Episode #100 | 3D Printing and 3D Tech in Pediatric Cardiology (Live Recording)

Imagine holding a child’s heart in your hands and seeing the exact path a surgeon must take before a single incision. That shift from uncertainty to clarity frames this conversation on how 3D printing, virtual reality, and advanced imaging are transforming pediatric cardiology. Our speakers show how AI-assisted segmentation, multimodality fusion, VR rehearsal, and rapid mixed-reality planning are reshaping preoperative strategy and improving communication with families.

Sarah Ptashnik of Materialise opens with the modeling perspective, walking through how CT, MRI, echo, and cath-lab 3DRA are turned into precise hollow heart models that guide baffles, conduits, and catheter routes. Nicholas Jacobson of Tangible Vet Tech brings the design and device lens, sharing how voxel modeling, hemocompatible printing, and cross-species research accelerate innovation for complex repairs. Dr. Ravi Ashwath of Baylor College of Medicine and Christus Children’s Hospital explains how advanced MRI, CT, and VR planning shorten procedure time and help teams anticipate complications in demanding congenital cases. Dr. Shafkat Anwar of UCSF Benioff Children’s Hospitals expands on fusion imaging and mixed reality for high-risk interventions, while Dr. Jenny Zablah of Children’s Hospital Colorado highlights how 3D tools improve strategy for pulmonary vein stenosis and other complex anatomies.

Together, they explore real cases in which 3D models reshaped surgical plans, revealed hazards that imaging alone missed, and enabled bench-testing of devices before entering the cath lab. The discussion covers sterilizable materials, device libraries, accuracy checks, and how VR and AR support rapid decision-making when there is no time to print.

If you are building or refining a 3D program, you will find practical guidance on quality control, when to print versus stay digital, and how to scale these tools across a health system. 3D technologies are becoming the standard for safer, smarter, and more human cardiac care.

See show notes and video highlights

Transcript

Copy Transcript Copied

Welcome, Mission, And Community Announcements

SPEAKER_05 0:00

Okay. Good morning, good morning. Okay, I just want to make sure uh you know every time I'm signing Zoom uh event nowadays the interface is completely different. I just want to make sure uh we are live. So the audience, if you can see me, please just let me know that you can actually see me in the speed first. Okay, good. Thank you. And uh glad that you know how to use the React button. Now, before people are trickling in, uh I'm just gonna say a couple words about 3D kills. Uh we have founded this company almost 10 years ago, and the goal, we have three missions. One is to educate the public about what 3D printing and 3D technologies can do in healthcare. Number two is networking. So despite the fact that we are mostly virtual, you can certainly socialize with the audience and the speakers by sharing your social media account or just shout out your name, tell us what you want. You know, it's a very casual environment right now, and you can just, you know, chat up using the chat box. And number three mission is our pitch 3D program. We help early stage startups, which means seed and series A, startups in the space with fundraising and no cost. And we connect with uh we connect you with institutional investors, which we have about 50 of them. And um, and these people are much more well-versed in the 3D technology space than many other VCs you perhaps know. So if you're interested, reach out to me, share me your uh your deck, and then we'll go from there. And in fact, Nicholas is one of the pitch 3D startup founders, uh founder several times. Um so that's our program. And another announcement I want to make is we are working on an in-person event in January on January 11th, right before the week of JP Morgan Healthcare Conference in San Francisco. The location is in San Francisco, and I will make an announcement using our newsletter and social media. And this program primarily targets investors and entrepreneurs in the 3D technology space. Okay, after all this announcement, um I want to come back to our main topic today, which is pediatric cardiology using 3D technologies. And the reason I expanded the topic from not just 3D printing, but 3D technologies is not only they're complementary, they're really kind of augment each other's ability. It augments the physicians uh and technologies in taking care of these complex uh conditions. And so the first time I interact with 3D printing was actually uh in 2011. I remember very clearly I was in RSA. That's the first time, which is radiology society uh conference. I'm a radiologist, I don't know if you know, but in 2011 is the first time I used my I touched 3D print. And it was a box of congenital heart disease mottos. And that experience kind of imprinted on me on what can what can happen is that we can extend what is a 2D image in the cyberspace to the 3D world and even in the physical world. So I really feel this is a very quintessential topic in the 3D technology space. So I'm happy I can host this. So without further

Why 3D Tech Matters In Pediatric Cardiology

SPEAKER_05 3:32

ado, I'd like to introduce our first speaker uh who also represents our sponsor, uh, who is Sarah Tashik Tashnik. And uh Sarah, why don't you take it away?

SPEAKER_06 3:44

Yeah, thank you, Jenny. Um I'm gonna share my screen here very quickly.

SPEAKER_05 3:55

Okay, I'm gonna go live on YouTube as well, just so you know, don't don't be bothered by the announcement. Okay.

SPEAKER_06 4:05

All right, and Jenny, can you confirm you can see my screen?

SPEAKER_05 4:09

Yes.

SPEAKER_06 4:10

Okay, great. All right, yeah, thank you everyone for joining here. Um my name is Sarah Tashnick. I'm an account manager at Materialize, um, and I focus with helping out the hospital markets set up 3D visualization or 3D printing initiatives or lab within their institution. I have worked with many different departments over during my time here, but one of my favorites has been the pediatric cardiology space, just because of how fascinating um the all of the different cardiac anomalies are and getting to know and learn from the pediatric cardiac

Materialise Overview And Hospital Workflows

SPEAKER_06 4:56

community. So I'm excited to be here today. But before I begin, I did want to mention, um, of course, if you have any questions throughout the presentation, feel free to put it into the chat box. There'll be a quick QA afterwards. But otherwise, feel free to scan this QR code if you even want to have a further discussion and someone from Materialize can reach out to you. This QR code will also be at the end of the presentation if you don't get to it now. So the first thing that I wanted to talk about are just some general challenges in pediatric cardiac care. So, you know, if we start off thinking about some of the clinical challenges, um, the first is of course the complexity of the conditions and and the wide variety of the anomalies. Um, and then also making sure that you're diagnosing early enough and accurately, and that can make such a drastic difference in patient prognosis. Um, and then of course, the strategy of treatment can depend on the patient's age. And then, of course, timing can be a very critical factor. But it there's not only the clinical challenges, but thinking about the emotional challenges, especially especially when it comes to pediatrics. I mean, they can be very, very hard, very difficult emotional conversations that you're having with the families, with the parents, with the patient themselves. Um, and then of course, the emotional burden or psychological burden that that can carry, um as well as any prolonged support even afterwards, any post-op surgical support. Um so it's important to keep these in mind and also kind of figure out how we can, you know, I guess fill fill in the gaps to these challenges and and make it easier during these processes. I think it's also interesting to think about, you know, if we're if we're talking about patient-specific models and how that can kind of um shorten the gap of those challenges, who it's exactly making an impact on. So of course, there's the patient, you know, once you have an actual 3D or patient-specific model of their anatomy, it can help them feel like they're making more informed consent um or informed decisions on what they want their care to be, as well as just overall improved patient satisfaction. And then on the clinician side of things, when they're going into the operating room, it can help greatly increase their confidence and there's less stress. And you know, ideally there is an improved patient outcome. And also when you're going into the operating room with greater confidence, sometimes that can also shorten the OR time and therefore the OR costs every minute costs a lot of money. Um, and by that it has a positive impact on the hospital itself by saving some money and by also not just due to the shortened OR times, but by reducing any instrumentation waste, for example, if you can make that decision ahead of time rather than intra-op. Um, so before I get into an example of an actual workflow and what that looks like, what the models look like, how it looks like to get those, I wanted to just go through some examples of some studies that have been done and the conclusions that have been made from those. If you have any questions or want me to send you any of the references to these, just let me know afterwards and I can do that. But essentially I just want to highlight more on the value that 3D visualization and 3D modeling has added into the clinical space. But the first is, of course, offering a better surgical outcome and minimize complications on the way. So 71% of clinicals found that the virtual 3D model has improved their ability to manage potential post-op problems in surgery for congenital heart diseases. And there's been a 41% of shorter stays of dwarf patients in the intensive care units when using 3D printing. Having a 3D patient-specific 3D model can also help to plan critical clinical decisions more accurately. So there is an 8.5

