Jay Trautman, Invenio Imaging - Stimulated Raman Histology Laser Imaging System | LSI USA '24

Jay Trautman  0:03  
I'm Jay Trautman, as was said, co founder and CEO of Invenio, where we've developed the Neo laser imaging system for real time imaging of fresh tissue in the LR, it's been used in over 5000 clinical cases to date. And we're now adding an AI layer for immediate analysis of those images in a five center lung study, founded funded in part by the Johnson Johnson lung cancer initiative, and the NIH. So why are we working on lung cancer, it's still the leading cause of cancer related deaths, to try to reduce those rates, there's screening programs in place. So high risk patients are CT screened. That results in a million and a half nodules being identified, which is about three quarters of a million procedures every year to establish a diagnosis for those patients. But about 20% of those 750k procedures have failed biopsies. So a lot of money has been thrown at this problem, and has improved the the outcome. And along here at the bottom some some of the investments made in the area in robotic bronchoscopes and navigation. But probably the best way to improve the outcome is to examine the biopsy when it's taken with a real time technology. And so that's what's recommended by us cap that a psychologist be present in the in the procedure and provide a real time read. But it's not widely available, those people are hard to hire. And the quality is, is low, as well as it being labor intensive. So our technology is based on a laser spectroscopy and imaging method that came out of a lab at Harvard, my co founder and our CTO is CO inventor of the method. So it allows us to take fresh tissue, it's placed in our consumable and dropped into the instrument and in a couple of minutes, you get a high quality histology image with no processing of of the tissue have tissues not damaged in any way. And so it can be used after the fact for downstream analysis. Now we're adding the instant AI interpretation, which I'll tell you more about. As mentioned, it's been using over 5000 cases, we have 16 systems running around the world right now. That's actually from 12 hospitals for those hospitals bought second systems to branch out into new disciplines like the technology so much, and that's without the AI. That's due to using it with with existing human reads. So it works across all these different sample types. We began in brain because our our initial collaborator was a brain tumor surgeon. But as I've indicated, we're focused primarily on lung right now for the AI development. We really strove to get the best possible image quality when when we built this system, because that's what drives diagnosis in histology. And so the other technologies that are being used to attack this problem, have poor image quality than ours. And, I mean, that was done initially because we were worried about pathologist reading the images, which they of course can, but it also makes a huge difference in the AI. So another feature that you get because the there's no processing of the tissue is the report. reproducibility from site to site is just amazing. And so when we've built a number of the API's, we've got data coming from multiple sites, and we use relatively small data sets to complete the API's. So these are examples of API's that we built differential diagnosis. Molecular markers is particularly interest One interesting one where the markers were identified based on histology, and these diffuse glioma samples, those are both academic collaborations in Nature Medicine publications, what we're focused on is the simple detection of cancer in the lung biopsies for our first FDA clearance. And so the lung cancer initiative, Johnson Johnson is supporting the study, as well as the National Cancer Institute. Where we stand in the study is, is we're up and running at at the various sites, we spent a good deal of time with the FDA over the last year, year and a half, to get to the point that they were accepted, accepting of our endpoint. This is a de novo route, because there is no predicate for the for this technology. So we're moving fairly quickly now, through the roughly 1000 patients from these these five sites that we need to get AI clear or clearance of three different API's for the three different sample types that are taken in the study procedure. So I think the value is pretty obvious. I mean, for those 150,000 patients, if they didn't get an initial diagnosis, some of those may had to go back for another procedure. Some just were put on a waitlist, and may have actually had the condition but it wasn't known. So getting getting getting diagnosis for people immediately is a benefit both financially. And you know, for the patients, psychologically, where we go next with this is the push into real time diagnosis, enabling same day treatments and even same procedure treatments. So we've been exploring that both in lung and at length with one of our collaborators in prostate. So there's there's quite a market just in the biopsy end of the interventional pulmonology. Moving into procedures will open up even more. urology is the the area that we'll go into next with studies that will take to the FDA. There there's focus on targeted biopsies. We've already got some interesting data about optimizing ablations as well. So the team is small, even though we produce this so manufacturing is in Santa Clara AI is developed in house and we run the clinical study. So on top of the non diluted monies from Johnson Johnson, NIH, we need to add some private capital to get all the way through this study and clearance FDA of the first day I and submission of the second two AIS, late in that process. There are a number of strategics that we're working with so we'll explore commercialization, you know, with the strategic and alternatively, put together a B round for commercialization ourselves, depending on what looks like the best option. Thank you