the extra arms right.
And so this notion of, well, how do you just make it an extension?
And that's what AI is so good at.
Yeah, absolutely.
But at the same time, as a patient, would you rather have arms from a platform like ours that behave like a surgeon or that are just, you know, systematically controlled by an algorithm that Oh 100?
Yeah.
The human expertise.
Yeah, absolutely.
But that's the whole point about humanizing the platform.
Right.
And turning... with that surgeon who is operating today on that particular patient.
All right.
So how did you come to start working with NVIDIA?
And maybe tell us a little bit, if you would, about the partnership.
Yeah, so working with NVIDIA has been quite an incredible journey, which started, I have to say, by luck and serendipity.
I mean, totally right.
I mean, NVIDIA has very talented scouting people out there in the field.
And one of them, you know, happened to be in Paris and focused on healthcare.
And reached out very early in the life of the company.
So clearly someone who had been, you know, keeping the pulse and doing really well.
But you know.
So we got into the Inception program and got familiar with the capabilities and what we could access.
We, as I said, had this vision that we wanted to equip our platform with very, you know, extensive sensing from the beginning.
Whether or not we would use it immediately or further down the road.
We wanted to make sure that the infrastructure you know accounted for that.
And similarly, we wanted to equip our platform from the get-go with very edge computing, right?
We wanted to make sure that it was able to, you know, basically live for many years without changing the hardware and systems in all these evolutions, et cetera.
And so we sort of embarked on this technical bet with the NVIDIA RD team where we were like okay well, you know, let's assume we're going to put a medical grade GPU in this thing and we need it, you know, by that day.
So, you know, can you guys make it?
The NVIDIA team was incredibly responsive and reactive.
Amazing, yeah.
To use this as a pilot to really, you know, get familiar with this industry and write specifications together and test it.
And so we were on our deadline, they were on their deadline.
We kind of made sure that the operating plan sort of coalesced at some point.
And it worked, right?
So you know, when we got to our commercial product and we were ready to submit it to the FDA, we had the NVIDIA PPU in there, right.
Right, right.
That's fantastic.
And how the product was approved.
And so what this has enabled us to do since then is really built on that.
You know, it is the training environment.
It is now the simulation environment and developing these features that are enhancing the product right.
And we've been deploying some of them over the last few months.
I'm speaking with Anne Asdwa.
Anne is CEO of Moon Surgical, whose maestro surgical platform is really, as Anne's been talking about, revolutionizing the concept of being a surgeon, being a super surgeon, augmenting the human surgeon's capabilities and kind of just rethinking what the operating room of the future is.
I was going to say will look like, but really it's an is at this point which leads me to want to ask you about the impact so far of everything Moon's been doing and Maestro.
What does Maestro allow surgeons to do and deliver to the process that just wasn't possible before?
Yeah so Maestro, allows the surgeon to do more in a way that is more efficient, but also more specific, more tailored to that patient, more tailored to the way they do things in the best way, with fewer resources, in a way that is more autonomous and that is going to deliver greater quality care in a way that is very accessible.
And so typically, you know, surgical robots have been implemented in select hospitals and indications because they are very, you know, well designed for complex procedures, right.
They tend to slow down the operating room and the workflow.
And so we were really attached to developing a platform that's would be easily accessible, easily adaptable.
And that would be basically an asset to the surgeons and their staff in high throughput environments.
Right, right.
Makes sense.
And so some of the initial results that I was looking at before today show reduced variability in procedure times and then also increased surgical quality, as you were talking about.
Can you talk a little bit more about those findings and specifically you know, why they're so important?
Yeah, absolutely.
Some of it goes back to what we were saying, right?
So basically, if the surgeon is in control of all these different instruments and elements during the surgery, they're going to be managing all the different steps, all the different transitions between those instruments, right?
So, rather than having to coordinate with someone, anticipate et cetera and communicate, it is seamless, right.
Right. system behaves specifically to that surgeon.
The surgeon is in control.
Some of these tasks are automated, leveraging AI, as we said.
So it really makes it more consistent, more efficient for the surgeon to go from point A to point B.
An analogy that we use that is fairly simple but I think illustrates it really well is the analogy around how you learn to drive a car right.
When you're in a car, you have two people in the car and you're splitting roles and responsibilities and functions between those two people, right?
And as a result, it's a little bit clunky, right?
Sure.
Because you would need to coordinate everything perfectly for that drive to be Right.
And then, at the minute that the driver has control over the gearbox, the brakes, the steering wheel and the vision, it is, of course, a lot smoother.
It's a lot faster and it's also a lot more consistent when they point A to point B.