Clinical Impact Evidence And Outcomes

SPEAKER_06 9:49

out of 10 score given by surgeons involved in congenital heart operations for usefulness of 3D models in pre-op planning. And 66.7% of interviewed cardiac surgeons, cardiologists, and radiologists find patient-specific 3D models helpful in pre-op planning. It can also help to realize that you can treat patients who may have been previously untreatable or thought that they were previously untreatable. So there was a study done where after assessing 15 different patient CT scans, only 13 cases were consistent with real treatable options and two others were not. But with the addition of seeing an added perspective of 3D, they realized afterwards all of the 15 cases were consistent with real treatment options. And as I mentioned before, you know, having a model, especially from the the patient or the patient's family perspective, it allows them to really feel like they can give true informed consent. So 9.5 out of 10 average rating by parents for the usefulness of a 3D model when explaining the congenital heart disease of their child. And we can also even think about how it can improve communication between the surgical teams. Um so 92% of clinicians are moderately to highly satisfied with the ability of 3D models to enhance the understanding of the anatomy of patients' congenital heart defects. So one of our um one of our groups that are using 3D modeling for almost every single one of their pediatric cases, uh, Dr. Hogginson from Boston Children's, um, we did an interview with him and we were able to get some insight into his perspective as a pediatric cardiac surgeon and his use. Um and later on, I will show you a clip at the very end of this presentation on so you can see visually how exactly he's using the 3D models uh during his pre-op discussions and as a guide in the OR during surgery. But now we'll get into a little bit so you can actually see what a workflow might look like in the 3D in a 3D software.

From DICOM To Hollow Heart Models

SPEAKER_06 12:26

Um, you know, the question is how exactly do you go from having a DICOM to a hollowed model that you can use for your planning? Um in this case, I'm just gonna run through, you know, having a contrast enhanced CT scan of a dwarf case and how you can get that model, how you can get it hollowed and use it, you know, in the best way for the planning. So obviously, our whole world is shifting towards automation and AI. Um, here I have an example of a an AI-based CT heart tool. So you can see that, you know, and I'll show a video afterwards. After you click apply, it's essentially automatically outputting the different chambers and the different great vessels of the heart. And then, you know, if you need it, if you have a more complex case and you need to apply more more manual feed points, for example, that is an option as well. But I'll play this video, it's a super quick 10-second video, but you can see the the output that each of the chambers and the vessels are separated, and you can also see that there's the overlay on the CT scan, which can help you decide, you know, if you need to do any further editing afterwards. So let's just say we're gonna move forward with the scenario of having the whole blood pool instead of each of the chambers um and vessels separated. What we can do after that is hollow that model. Decide where exactly you want to make cuts so you can view inside that model for the intercardiac anatomy. And here you have your completed hollowed model. So the question is what can you do downstream after you have that hollowed model? There you can add design components, so create a patient-specific baffle. There's also the ability to, for example, um use for the font hand procedure to create patient-specific conduits. Also take different measurements um along, you know, if you were gonna do some sort of implant or or device implementation. So here we have like various measurements that you can take along that hollowed model as well as mapping out maybe the the catheter pathway through the fossa, for example. And then, you know, outside of even just trying to figure out the best like design uh device fit, um, you can also do many different measurements, centerline measurements, for example, for more research purposes. So the other question is once you actually have that completed model, what is the best way to communicate it to the rest of the surgical team for approval or for review? So you can upload them to a case storage or a case sharing platform. So you can see an example of all these different cases that are already listed. And if we click one and open it up, you can see the various case details listed, as well as the ability to add any other files that are relevant for review for that case. You can share that case to any various stakeholders on the team and then open up that case to actually view it. So once you actually open that case, you can see that you have the completed model information, you have the the DICOM images to review, the contouring over that model, you can take various measurements or annotations.

Sharing, VR Options, And OR Use

SPEAKER_06 16:33

You can also take that into a virtual reality space or augmented reality space if you prefer to go that pathway rather than printing per se. Um but here we'll wrap up. I know we're kind of running short on time here, but I mentioned that I wanted to bring it back to a video example of how Boston children might actually visualize their completed models during their review process and during the actual OR. So they will have this, you know, they'll set up a call with their whole team and kind of go through go through the review process, but also have this open during their time in the OR. Um you can see in this video, you can toggle on the different chambers that are maybe are necessary to review, or you can also set default views that are relevant, maybe the the most common perspectives to review. Um, and you can see that you can add different measurements, or you can clip through the cross-sectional area to kind of follow your way through the anatomy. Um, whatever way is going to be the most relevant for you while you're in the OR. So the rest of this is kind of going through the different viewpoints. Um And you can also see, you know, any any design component. So if you added a patient-specific baffle design, you can also bring that in and overlay it, for example, during the different different phases of the cardiac cycle. So with that, if there are any questions, if there's time for any questions, um, feel free to put those in the chat. Otherwise, I am happy to to reach out to you directly and feel free to scan the QR code afterwards as well.

Q&A: When To Print Versus View

SPEAKER_06 18:39

Thank you.

SPEAKER_05 18:40

Uh yes, actually, Sarah, feel free to put your email and stuff in the chat box as well. So people can just directly email you. Yep. Um, also love to have the reference uh links if you have it, and um, I can share it with the our post-event email. Uh, you know, all those numbers that you shared. Quite interesting in terms of the effectiveness of the 3D technology. I think people want to know how these outcome results were generated.

unknown 19:07

Okay.

SPEAKER_05 19:07

Um, so I don't see any questions in the QA box uh since we're kind of tight today in terms of schedule. I'm gonna move on to our next speaker, if that's okay with everybody. And also, oh, actually, you know what? I just saw okay. Okay, guys, put the uh question in the QA box um uh if you can, because um, as we go, there are gonna be more questions. And I'm just gonna lose there's one question. So when do modos generally get printed versus just viewed? That's a good question.

SPEAKER_06 19:36

Yeah, so I think it depends on a few different factors. The first one could be, you know, what resources you have. I mean, do you already have the printer and the material in place to print? Um also how much time do you have? Do you have the time to wait for the model to be printed to do any post-op care for the model? Or do you just need to quickly view the model? Um, and also maybe how many people need to be involved in reviewing the model, for example. I mean, when you print it, how easy is it gonna be to take that printed model and pass it, you know, along to get everyone in a room at that same time to view that model together? So I think it just it depends on maybe that specific scenario, timing, resources, and how easy it is to get everyone together.

SPEAKER_05 20:26

Thank you. And maybe later on, I'm gonna get some um feedback from the clinicians is the tactile component that was added by 3D print itself. Like I don't know how much more value it adds sometimes. So that that is a question on my end, since I never get to touch any patient. Um all right, okay, that's a good question. Let's

Apples, Variability, And Design Gaps

SPEAKER_05 20:46

move on to our next speaker, uh, is uh Nicholas Jacobson. And he has very extensive knowledge in 3D technology in general. And I think he used to work a lot with uh University of Colorado or University. Yeah, okay. Sounds good. Uh and he and he works with uh Dr. Jenny Zabla, who will be joining us later uh at this um in this webinar.

SPEAKER_03 21:13

Yeah, um you can see my screen, you can hear me, everything's good.

SPEAKER_05 21:17

Yeah.

