Yeshwanth Pulijala, Scalpel - AI-Powered Surgical Inventory Management | LSI USA '24

Yeshwanth Pulijala  0:02  
Imagine you're in an state of the art operating theater. You're working with a surgeon who is an expert in the field. For the past 30 years, the operating theater is serene. Everything was going quite well, but halfway into surgery, something goes wrong. Surgeon is holding an implant. That's not for the patient on the table, but that's for the patient in the next room. The Medical Physics has to come in. Nurse has to check these instruments. The next surgery has to be canceled. Wait, this is not a medical drama that I'm narrating here. This is my personal experience. I was standing next to the surgeon, and this happens every single day. Luckily, that patient was quite safe. But this is not the common thing that happens every single day where you have some kind of instrument that is either left inside the patient's body, you've got either a wrong instrument or a wrong implant inside operating theaters. Why does this happen in the first place, and is it really as big as a problem? My name is Dr eshwan purijala. I'm the co founder and CEO of scalpel. At scalpel, we make surgery safer and smarter by ensuring every surgery has the right surgical equipment in it at the right time, right place. The way we do that is by employing computer vision and robotics at different parts in the surgical supply chain, right from the warehouses till the point where instruments are used and outside into the stellar service divisions. What if I told you the next groundbreaking change in surgery is not one more medical device, but actually help you manage your surgical supplies? Don't believe me. Well, look at this. Every single day, instruments are misplaced lost, and this costs hospitals billions of dollars, and device companies alone spend tons of money in terms of ensuring that their assets are managed appropriately. Surgeons being surgeons, keep asking for instruments which are not inside the operating theater, so nurses spend more time outside the rooms fetching them that directly impacts patient care. 16,000 patients in the US alone have a retained foreign object directly affecting not only patient care, but also the operational efficiency. And this is one of the biggest pain points. And if you actually wonder why this happens in the first place, let me walk you through the journey. This is the image of a back table of a surgery. Can you guess the number of instruments in there? For an average knee replacement procedure, it takes anywhere between eight to 10 different tray types coming inside the operating theater, each having 20 to 30 objects within it. It's a huge manual process checking every single object that it is the right kind of object for that operation, as each one of us have different knee sizes and slightly different angles. These instruments have to match for this patient requirement. But wait, that's not all. Every surgeon has a preference card, by the way. These are not changed once the surgeon joins the hospital, and those preference cards are often manual. They are fixed, but somebody has to read this tight as this information, then pack a tray that looks anywhere from here to here. So surgical trays are quite complex, and it requires tremendous amount of training, firstly, to manage those objects in the first place. With increased number of operations in waiting lists, shortage of skilled staff members, the errors are imminent, and that's what we keep seeing happening. But this is one part of the problem, which is to get the right instrument in the train. How do you bring the train to the operating theater? This is more like you keep comparing surgeons to the athletes. This is more like relay runners. In a relay race, you have a baton being passed between different runners, with every second being so critical, you want to make sure that the baton is positive the right time. Any inefficiency means you lose the race. Let me tell you, getting the right train to operating theater is 100 times more complex than this, because the process starts from the device distribution company actually getting the tray from the striker, or Medtronic of the world. The device distribution gives it to the Driver. Driver driver gives it to the rep. Rep gives it to the sterile Service Division. Sterile service cleans the instruments, gives it to the scrub nurse, and then goes back again. Throughout this journey, across this journey, there are 10s of people touching the surgical instrument. We call these handovers. Every single handover, there is a chance of error, and that baton can be dropped, affecting Patient Safety directly. This is the problem we are addressing at scalpel. At scalpel, we have an AI technology, and we are currently applying computer vision and robotics to automatically pick and pack trace at different points in the surgical supply chain to ensure that every object is right for the procedure. The best part is, you don't need any physical tags on these surgical instruments. We've trained our computer vision algorithms on millions of surgical instrument data at different points in the surgical supply chain, right from warehouses to sterile service divisions, and this is how our system works as of today, firstly, by improving staff efficiency, reducing the cost, by eliminating any kind of. Errors, but more importantly, reducing the onboarding time of staff members. There is training that you can do preoperatively, so somebody that could be trained before they get onto the job. But the biggest part of the training is on the job. When you keep seeing new trace every single day, how are you going to help these people to pack those trace correctly? That's one of the biggest places we make it, where we make a huge impact. Another key area which is often ignored is the amount of waste that happens inside operating theater. Only 10 to 20% of instruments within a tray are used in surgery. What about the remaining 70 to 