So it's a similar concept.
Right.
No, that's great.
That makes sense.
And from the patient view and, as I mentioned before, I've had a couple of surgeries that were long enough ago.
Now that all of this is just man, why couldn't I?
But I have kids, so I'm happy for them that they'll benefit from this.
But what are some of the specific benefits you've seen from the patient point of view?
And then as well, you spoke some to the other people in the operating room, maybe from the hospital's perspective as well.
Yeah, I mean, the patient is central, right?
I mean, everything we do is ultimately about delivering better care for patients.
We've treated close to 2,000 patients, and it is a daily source of satisfaction, of course.
And so from a patient standpoint, as we said, it's about access to the best quality care.
And if you think about it and more specifically, what does reduced variability mean?
Well, reduced variability means that, as you were saying earlier, two patients getting in for the same procedure twice are probably going to have a more similar outcome than in the past, which is what we want.
You know, complications or outcomes in surgery.
It also means that the surgeon is likely to end their day on time and basically get you on the schedule as anticipated, which nobody likes when things get delayed or you're kind of rescheduled etc.
So it means a lot more consistent patient experience when they go through the surgical journey.
It also means, for instance, in emergency cases, making sure that the surgeon can have the resources to operate in a way that is minimally invasive.
There are a lot of times where, during nights and weekends, the surgeon doesn't have the staff that they would need conduct minimally invasive surgeries, in which case they would either do an open surgery, which is a lot more invasive, or they would just table that surgery to the next day.
So you know, giving the surgeon a lot more autonomy is also a way to ensure those procedures can be done in a timely way.
So all of these things are important for patients.
When you mentioned a moment ago helping to keep things on schedule and ending the surgeries on time and the doctor's day the surgeon's day, on time.
And then you were speaking about the patient, but it made me think about the doctor's point of view, the surgeon's.
Is physician fatigue.
I don't know if it's specific to surgery, but it's something I've heard about outside of this context.
How big of a problem is it?
And I would imagine it's something as you've been talking about that, a system you know that maestro can really help alleviate.
Physician fatigue is absolutely real.
You know, we did our first in-human study in Brussels in Belgium, with a surgeon and he used the system over 50 cases.
And he told us after a few weeks hey, when I get back home in the evening, my wife tells me that I'm, you know, a lot Nicer than before.
So like what's going on?
And, you know, I mean, he attributed that just to his own fatigue level, right?
Like, you know, I end my day in a way that is a lot more relaxed.
It's about both the physical and the mental load.
The mental load is about constantly adjusting coordinating, communicating with this assistant resource, which is very taxing.
And even with the best resources, it's going to be imperfect and frustrating.
And then the physical load is really about, you know, being in position that are not ergonomically optimized, because you're sharing your workspace with someone and because you don't have easy access to everything.
And a lot of surgeons have musculoskeletal pain and you know they have to get infiltrations and this and that.
And that is very training as well, right?
So, yes, we absolutely have many reports from surgeons telling us about their fatigue.
It's not the easiest thing to quantify, but you know, operating rooms are short staffed, which is about the nursing staff but also, to some degree, the surgeons.
I know it's not technically quantified, but if you're coming home from work and your loved one, your roommate, whoever you share your home with, is saying you've been in a lot better mood lately.
I mean, I think that's saying something.
But those?
I was going to call them ripple effects, but then I thought well, they're not, because as you described it maestro, is about humanizing the whole experience.
And so these things I wouldn't have thought about the ergonomics of sharing an operating theater with you know, assistants and other people.
But I can relate to, you know, my back hurting at the end of the day if I've been in a bad position.
And so it then sort of you think about the trickle down to.
You know the surgeon's happier.
They're more relaxed and less fatigued.
So they're no doubt giving better care, even if they were already excellent.
It's always good to be rested.
Patient has a better experience.
The hospitals are more coordinated on.
So it just seems like this virtuous cycle which I was thinking about when you were talking about building the platform with the sensors, even if you knew you didn't need them right away.
AI is all about collecting the data and using the data to learn.
And so You know, it's just it's it just sounds so great.
I do have two parents who are retired medical people, so maybe I'm a little have a soft spot for it.
But it's just it's great to hear about.
So as you look ahead a little bit, we like to end on kind of a future looking note.
What do you see as the potential say, over the next five years or you can shorten or lengthen that if it's better, but the potential for Maestro and NAI more broadly, to transform the operating room.
Yeah, so it's a very exciting journey, right?
I mean, I mentioned we equipped the system with a lot, but we're only scratching the surface of what we're leveraging today.
So one of the first things we did was to automate the camera movement, which is something that is now in our commercial product and, you know, fully feared by the FDA.