SPEAKER_03 21:18

Just give me a thumbs up. Cool. Yeah, my name is Nicholas Jacobson. Um I'm the co-founder and CTO of Tangible Industries. Um, I worked uh and ran a biomedical research lab for seven and a half years at the Anschweiz Medical Campus and worked really closely with Dr. Jenny Zabla, who'll be talking today. It's very cool to talk with her. I don't get to do it that often. Um, but Jenny and I represent two sides of the same coin. When I first started getting into, you know, my background is as an architect, I know it's very bizarre, but I'll kind of explain that. When I got into this world, um, I saw that there was a big gap between engineers and clinicians. Engineers had such cool tools to be able to do uh amazing things, but they didn't have the clinical experience to be able to put it into action and really have an impact. And then I saw clinicians who would have a 3D printer in their office, and they were able to do very, very simple things, but not take full advantage of all of the technology that was out there. And so Jenny and I really represented, you know, one of my first forays into the medical world of embedding myself into the clinic, having Jenny come over to the lab and um work very closely to make sure that all of our work was fully integrated and had the most impact. Um try to explain that today. I'm gonna talk about a lot of things, um, but uh, and I'm gonna talk about apples. Uh, that's one of the big ones. My background is as a farmer, and I was just back in Wisconsin and I was thinking about um the big difference between interventional or between adult cardiology and pediatric cardiology. And in Wisconsin, everything is a factory farm, everything is uniform, and it makes things very easy. You can make tools that are replicable, everything that that comes through fits into um you know the current manufacturing paradigm where you can have mass manufactured parts. In pediatrics, though, everything is like an apple. I don't know if you know this, but there's 7,500 apples in the world. And every day in pediatric cardiology, a different kind of apple comes through. Nothing is uniform. And um, it requires having to use these off-the-shelf tools that are made for adults and to be able to address this variation. You never know what you're going to get. And so this is an area that is ripe for um biomedical innovation. Really, it's a, you know, our limitation in being able to treat all of these patients is really a limitation in the manufacturing and design process. Now, when I first started uh doing this work, um, we started doing very simple models like this for uh cardiology and surgery, and we had wonderful results. Um, you know, I can go on and on about the wonderful results that we had. Um, I was able to embed myself into the operating room and learn a ton about what was going on on a daily basis, and our models really hit the mark. We did a lot of this kind of work. Um, and at the same time, you know, we we've been having this conversation about reimbursement and what's going to happen with reimbursement for these models. Um, in pediatric cardiology, it's a really challenging case to try to get reimbursement because most reimbursement works on trying to reduce operating time. But the fastest surgeons are not the best surgeons, and speed is not really the goal in pediatric cardiology.

Printing Blood Flow And Fetal 4D MRI

SPEAKER_03 24:55

It's being able to successfully do the surgery, understanding what you're going to get ahead of time. Sometimes it's faster, sometimes it's slower, but it's a really challenging case, although we know that it's highly beneficial. Now, I'm a little bit of a weirdo in terms of my background. As an architect, I started to see that there were issues in um in cardiology that went far beyond just modeling. And in architecture, these were problems that we had solved a long time ago. And so I started to look at ways that the technology that I had could start to solve, you know, and things that we had solved in architecture that were very easy, could help to solve some of the issues in pediatric cardiology. And I found out that, you know, one day I walked in and there was a gentleman named Alex Barker who runs the advanced imaging lab, and he was figuring out how to get um four-dimensional flow, um, the actual blood flow and velocity in the heart. And I saw this uh from MRI, and I saw this on the screen, and I thought, well, you know, first and foremost, I'm an artist, so I thought this was really beautiful, and I wanted to figure out how to print this. We've got this weirdo printing technology called voxel printing technology that allows us to capture things in ultra high resolution right from um imaging. And so we were able over six months work back and forth and figured out how to be able to take this work, uh, this data, and transform it into a physical model. This work led to some early cases of single ventricle outlet where you really want to understand what kind of intervention you're going to make, what the impact is going to have, and what the long-term ramifications of your intervention is going to have on the heart. More recently, this work has led to um fetal MRI work. Um, Dr. Barker, and this image is really um, sorry, bad on the screen right now, um, but a few email mails and just some better ones. Um, Dr. Barker has figured out how to uh get 40 flow uh in utero, to be able to do surgery in utero. And so it's been wonderful to be part of the team to try to figure this out. We've been taking um this 40 flow and um printing it in all kinds of different forms to both study the science of it, but also to study the uh the ability to do the surgery. And here's a few more images. I just think that these are really pretty. And they're tiny, you know, they're they're they're three, four inches tall. Um it's pretty amazing stuff. Um I also started working on understanding, um, you know, we work a lot on the pulmonary, um, on the pulmonary valve. It's one of the most common valves to work on for kids. And we started understanding compliance of the tissue. Wanted to really understand morphology. Every kid that comes in is so different. So if we could figure out, you know, what some common themes are, we can maybe start to better understand the types of interventions to make and maybe even understand the types of implants to put inside of that. Because the one size fits all really excludes the vast majority of kids. And the biggest need in pediatric cardiology is for patient-specific implants. We started to do a lot of work on um the materials. Um, you know, moving into the implant space means having a hemocompatible material, and that's been by far the biggest limitation over the years. The reprinting's been great, but the vast majority of these materials are highly toxic and you wouldn't want to put into your bloodstream. And so we worked really hard with a wonderful polymer chemist to develop hemocompatible material and have been doing a lot of these studies um ourselves on campus. Then I started to push really hard. And over the years, um, you know, uh gave a lot of talks about all of the work that we

Materials, Regulation, And Disruption

SPEAKER_03 28:44

were doing, started to do some clinical trials, investigator-led clinical trials in other areas in 3D printing. It was all over the map, from pediatric cardiology to um to epilepsy um to adult work. And um basically I got on the radar of the regulatory team at our campus, and they took a look at what I was doing and got really nervous in the way that 3D printing right now is vast, is is basically unregulated. The FDA hasn't made a lot of decisions and put out a lot of guidance on what's going to happen. The materials, the printing process, computational design, what happens with data, the fact that I'm an architect in the operating room, all of these things that I was doing, it became really concerning to the regulatory team, to the insurance team, because it's all so new. It's wonderful to be able to do pre-surgical planning models. But once you start to go beyond that, we enter into a world that is brand new and very complicated and would require an enormous regulatory team. Um, so I got labeled the disruptor. And I realized that the pathway to be able to move forward with any of this work. Was not just going to be time consuming. It was going to take maybe 10, 15 years to be able to move things forward. But it would be very expensive. And so I started to look at other ways of moving forward. And I realized that, you know, I had this wonderful, you know, band of things that I was working on, apples and babies that were all one-off and trying to bring, you know, the address this vast majority of diversity and complexity. But there were some other people, you know, I discovered about a year and a half ago that I could join and could join our little party here. And those are dogs. Up at CSU, I'm here in Denver. CSU is one of the best veterinary hospitals in the world and have been working on interventional cardiology and surgery on canine hearts for a long time and really leading the way. They have incredible resources, and people are flying from all over the world to be able to get their dogs treated. And a dog's heart and a child's heart is very similar. They both have an incredible amount of complexity and variation. There's a lot of uh the sizes are pretty much the same. Dog's heart is slightly different, but to the untrained eye, it looks very much the same. Um, so there's a lot of things that we can do and we can study um dogs. And in fact, the advent of open heart surgery um invented by Walt Lilihai, um, he first studied dogs and ended up sacrificing many, many hundreds of thousands of dogs to be able to create um the heart and lung machine bypass and be able to bring that for the first open heart surgery that was done on pediatrics, which is a VSD repair. Now, the cool thing about interventional, about about um the veterinary world is that they have a wealth of resources. People love their dogs, and the access to all of this resources is enormous. CSU has one of the world's largest libraries of advanced imaging, from TEE imaging to MRI and CT. Um, and they've been doing a lot of models for a long time. So I was able to continue my work and actually get better imaging to be able to move all of this research forward. Here's a um cinematic rendering of the heart, and we're fusing together TEE data to be able to understand and get full view of the leaflets and the cardiac cycle together. Really exciting stuff. And again, tremendous amount of for a long time they had one of the best CT scanners in the state, and they have one of the best operating rooms in the state. Um, it's very exciting. And we're also able to move forward this idea of the valves. One of the most common problems in dogs is the mitral valve. Um, there's 8,000 or 8 million dogs in the United States that have this disease, and and they've been pioneering this technique to be able to go in, they go in through the apex of the heart to repair um uh the mitral valve. And um, we've long in pediatric cardiology and and as you'll hear from Dr. Zabla, um, have been trying to push on interventional cardiology, a um vascular approach. And so it's wonderful because this is one of the most common issues in dogs, and we can take the same tools that we use for