80% of objects they are going back into the circulation? So that's not only a cost, but also a carbon dioxide footprint to hospitals, and that's what we address. With respect to our technology. Majority of the current systems have some form of track and trace or data. The problem is not not having data. We have good amount of data. The problem is able to create useful analytics that people can take actionable insights from that, and that's what scalpel is able to deliver, but not only in one part of the surgical supply chain and being able to connect different other players making it that you can connect it to your ERP systems, EHR systems. So imagine a future where you walk into an operating theater and you've got the confidence that you've got the right tools in your hand, and that's what we are able to offer with respect to scalpel. I hope this video plays where we demonstrate how scalpels technology is used at different parts in the surgical supply chain. Here we are showing you two of these examples in a stellar Service Division. All it requires for somebody to pack these trays is take the tray out of the washer, place these instruments in the field of view of the camera, and the system identifies these objects instantly and validate it's against a checklist. And once the checklist is validated, the tray can be confirmed that it is ready and moved into the operating theater. The tech could be deployed also on iPads so instruments can be counted before surgery. We are able to track instrument utilization during operations by deploying it on cameras which are present inside the operating theater. If you already have them, make sure that you are have a tally before and at the end of surgery, and also measure how many instruments are actually picked up and used inside surgery. This is how we are envision the future of surgery, but I'm really excited to share something that I've never shared before. We recently completed a case study at University College London hospital, and the results are staggering. So what we have done here is we have compared human and human versus AI setup. So we have a human on the left, which is basically doing a standard manual processing of surgical instrument trace. We compared this workflow with a human powered with scalpel AI, where they are able to have an AI assistant in assisting in packing trays. We measured various key factors. Look at this, we demonstrated that when human uses AI in the process, the error rate comes to zero. Today, on an average, the first time somebody packs a tray, there is a chance of three to 4% of error rates. Second check or third check brings that error rate to less than 1%

with scalper AI, that goes to 0% altogether. Second case study, when it comes to efficiency, on an average, a technician takes four to 4.5 minutes in packing a tray. We have cut that time down to half. What this means is technicians are now able to pack twice the number of trays, meeting the unmet demand for more procedures. And what I'm really excited to share is this, within five weeks, and novice is able to perform at the same quality as an expert, five weeks. It takes six months as of today. So this is one of the biggest problems for hospitals today, with reduced staff, and, you know, the huge onboarding time. But that's not all. We demonstrated the value on of AI assistance working with them in terms of improving confidence in their ability to learn new equipment, and then in their ability to onboard news new staff quite quickly. And this is groundbreaking for various reasons, because of the urgent need to hospitals, but also for device companies. When we deployed the same technology in instrument warehouses, we demonstrated that device companies actually need less inventory in the field when they use AI to monitor how many instruments are actually used inside the operating theaters. And this is grain changing for the device companies, providing immediate and long term financial returns, not only for the device companies, but also for the hospitals by working at different parts in the surgical supply chain. And what we are able to do with scalpel is today is just the beginning. We are creating a digital footprint for every physical instrument, and this digital twin is now being added with instructions for use training guides so you can train new technicians. You can provide additional insights about every single implant that ever flows into the surgical supply chain. Automating this workflow is the next big step that we are working with various hospitals in the UK and the US. I'm really excited to share this vision with you, where scalpel is this end to end pipeline, connecting different parts in the surgical supply chain, providing critical insights to different players. Not only. On an AI guidance assistance at different points, but an AI automation assistance at different players, to the surgical supply chain. And what we are really excited about this vision is in terms of how what it takes to actually build the technology, but also our unique value proposition and our proprietary data sets. We have created a team of machine learning computer vision engineers based in the UK. We've been working for the last six years in with various universities and research bodies to build this team. We have a strategy group of organizers understanding what it takes to take this technology into healthcare. But also, we are currently raising a pre series, a funding round to secure additional team members, scale our team in the US and grow this across multiple sites. We're already across six different hospitals and stellar service divisions. We are working with the largest device distribution company in the US, and we are looking at expanding that to double our market here in the US. We are on a mission to make surgery safer. My name is Yash, and if you are interested in terms of how you can use AI to replicate the current workflows and automate them speak to me further. Thank you very much. Thank.