It is the first physical AI product in the operating room, which has been incredibly exciting.
One of the things that we did a few weeks after that was get a regulatory clearance about our ability to evolve that AI algorithm over time without having to go back to the agency for approval each time.
You can imagine that regulatory bodies like things that are frozen.
Not to get us off topic, but how novel of an idea or of a concept was that on their end, on the FDA's end?
It was the first time they were seeing it.
Yeah, okay.
Yeah, yeah.
So it doesn't come out of, you know, education on what was needed to get this through.
So you know, I think it gave us a lot of knowledge now on how to get additional features into the product.
And so I think the plan is to leverage that sensing a lot more right and turn this into workflow efficiencies more.
It is about things such as you know, enabling operating rooms to have dynamic scheduling right, knowing predicting the end time of a procedure, making sure that this is adjusted based on how you know, the previous surgery is going in real time.
Delivering those notifications to the staff so that they can adjust when the The downtime.
It is about optimizing the staffing and the resources in the OR right.
We will know when a given surgeon is able to do a chunk or a procedure without assistance, right?
So based on that, we can optimize how staffing is deployed over the operating rooms.
It is about helping them manage their inventory.
We can see what's going on in the OR.
We can see what's going on in the abdomen.
We know what they're using.
Based on that, we can help them with instrumentation and inventory management.
It's about, you know, case notes.
I mean, I don't know.
You said you went through surgery.
Typically, you don't have great surgical reports, right?
I mean, these things don't really happen because surgeons are busy doing the surgery, right?
These are things that you can absolutely do. you know, automate and get value out of, right?
Absolutely, yeah.
So there are many things.
And then providing feedback to the surgeon and the staff.
What makes you a better surgeon?
In which cases have you been more, you know, efficient or have you delivered, you know, better care?
And what was that based off?
And it's something we're very excited about is that continuous training, right?
Really creating that feedback loop into the operating room, which is incredibly exciting.
Yeah, I can only imagine from your perspective, but it's exciting to hear you talk about it.
So, with all of this and you've talked about this throughout the human, the surgeon, being central to everything.
But sort of just to land on this initial concept of the extra arms issue.
How do you think about balancing all of the innovation that has happened, that Moon's Moon's been able to accomplish to date and everything you were just talking about with?
You know both the physical operating process, but then all of the data and the background and all the things that you can do with it.
How do you balance that with keeping you know just the physical reality of the surgeon's hands are healing another human being, right?
How do you think about maintaining that balance?
Yeah, it's a great question.
I think two aspects to that.
I mean, first, you know, we're in a regulated industry, right?
So there's only so much that you can change at a time.
So we're pacing ourselves.
But also for us, the thing that has been incredibly critical in getting the product and the additional features through regulatory bodies is the fact that the surgeon is at the OR table in the operating theater and can control everything manually and override anything at any given point.
And this is really a safeguard, right, for regulatory agencies.
If anything goes wrong, you still have the surgeon there.
You've been trained to operate like this.
You have not put the surgeon behind a console on the other side of the wall.
Right, right. back end of the room, right?
And so this has been incredibly helpful in terms of convincing about the risk profile, right?
The surgeon is still operating with their two hands.
We're assisting and enhancing them, but they're there with their traditional instruments and training.
That is really where these workflow efficiency improvements come from, and things such as managing staff, managing the inventory, managing scheduling, providing feedback and really continuously improving what they're doing in the OR, with insights for the staff, insights for the surgeon, insights for administration.
And the beauty of all that is that it's not regulated, right?
It's not during the surgery.
And so we can deploy those things at a pace that is about greater.
Right.
Fantastic.
Final words of wisdom, or just a message you might want to leave to surgeons, medical students, aspiring surgeons and patients, for that matter, who might be listening on the future of surgery and robotic surgery.
What would you like to leave them with?
Well, I think what we're introducing is really a completely new way of doing surgery, but also training new surgeons.
And as we said, providing access to high quality surgery to patients.
So it is incredibly exciting.
I mean, I think we're really at the beginning.
And you know, this vision that we have is going to be, you know, deployed over the next few years and, I think, has benefits for all of these stakeholders.
Right.
And so I would tell them to.
You know basically, get excited and see what's coming with Moon very shortly.
Fantastic.
And for listeners who want to find out more about the company.
The website best place to go.
Social media where would you direct them?
Hi, it's LinkedIn.
We'd be where we're the most active.
Moon Surgical.
Fantastic.
And Anastra.
Thank you again for taking the time to join the podcast and best of luck with everything you're doing.
Thank you very much.