Veterinary Pathway For Faster Innovation

SPEAKER_03 33:32

kids to be able to address um things like the mitral valve and take a procedure like mitroclip, which is one of the most uh popular surgical interventions in humans ever, and apply this um to the canine world, all while starting to integrate all of this 3D work and hopefully bring this back into pediatrics. And so we've been working now for quite a while to be able to do an intervascular approach using pediatric tools and um our own custom um 3D printed clip, which I'll explain here in a second, to be able to um address not only dogs, but then within the first, you know, we think within the first two to three years, be able to gather thousands of cases that we can then take back and move forward all of our work in the pediatric world. Here's a little view of our of our clip. Um in the there is a very um enthusiastic patient population. People are paying um cash and flying from all over the world looking for anything that can be done to help their dogs. And so the regulatory environment in the canine world is so minimal that it allows for rapid innovation. Instead of this 10 to 15 year cycle for pediatrics, we believe that we can do our first interventional procedure within the next year and within two to three years get thousands of cases, efficacy cases, compassionate cases, where we're not putting down dogs, but we're doing this compassionately and iterating and innovating all along the way to address a large population. And then ultimately bring this back to the pediatric world. It seems like the only way that I know of to be able to start to bring 3D printing and all of this technology back to the human world as quickly as we can. Um so that's that that summarizes where we're at right now. Um I never expected to get here. 3D printing has been quite a journey. Um and uh stay tuned uh for for where we go after this.

SPEAKER_05 35:46

Thank you, Nicholas. I'm always on this journey with you, and every time I see your presentation, I guess um who's a nas. And uh I was like, oh wait, this is not a veterinary webinar. But then okay, thank you for coming back. Um and also please remind me to introduce you was a guy I recently met who actually will uh who has a startup focusing on clinical trials and docs. And uh I think I think you guys are gonna be a good fit to talk to uh each other. Um, so let's see. Do we have any questions? Uh let's see. Okay, not yet. Again, if you have any questions, feel free to put it in the QA box so that I can address. And if not, I am gonna move on to our next speaker. Um, next speaker is Dr. Ravi Ashwath, um, who really graciously offered his time to teach us what he's working on. He's currently professor of pediatrics at Baylor College of Medicine and a division chief of pediatric cardiology at Christus Children's Hospital. And I can, even though I haven't met Dr. Ashwath, I can tell that he is a very passionate educator and also innovator. And um he has many uh innovative approaches in pediatric cardiology. I'll let you take it away.

SPEAKER_01 37:07

Thank you very much, uh Dr. Shannon and uh all the people and uh thanks for inviting me. I'm gonna skip a little fast, and if people have questions, they can obviously answer. I just wanted to show all the things that we are doing here. No financial disclosures. And the things that I might want to just touch by is role of 3D printing, role of virtual reality, virtual surgical options, virtual interventional options. We also had this thing about a QR code, how to incorporate into patient charts so everybody can access the models at a higher level, large-scale, patient and family education, and also closer to the cardiac but non-cardiac uses, and several others completely non-cardiac, but I'm going to stick to cardiac today. So, examples of 3D models going back when, like Dr. Chen said, 2011, she got introduced to 3D printing. Similar to that, around when I came back from Boston Children's, I think, in Cleveland. This is when I started getting in. So somewhere around that area, and long back, this is how we used to print in the uh solid materials, just to educate patients. So that's how we started, just to say how we can print in different materials, make it colorful for teaching and understand. So that's how it started way back when. And uh, we did not wait too long to introduce virtual reality, so I'm just gonna play this quick uh video. So, this is to teach people how overlay cuts of echoes, axial imaging works. So we could have done it in virtual reality, and this is just to show uh this one. So we'll switch to some case presentations of how

Case-Driven Pediatric Planning Examples

SPEAKER_01 39:52

we have used 3D printing as the years have gone by. Some of them are cases that we just wanted surgeons to see if they could do something different than what we had thought initially. This is uh eight-month-old female basically was presented to us pulmonary hypertension and pulmonary vein stenosis as the diagnosis and a CT scan that we did did find some amount of left upper pulmonary vein stenosis, but that was not all. It was a huge inferior sinuspinosis, uhteral septal defect, which we thought was a new finding since we were not told about that, and also quite anomalous pulmonary veins, with the right side being totally abnormal and there's some amount of aortic arch hyperplasia. So I think the initial plan was get them to get in left upper pulmonary vein, balloon angioplasty, and then surgical repair of the PAPVR and the inferior sinus venosus AST, but the complex anatomy of the inferior sinus venosus with two veins in a different location, and then to baffle them back into the left atrium is a task which is going to be probably met with some obstruction of the baffles. So we wanted to understand more. So we did the CT scan and the rendering of that. I think the things that I want to point is this is the IVC coming posterior view, and this is the inferior sinus venosus AST right atrium, upper pulmonary vein draining directly to the right atrium, right middle, right lower to the junction of this. So to baffle this to this side, close the AST and baffle this to this, you know, it's not this space is not a whole lot of real estate. So looking at that in this cut view, right upper pulmonary vein, this one, that's your left atrium. So it's going to be a complex one. So we printed that to exactly get a sense. I have labeled everything here, right? Superavina cava, right atrium. Most of the things is for the surgeon to see how you could do all the baffles with these veins going back to the left atrium, three veins. And we found out that I think baffling of this and keeping the IVC where it is was going to be a task that was going to be met with IVC obstruction. So we suggested to the surgeon to take a different approach, and we went up with the reimplantation of the IVC into the right atrium very separately, not trying to do a big baffles, and the kid did very well. Um so next is uh ten-year-old who has been prepared for transposition of the great arteries with uh atrial switch, sorry, uh arterial switch operation and the licompt maneuver, which is the usual operation. The kid then had heart block, so went back for a second stronatomy and then had to get the pacemaker. So now the pulmonary valve, which is the uh used to be the aortic valve, is having quite a bit of regurgitation, and that is the field that most of us are working in terms of getting the pulmonary valve placed. So she was considered for pulmonary valve replacement, and she's 10 years old with a third stronatomy and was referred for a second opinion, and a CT scan was performed. And in the CT scan, the things that we need to notice here is this is the right ventricle, and this is the main pulmonary artery and the branch pulmonary arteries. As we know, a melody valve, if it has to go here, has to sit here and obviously make sure it doesn't protrude into the branch pulmonary arteries. As you see here, the angulation of the R V OT going into the MPA, main pulmonary artery, and the branch pulmonary arteries would not make it very easy to deliver a valve here unless we understood it a little better whether the valve was going to stay there with a landing zone and also not protrude into the branch pulmonary arteries and cause. So we modeled that and then we took the model into the cat lab, just like how we would take a patient and then actually did introduce a valve in the stent and made sure that it stays there and does not occlude the branch pulmonary arteries. So once that was proved, you know, taking the basically the model and working on it, showing that it could it could actually hold the stent. So we took the kid then to the cat lab and then put a 22 millimeter melody valve, successfully implanted, and obviously we all know that if it was a kid that was going to get a third strenotomy and a valve, it was going to be at least two weeks, ten days to two weeks of stay. Patient was discharged on uh day number one after the operation or the procedure. You know, as uh Sarah was mentioning here, you know, the operative time and also the length of stay has been quite uh remarkable in our experience because we know a lot beforehand and we know what could go wrong. Those are the two things, at least in the surgical field, I tell the surgeons that not because uh you cannot do this stuff, is how easy it will be for you to do it and what are the things that you can encounter when you go there. So now I think uh switch to the adults. Uh we had uh this is a paper that we published, a transcatheter valve implantation in a severely regurgitated apocalypse aortic conduit. It is very, very rare. So I presented this because we had basically an aortic stenosis in the patient, severe when she was uh in the younger part of her age. So she got from the LV apex to the descending aorta conduit, which was obviously sending blood to the descending aorta, and the ascending aorta could get some blood through the aortic stenosis. So obviously had to have some replacements a few times, and now there was continuously regurgitated, and she had problems with fatigue and everything like that. So we said instead of taking and revising the conduits, can we actually go put our valves, melody valves, or into the conduit here? And after all, the date of uh is it severe regurgitation? Is that the problem? After all, the CT scan. We did uh construct the images. Actually, we printed this was a huge print because the patient was an adult patient and we actually practiced on that. And then once we did that, we went in actually by cath route and actually put in the valve right in this area. So this was one of the things that we also could avoid a surgery and also an innovative technique of not putting wells the usual place, but this is a very strange case. Then we moved on to a 67-year-old grandma who

