Tamir Wolf Presents Theator at LSI USA '23

Transcription

Let's talk about the future of surgery. For those of you who listened to Fred molar earlier on today, this is the immediate follow up. So we've been talking a lot about hardware. Today, let's think about what we can do with actual data. And with software at a scale that I think has never been seen before, surgery, if we're talking about the world of surgery at large, it's still very much an apprenticeship model hasn't changed in hundreds and hundreds of years. As an apprenticeship model is very limited. And it's very limiting. It's individual. By nature, the impact is patient at a time. And leveraging data and leveraging artificial intelligence, I know there's a lot of buzz around it, we can actually do things very differently. And at Theator what we've started to do is actually augment human surgeon performance by leveraging these capabilities. Now in the future, the idea like everyone's talking about robotic platforms, again, around the hardware. And we may have in the next several years robotic platforms that are able to pass the future autonomously. But what happens when they pass the suture in the wrong plane? What happens when they nick a vessel? So these robotic platforms are hardware absent intelligence, and we're calling it surgical intelligence is not really going to work. We need to augment the hardware capabilities, and leverage intelligence, which I'll talk about in order to make things really happen. And so this is the world that we're starting to build at Theator today. But what really makes a successful surgical procedure or what makes a great surgeon. So several decades ago, Dr. Frank Spencer, is known to have mentioned that it's really about situation awareness and decision making. That is really what differentiates an amazing surgeon, from one who's mediocre. Although there are no mediocre surgeons, obviously, they're all great. That is really what differentiates like someone who is who is the best from from all the rest. But it's very hard to understand what best practices look like today, especially in a world where there's significant variability in in the way that surgery is performed. There's variability within, you know, between geographies there's variability within with within a city, these are two examples just out of New York. And there's also variability within hospitals. And so this is really what we need to understand in order to then guide human surgeons and in the future robotic platforms. And so more and more hardware is only part of the solution. If we want to have impact not at on a patient at a time, but we want to have impact at a larger scale. In order to understand best practices and disseminate them in the world in ways that are very, very different from what's being done today, we need to think differently. And that's obviously where software comes into play, because software is limitless. And so this is everything that we're really leveraging at Theator, and doing. So if you agree with me that data is powerful, then this is going to be mind boggling, most of the data today is lost. And you know, we have millions and millions of minimally invasive procedures that are done under visual guidance on a daily basis or on an annual basis. Across the world. Most of them are not captured, they're not stored. And they're definitely not analyzed. And so ultimately, the status quo or what we have today is that we have really good understanding of outcomes, but we have very limited understanding of what leads us to them. So we're not talking about sci fi, we're talking about like the world today. And so Theator, what we're doing is starting to connect the dots, we're enabling routine capture of intraoperative performance, and Lincoln, linking it to who the patient is and what the outcomes are in order to understand what these best practices look like. So data is powerful, gathering it is important, but we need to make it actionable. Because gathering hours and hours of intraoperative footage is meaningless. And so what we've done at Theator over the past several years, is create a platform that automatically and hundreds of procedures across multiple specialties, identify key components or key moments within surgical procedure. So these are steps, events, milestones, complexity analysis, ultimately decision making points within a surgical procedure that make it or break it. It's all automated. And it's all done in in real time. So this is an analysis my team did we actually have more data than Netflix at this point in time. And all of this data is, is really structured. So it's all the meaningful aspects of intraoperative performance that can make a difference. All of them are identified Now technology for the sake of technology is really useless. We need to leverage it in order to have actual impact on patient care provide value at an individual patient level, obviously at a departmental level and at the level of a health care system. And I'll give you a couple of examples. The first is a year and a half old, female patient that had to undergo surgery. During surgery, there was a moment in the procedure or something happened, the surgeon didn't notice it, software picked up on it. We're not providing real time information just yet. But after surgery immediately after surgery, the software provided that input, and they were able to take care of the patient in ways that were much better and avoid complications that otherwise would have gone missing until something really bad happened. This is a patient's at a patient level. Another example, one of the best practices in hysterectomies, is observing the ureter in order to avoid injury. At a department that had issues with injury to the ureter, we actually asked the department chair, so what percent of the time do your surgeons take a deliberate view of the Riveter in order to avoid it? The answer immediately is like close to 100, we have an algorithm that automatically identifies whether or not that happens. And when we took a look across the board, over four months, we saw that that's really far from reality. And what you see here in the graph is actually the percentages of the percent of time over four months that surgeons actually took a deliberate view of the rotor in hysterectomies, a commonly accepted best practice, far from 100%, some don't take a look at it at all other take a look at it some of the time. And the idea is that once we surfaced this information, this actually push them to a cultural shift, where they started looking at things a lot closer. And so it was there was an increase in achieving the safety milestone. It was interesting, it was also coupled to more efficient procedures, flow surgery was done better. So quality was coupled to efficiency enhancement in the operating room, obviously, very important. And when we take a look at two different hospitals, we see that some hospitals or departments are clustered. So they they're everything is standardized. They take a look at you know, the specific metric or best practice. And they do it routinely each, each and every procedure very early on. And in others, this doesn't happen. So obviously, it connects with variability that I was talking about. And our approach to standardization is something that you can see here. And when we take a look at entire enterprise systems that have variability within them, but they have no idea why they're happening, this is where it really gets interesting. Don't have too much time to talk about this. But you can see here, two different hospitals under the same logo, or two sites under the same logo, one with very high complication rates, and another with low complication rates in the same procedure. Best practices are not disseminated. And they don't have a clue as to why things are happening. We're helping them here. In order for things to be meaningful, you have to have them in real time. So this is an example of a conference a few months ago, where for the first time we demonstrated in real life procedure being performed, and transmitted to the stage. We we demonstrated real time identification of key components of the procedure. And this isn't this was cool. It's kind of sci fi. But like the really important aspect of this is that if we are moving into a world where we will have real time decision support, identification of key components of surgery has to be done. Pretty much like with very low latency and in real time. And so the idea is to change this apprenticeship model instead of having one surgeon behind my back telling you what to do in this case, to have the experience of 1000s of surgeons in many procedures right in front of me, showing me the impact of the decisions that I make as a surgeon in a procedure from a clinical perspective, financial perspective, etc. This is like decision support as we envision it in the future. And so hopefully, I've been able to share with you what we do at an individual surgeon level. Obviously the impact on patients as well as the impact that data structuring artificial intelligence analysis can actually provide today to health systems. 20 seconds over. Thank you.