VR Baffle Design And Team Collaboration

SPEAKER_01 46:23

was short of breath, already had CAD, coronary artery disease, and had PCI and cervical cancer, diabetes, hypertension, a lot of things. So the surgeon did not want to go ahead and repair the superior sinus vinosis ASD, which I think is the standard of care for the most institutions. I know uh some institutions are working on getting it closed through the stuff. So we thought in Iowa this was the first time we had thought Dr. Aldous was a cat. So we obviously wanted to show where the veins are. The two things that you want to make sure is the veins get back to the left atrium, and the SVC will drain into the right atrium, and the AST gets closed. So here we are showing the AST, and here should we are showing the anomalous pulmonary veins. And obviously, the transcatheter closure is basically delivering a covered stent to serve two purposes: that is, covering the ASD, not occluding the SVC flow, and then not obstructing the pulmonary veins that go back to the left atrium because the pulmonary veins are right behind this one. So we printed the model, we did the same thing as delivering a covered stent, showing that it can cover that, and also it would not obstruct the uh pulmonary veins in the back. So I'm just going to show, and in the interest of time, I think we'll move on. So now, obviously, we also started using some VR as things go by, uh, and this is a short video of the same patient to show that that's your SVC iota pulmonary artery, that's your AST right there, and then the pulmonary veins are coming back here. So you want to put a stent which is there, deliver that into the SVC right there, put there, this should be a covered stent to make sure that the blood flow goes there. Now it goes right into this, it covers the atrium because there's no flow going through the AST and the pulmonary veins are not obstructed. Once we approved all that stuff using TEE, we could deliver it, and the patient went home quite successfully. And then virtual procedure was done, and then you can see post-op a beautiful stain, just like the one that we had showed in the VR, pretty much resembles like that, closing the AST and the pulmonary veins were not obstructed. And we also did a CT scan a little bit later, uh a month later, to show the pulmonary veins were wide open. So moving on to the virtual reality in complex cases, uh, we also worked with Realize uh Medical to say vision to reality of uh baffel planning. You know, not only in the double outlet right ventricles that we hear so often in pediatric cardiology, we also, it was not just the subiotic VST, which is the easier component, which is also the subpulmonary VST, where the baffles have to be bigger, the baffles can take up the right ventricular space. What is the distance from the tricuspid valve? What is the distance from the VST? So there's a lot to understand here. It's not about just creating baffles. So in this, these all are the annulus of the valves, four valves, and one of them is the um VST in terms of the borders of the VST. Here in concept, what we wanted to uh show people is that creating the baffle includes a good path, not obstructing any of the ones on the LVOT or on the right side, and the amount of volume it takes up on the right ventricle before the patch and after the patch, does it suffice for us to go for a biventricular repair? And all of this can be done. And uh this also can be done as not only one person, but we can also collaborate and do this in complex cases. So, a small video, I might have to run this a little bit. So basically, we are gonna do this segmentation, we're gonna get the interior, and we have to show the analysts of all the things, the uh VSDs, how we are gonna mark the VSTs, and all the uh measurements that are needed for us to do the procedure so the surgeon will not get into trouble in terms of putting a patch. We get that patch, and then we also look at both outflow tracts, volumes on both sides pre and post, to make sure that the amount, even though it's reduced, does it still confer to the 25 mL to 30 ml per meter squared? So same thing with the two viewers now. Then switching gear to one of the augmented reality in pediatric cardiology, you know, Dr. Sha Anwar has joined and he has a lot of work done. For us, I think the goal here was we did printing, and then it's only a few people that can hold the model and feel it. Then we went to the VR where we

AR With QR Codes For Bedside Teaching

SPEAKER_01 51:06

don't have to print all the models. We could probably watch in it, complex cases, take them to the uh uh VR environment, sorry, to the 3D printing and then work on it, either virtually or the surgeons can work. But then how about all the people that we talk to, the parents, the uh residents, the medical students, the whole ICU team? Can we show them the model when we teach them? So we decided to go on this kind of thing is basically we do a 3D STL segments, make it in such a way that we can actually store it and then point it with a QR code for every lesion that we have. And during the rounds, we will actually share the QR code. I mean, obviously it's a password protected, and then if you scan this, if people actually on the call can actually scan, and if it is still valid, you know, this is a subscription thing, that you can actually get the model of the ones that we are talking about. So this can be either incorporated to the bedside or can go into the patient charts wherever they go. People hopefully can understand the 3D models better. The way to work on this workflow is we do the segmentation through materialize and the three-matic to help, and then we use Blender for some of the color assignment. The uh QR code access cannot be huge files, so we have to do some decimation and scaling down, and we can actually label them all well in terms of everything for teaching, so they don't even have to uh somebody there to help them. And the usage of this on their iPhone is not something that anybody has to teach them because the zooming, the uh you know, making it smaller and bigger, turning it is all pretty intuitive for people. So this is going from there to there, and it's a small video if uh I play. So if you scan this in your phone, basically you scan that, and you will get things to say, open this in this, and you can open it here. And when you open, and this is an iPhone and the Android, sometimes it's a little difficult. You have an augmented reality mode. You can go and lay on the patient's chest and show parents, or you can go to the object mode, which then you can do the zooming, you can turn around, and you can actually learn everything. This is a total veins, so you could see the vertical vein going to the nominate, and you can also put the trachea and you can name this. I think we found it very helpful for uh people who can understand and learn and also show it to the parents in a very easy fashion. So I thought uh this was uh very helpful for us. Then uh, you know, in terms of the surgical planning and uh virtual surgery planning, some of the uh few papers that we wrote basically is looking at not just uh printing and then you know working on sutures, which uh people did, but we wanted to see what's the accuracy of the sutures. If two surgeons do patch one, patch two, you know, what are the things that they take into account? Can we minimize that? So this is the novel application, it was novel at the time, but application of geometric morphometrics to cardiac surgical simulation. In short, I'm not an engineer, uh, but in short, what we are trying to do here is if there is a lesion, in this case, it was uh PAP VR with an AST, in the surgeon and the resident would actually do give the two models to them. They would do patch one, patch two, and then we will compare them, you know, directly side by side like that, but that comparison is not really very helpful. So we use what is called as geometric morphometrics, which is basically a science that actually compares it point to point and tells us what exactly is the difference, and we can actually make patches based on that, and then probably can use it better. This not only does in 2D, it actually does it in 3D because on the patches are not always very well in 2D. You need some redundance and you need to do that. So I think from this, what we were trying to help is the formation of the patch can be a little bit more ideal. If I'm running out of time, please let me know. Um and this is the same thing that we use for uh interatrial rerouting. And one of the things that we did was actually went to the surgical specimen here, and then our specimen, and then we compared our uh prints here to the things and we labeled it to show that it was exactly how if you look at this specimen here interoperatively versus our printed model, the accuracy was uh, you know, really remarkable for the surgeons to come back and actually feel like yes, what we do is really helping them. And that's basically what you want to show. And the other uh little project that we did was give the two surgeons five or six complex models, make their own patch, and we designed a custom design patch based on the geometric morphometrics, and asked them to see if this suited themselves versus if this suit ours. What would the difference be? At the end of the study, this was a qualitative study, not in a very objective study. And the surgeons themselves felt that the custom design patch was actually a little bit better because there was not much of redundancy and there was not much of um you know wastage on the sites, and the patch

Surgical Simulation And Patch Accuracy

SPEAKER_01 56:27

would look like that versus ours would look nice and taut, and we could uh you know give them a little better things. So this is just uh the same thing in terms of using some of these warping graphs and the you know, using the centroid methods to help them do all that stuff. And then, you know, when people say, how does it help with the teaching, you know, and the parents and the patients, so uh one of the medical students did a little bit of a uh study with Likert Scotts to say what is the impact of 3D printed, and this was not comparing to 2D or sketch diagrams, which we had done before, and proved that 3D printed was quite good, but we were trying to figure out if 3D printed versus 3D virtual, which could be either on the computer screen or on their iPhones, what difference did it make? I think people came back to say that actually 3D virtual pretty much they understood very good as much as in 3D printed for the most part, but better than the 2D or the line diagrams or the diagrams that we put on the paper for them to understand that one. So all four groups believed that 3D models would be beneficial, but interestingly indicated that they preferred patient-specific model as opposed to a generic model also. So the two things that they came out was they like their own heart disease in the exact same condition, because as uh Nicholas was pointing out, about 7,500 types of apples, we cannot just take a VST and show a patient when it's a complex anatomy. We want to exactly show what the patient has. And then um, you know, in terms of other things that we do, congenital portosystemic shint, we have helped by showing, you know, where exactly that is and how they can go coil it, if at all they can. And they tried this by uh intervention, the pressures were too high, so it had to be done in two settings, and the surgeons took it, and uh, at least they have the landmark of where they have to go and ligate stuff. So that was one of the other things. ENT was a big thing that we did in the University of Iowa because it was a big center for tracheal reconstructions, subglottic stenosis. We had helped the patient, and this was a patient was a hard time to uh decannulate, but they had several questions. We printed it, we did the internal diameters and all the stuff to say that if they decannulate and actually do a graft or patch it up, the patient probably can get uh come out of the tracheostomy, and it was a collaborative effort, and then we printed these things to show all the internal diameters, radius, and stuff. The most important that we are trying to use this for tracheal was whether they should go on a rib graft or they can go with end-to-end anastomosis, and with the measurements we would go with one of the two routes. In this, we decided it shall actually go and actually decannulate. And the patient actually was decannelated and postoperative number two, and then third day the patient went home. Last one at tracheal stenosis, same thing, same concept. You can see that this patient did several operations, and this is the way the tracheal stenosis was, and also the blood vessels that were there when they had to go on repair was not something that they could uh avoid. So I think the two things that we helped here was that the patient actually could have end to end anastomosis, and also they might need a cardiac surgeon to help them because of the blood vessels and the complications that might occur if they just did it by themselves. So basically, a strenotomy and cardiopulmonary bypass had to happen, and then they could do an end to end anastomy. Astomosis without doing a rib drafting, and that was the initial point. So, in conclusion, I think 3D printing, you know, as we have all been doing for a while, is exciting and promising technology, which is feasible in our field. Models can be very effective and it can reduce procedure times. Now, as been shown by articles, increase the safety and quality, which I vouch for from all the patients that we have treated. And also being in educational institutions, I think cardiology is very hard to understand. I would not understand the echo and stuff in my fellowship until I figured out 3D and CT and MRI. So I don't think we have to have fellows and uh residents struggle like I did. So I think it's okay for them to go to the next level and learn. And it can be applied not only for cardiology and other fields. And of course, the goal is personalized medicine and regenerating organs like people are working on. And I'm happy to take questions, but um very glad to be here and thank the opportunity.

SPEAKER_05 1:00:59

Thank you so much. This is such a great presentation. I totally agree that residents may not need to struggle as hard as we did in the past. Um, just like you don't need to ride a horse because now we have cars. Um so we have one we have several questions in the audience because due to time constraint, I'm just gonna address a real quick one, uh, which is from Brian.

SPEAKER_01 1:01:21

And he asks, Do you is there any privacy con concerns uh using the air code uh that's we have not we have not made it completely into the uh epic arena at this point because then um already when we started the conversation, people just put a card watch right away saying that it has to go through several layers. Right now, the only way we have used it is for education, you know, during right now, not putting it anywhere into the charts and everything.

SPEAKER_05 1:01:50

So that does look amazing. Um I I honestly think all the patients would love to have one. Um so thank you very much for sharing. And okay, maybe one more question uh is in terms of quality control is how what is your you know, quick brief review of how do you guys decide if the model is accurate enough?

SPEAKER_01 1:02:11

I think uh Chafka Thanwar can answer this because he does a little bit more for me. Uh for me, it has been printing the models, sitting with the surgeon, going through the echo CT and the models together and explain to them, and then come back and ask the surgeons what have you found different. That is the that's been the quality control for me. And I feel like that probably is one of the best ways because the surgeon comes back and says completely no or yes. And we have had very good success in them coming back and saying that it was as billed as dubbed and as you showed it. So I think that's that's been our one, and not a specific way, but I think a practical way.

SPEAKER_05 1:02:52

Thank you so much. Okay, so since we're already mentioned our next speaker, uh without further ado, I'd like to introduce Dr. Shafkat Enward, who has been with us many times before. Uh, he is at UCSF, so right next door. So uh Dr. Enwart, please take it away. Okay, hello Danny.

SPEAKER_04 1:03:09

Can you hear me okay? Yes. Okay, you um you set me up very it was a very difficult task because I have to follow Dr. Ashwat. Uh and I just wanted to start with a quick story. Uh Dr. Ashwat and I overlapped in uh Cleveland. Uh he was my attending when I started out in fellowship. And one day I was just kind of bemoaning uh because you know LeBron James was there, and I was like, wow, LeBron is such a great basketball player, and he's leaving Cleveland or something like that. And um, Dr. Ashwat told me LeBron's a great basketball player, but he can't read an echo to save his life. So that that just put things in perspective for me. We're all 3D modelers here, and uh LeBron couldn't 3D model like us to save his life. So I hope that's inspirational for all

ENT And Tracheal Planning Wins

SPEAKER_04 1:03:55

of you. And I have a very difficult task of following up Dr. Ashwat. It was actually really nice to see all he's doing over there uh to such a high level. Uh Jenny, can you see my screen okay? Yes. Okay, awesome. I'll get started and um please feel free to cut me off. I'm gonna try to keep this to my 15 minutes. Uh, here's my one uh lonely disclosure. Uh I'm Shafkat Anwar. I direct our 3D uh Center for Advanced 3D Plus Technologies at UCSF that I had the opportunity to co-found in 2018, and I'm also the director of our cardiac MRI program uh here at UCSF. Um so I wanted to start with a short story. This story in about a minute encapsulates the way we use 3D modeling uh here at UCSF, and I hope this audio will work. This is the right ventricle. We've been using 3D modeling here at UCSF for about the last four years to plan repairs for very complex heart conditions. We have our patients come in for high-resolution imaging, and what we end up with is something that looks like this. This patient we decided not to go to a medical repair. Sometimes options are not apparent without a 3D model because you don't have that kind of uh detail to figure it out, and that's why our surgeons like this, because they can just kind of look in advance and figure out the connections. We're elevating imaging to its fullest potential where you're giving the best information possible to get the best outcomes coming out of the old social model. I feel fortunate that I'm here at UCSF because there are a few places that can push the technology this far. It's extremely exciting. So that uh video basically in a nutshell encapsulated why we use 3D modeling. Um, just like the other speakers here, we use it uh mainly for clinical applications to design uh very um complex procedures, especially in the cardiac space, which is what I'll focus on today. Uh, we also use it for patient education, trainee education, uh, and for some research and innovation. So um I try not to be stuck in a paradigm of print or no print. Uh the way we'd set up our uh center at UCSF is to go with this label of 3D, where we're essentially providing a toolkit of technologies uh which include 3D printing and also virtual surgical planning, which is essentially digital surgical planning, um, and a host of technologies such as uh augmented virtual or mixed reality uh to provide solutions for our surgeons. At the end of the day, for me, the technology is less important than being able to answer the question. And so, in order to answer the question, we apply each of these technologies sort of um in a custom fashion. Uh a little bit about us uh so far. Um we're about 100 models, and actually, this slide is about a month old. We're closing in on about a thousand 3D models so far. Uh, we try to um serve um all the surgical subspecialties uh within UCSF. Our highest utilizing subspecialties would be cardiothoracic, uh, orthopedic, and plastic surgery, and then followed by a whole host of other um other surgical subspecialties. Uh a little bit about how we're set up. Uh we found it in 2018 and we're headquartered here uh in Mission Bay in San Francisco. Uh, and we're spread out uh throughout our entire UCSF health uh network. Um so we provide 3D models for uh pretty much anywhere where we're operating. Um Parnassus uh is our historical campus. Uh, we have a second lab over at Bennyoff Children's Hospital, so two physical labs where we make models and send the models out uh throughout the rest of our health enterprise. And we have a very active collaboration uh over at the VA Medical Center, who has their own 3D lab. Uh and this is the team. When we founded,

Building A 3D Program Across A Health System

SPEAKER_04 1:08:00

um I essentially had a choice whether I wanted to make a 3D printing program for cardiology, uh, but I we chose to go the harder route and try to make a 3D program for all of UCSF. Uh, this was an intentional choice because we really wanted to partner and collaborate with some amazing folks throughout the UCSF enterprise that were already 3D modeling. Uh, because so we didn't really want to reinvent the wheel, but we did want to collaborate uh and kind of bring everything under one umbrella. So we founded with uh cardiothoracic surgery, orthopedics, radiology, and more recently uh we uh involved plastic surgery in our leadership uh with Dr. Lin's recruitment a couple of years ago. Uh over time, uh we've uh had two engineers now. For years, we only had one, but we were lucky enough to hire Sagar a few months ago. Uh and we partner actively with the UCSF Health Enterprise. So this is my business partner, if you will, over at the hospital system, Maimuna, uh, who really helps with our operations. Uh this is a quick tour. Uh, if you were to visit me in the hospital in Mission Bay, this is a very, very small room, about just 200 square feet. But this is where all our digital modeling happens on the San Francisco side of the bay. Um, I'm just trying to highlight some of the technologies we have available, and this is not an endorsement for any particular technology, just showing you what we have. So we tend to segment on Mimix uh and do CAD work on Threematic and also Geomagic Freeform. The haptic joystick is super helpful to our operations. And then we try to have uh several platforms available for extended reality. Um so I like to use EchoPixel Mixed Reality for some of our quick volume rendering cases, such as uh VAD fit trials. And also uh we use uh VR. Uh, we also have the Eleusis product that Dr. Ashwat very nicely showed. And then our augmented reality system is an iPad-based system, which works really nicely for education. Um, as far as um uh that was on the San Francisco side of the bay. This is the Oakland side of the bay. Uh, we opened up a new 3D printing center there uh just about a year ago now in October 2024. So this expands our operations cross bay. And this lab was custom built for sterilizable 3D prints. Uh, we now sterilize models throughout the entire UCSL uh health network, but here we have a very fast um operation where we can sterilize and deliver models uh straight to the OR. A little bit about our technology because we did decide to collaborate with other specialties. We were able to bring a lot of 3D printers uh under our umbrella. So we have about 30 3D printers that we have if we wanted to use it. But to be quite frank, we use maybe five or six uh routinely as our workhorses. The rest, I think, is a mixture of um prototyping uh FDM type printers, uh, which are great, but not necessarily medical grade. So mainly our workhorses are our polyjet printers, uh, which is our uh Stratasys uh J850. Um we have a nylon powder bed printer, which creates a lot of dust, um, and then a bunch of uh form SLA printers that we find uh very helpful, and also FDMs that we uh tend to use for our orthopedic cases. Uh so I'm gonna dive into a few of the cases right now. For sake of time, I'm not gonna go into a lot of detail on the complexities of each of these cases, but I'm gonna just show you a spectrum of cases from neonates all the way to much older adults uh and their applications. And as I do that, I'm gonna try to highlight uh the different ways we use these 3D plus technologies. So, this is one of our smaller patients. It's a one-day-all patient with something called double outlet right ventricle, where both outflows from the heart were coming out of the right ventricle and an unusual configuration of ventricles with the RV sitting superior to the left ventricle instead of being left and right, like usual. This patient also had subaortic stenosis and an interrupted aortic arch. So this arch actually had a big interruption here and here and was supplied by a patent ductus arteriosis.

Neonate To Adult: Complex Case Walkthroughs

SPEAKER_04 1:12:18

The challenge with a patient like that is essentially to be able to route blood from this left ventricle that I'm showing here in pink all the way out to this aorta across this VSD and not get any obstruction. And then you also have to relieve the sub aortic obstruction and then reconnect this arch. Um, so this is not the smallest baby that we've done a model on. We've done uh babies as little as about 1.8 kilo, and of course, we do fetuses, but this is just an example. So if you were to come to UCSF today, this is what you would see in our surgical conference. I will pull up our digital 3D model. Uh, this is on the Mimix Viewer platform, and I'll essentially uh go over the case the same way as I'd go over a CT scan or an MRI or an Echo. Uh at our center, we've made these essentially standard of care. We don't really treat 3D modeling as a novel or an exotic tool. It's just another tool that we use to hopefully get the best surgical outcomes possible. So here I'm just showing the surgeons the external anatomy, and then I have the ability to cut into the model. I mean, this group knows this well. Here I'm showing the intracardiac anatomy, the hole in the heart, the area where the baffle has to go through, and I can scroll through whatever views the surgeon wants, and then I'll I usually end up with this view, which is a surgeon's view. You can imagine the patient with his back here, chest is out here, the sternum, and then I cut through the aorta and the outflows to show them the hole in the heart through which the LV has to be connected out to the aorta and the subaortic area that needs to be addressed. So this patient underwent surgery and the patient had the surgery that we had recommended and did really well. Um moving on, I'm going to be showing you uh an application of virtual reality. This is a 12-year-old patient. Uh, we can essentially think of this patient as a single ventricle. And just in complete transparency, we use 3D printing to plan this case uh for the actual case, but then I went back later and showed a virtual reality application of this. If you're interested, you can also scan the QR code, uh, which gives you more uh details about this patient. So virtual reality is amazing because just as Dr. Ashwat showed, you can step into the model. This is the same 3D segmented model that we've already made in Mimics, but now we've put it into the elusive viewing system. Um, as you know, in virtual reality, you can do many things that we couldn't do even with the physical 3D model. You have unlimited cut planes, you can look at the anatomy in very unique ways that we would never be able to do in the OR. And I love this part. You can pick apart the whole model. You would never think about doing something like this in the OR because you know it would be a devastating result for the patient. But in virtual reality, you can just take each of the segmented parts apart, and then again, showing the baffle pathway here. And like Dr. Ashwat said, you can create virtual baffles, uh, you can create conduits, and then you can go completely immersive. Here we are, stepping into the anatomy, looking out the outflows, and showing the surgeons, if they so choose, this is one way of looking at the information to be able to fix a very complex heart defect. Moving from virtual reality into mixed reality, I'm going to be showing you a case of a patient who had undergone multiple heart surgeries already. And to keep this simple, essentially the patient had several problems. Um, one is a very large pseudoaneurism that was arising from this patient's aorta. The patient also had a single ventricle. Um, and the configuration was that the superior venticava was connected to the pulmonary arteries. And they had developed clots in the pulmonary arteries, which I'm showing in this dark material here. And in the past, the PAs had been replaced with a graft, but unfortunately, the patient uh arrived arresting with poor ventricular function and with seizures. And what we showed is that there's a lot of clot inside this large pseudoaneurism that'll need to be resected. This graft here needs to be taken care of. And there's also a clot inside the patient's aorta, uh, which is right by the coronary arteries, that needs to be addressed. Um, this is just a CT scan I'm showing, nothing exotic here, but we put this patient's uh entire CT into our mixed reality viewing system to very rapidly plan for this case because we did not have time to make an elegant uh detailed 3D model that would take at least several hours. So, this is something I threw onto my um computer. Uh, this is the mixed reality uh Ecopixel product. It's very easy to use because all you really need is a computer and 3D glasses. Uh, you can scroll through different types of visualization modes. So here I'm showing the lungs where the patient has a necrotizing pneumonia. And then you can basically cut through the heart this way, as if it's a digital scalpel, and you have a cut plane that you're cutting through. Here I'm basically labeling for the surgeons the areas that they need to intervene with that LPA graft that's clouded, the large pseudoaneurism, the coronary artery. I mean, this is just a minefield. We gave the family a 50% chance of death in the OR. And so, you know, very high risk surgery. So I took this model, um, this plan, I put the video on my laptop and just took the laptop to the OR and just talked about it right there. And after a pretty quick planning session with the surgeons, this is what they found. Uh, this is the pseudoaneurism and the clot inside it that they resected. This is the LPA graph, just like we had shown. And this is the clot inside the coronary artery. All went according to plan. The patient made it out of the OR. Thankfully, the cardiac function recovered, and she left our hospital in two weeks. And when she had come in, we had a very grim prognosis for this patient, but thankfully she had a really good outcome. So that was an application of mixed reality. And we're not going to spend a lot of time here because Dr. Ashwat showed this really nicely. But as you know, you can use 3D printing for percutaneous procedures. And here are some papers showing that. Um, I'm moving now from congenital to adult structural cases, and this is a case of mitral annular calcification. This was a unique patient because we made different types of uh 3D printed stand-ins, if you will. This is a sapien 3 valve that we have modeled. Obviously, I'm not showing the annulus here, but the point of this is to show how big of a valve can we put into this patient with bad mitral annular calcification, what kind of fit we would have, so that we can do pre-procedural, essentially bench top testing. So this is with a 26 millimeter valve. So this is me putting one of the valves in place or a model valve. As you can see, I'm looking down the aorta, and a whole lot of the valve protrudes into the left ventricular alflow tract. That is not an outcome you want on the OR. So we um we made a surgical plan based on the simulation to have like a 40-60 approach where we would have 40% of this valve sticking out into this patient's left atrium, and we would flare it open, and then you would have much less protrusion of the valve stent into the left ventricular aflow tract. As you can see, this was an 86-year-old, extremely high-risk patient because of several comorbidities, but we were able to do this as a hybrid procedure between the surgeons and the interventional cardiologists, all guided with 3D planning. He did really well and left the hospital uh after a couple of days. I'm gonna make be very quick about this for sake of time. We've applied 3D planning for hypertrophic cardiomyopathy. This is just a very hypertrophied left ventricle that I'm basically showing the surgeons how much tissue to resect. So now I'm looking down the aorta and I'll be showing them the uh ventricular cavity. And all the pink here is what people need to resect. We went ahead and applied this for education. We made 3D models and then got pig hearts for a wet lab. And guided by the 3D models, we practiced uh myctomies on pig hearts. So you see our um surgeons here, a proctor from Cleveland Clinic, and also

Mixed Reality For High-Risk OR Strategy

SPEAKER_04 1:20:32

our surgical residents and fellows. Uh, so we used it for training training purposes. Uh, I'm gonna just all I'm gonna say about this case is extremely complex anatomy in a patient with a burnt-out single ventricle that needed surgery. Essentially, the heart was um all wrong in the chest with dextracardia, systemic venous anomalies. And basically, we figured out how to implant a normal heart into a patient with extremely abnormal connections guided by 3D modeling. Uh, this patient also did well. All right. I'll end on some uh literature and some objective data, and then happy to take questions. There's now a lot of uh literature actually on use of 3D modeling in cardiac applications. This is a paper we put out several years ago now, uh, included over 100 references that showed 3D printing is accurate. It's uh great for uh pre-surgical planning, it is quite precise, it leads to improvements in trainee education and technical skills. It also increases patient and caregiver understanding of the disease process. And back at that time in 2018, it was showing uh data from ENT and orthopedics uh as far as reductions in OR time, which is important. This was a quick follow-up study, but a large follow-up study, meta-analysis of a hundred, over a hundred studies done in 2022 that essentially had the same complications across multiple sites. And this is just something that's on the horizon. Um, my um I am chairing a writing committee uh that is going to be putting out expert consensus recommendations on cardiovascular 3D printing. Uh, we've submitted this to a major cardiac journalist currently under review. Over 250 publications on the use of cardiovascular or 3D printing in the cardiovascular space. Part of our goal with this paper is not just to provide a review, but also to provide um practical recommendations on how to start your own 3D printing program. Um, so I'm happy to keep the 3D Heals group posted on this particular paper. For the sake of timing, I'm gonna wrap up by saying that uh there was a question about accuracy. This is how we track accuracy. We get these feedback forms back from our surgeons. Uh, and you know, we actually um I this is a year-old data now, but our accuracy rate as assessed by the surgeons compared to their anatomy in the OR was about 4.7 out of five. And this is across two centers out of eight years. And surgeons report that the 3D models helped them plan their cases for uh 95% of the procedures that they undertook. Uh, similar findings were seen in this multi-center trial back in 2017. You're welcome to snapshot this and read a little bit more about it. Patient and caregiving. I think it's great. I will leave it at that. And those are quotes from our own families. And for education, several um publications now out in the literature showing that 3D printing and 3D models help improve our understanding among our trainees for very complex cardiac anatomy and leads to greater satisfaction with their learning. So, my conclusions are is that 3D technologies in the cardiac space are here to stay. I think I personally believe they should become standard of care because they are accurate. Our surgeons and interventionalists clearly love it. We know that based on over a decade of feedback now that we've gotten, they are helpful for trainees and they're great for patients and our families. And there has been rapid growth and adoption worldwide. Um, a couple of QR codes, if you wanted to snapshot it, this will lead you to our uh center website. And if you're particularly interested, we do have a WhatsApp group for uh people in the cardiothoracic 3D plus space. There are many people I need to thank for all this work, especially our engineers profiled up here, our surgeons, my co-founders, and of course our families. Thank you so much.

SPEAKER_05 1:24:28

Thank you so much. That's excellent, wonderful presentation. Um I don't know when you moved to San Francisco, um uh Shawket, um, but we organized our first in-person event in 2017. And uh at the time we were collaborating with INOR, which is your orthopedic champagne there. And uh UCSF have five different 3D printers or five different centers that has 3D printing and they never talk to each other. And seeing your progress in terms of leadership and creating the center is really heartwarming. And uh it a lot of times it is true that technology may be the second important thing, and then the first important thing is people, and you seem to do a really wonderful job. Now, there are a couple of questions from the audience I think you already kind of addressed. One is uh from Charles Haven uh Havens. Um, how do you you know, when you first arrived at this brand new institution, there must be a lot of barrier. You want to share a little bit experience on

Bench Testing Valves And Training Labs

SPEAKER_05 1:25:30

how you tackle the problem to get the buy-ins?

SPEAKER_04 1:25:34

Yeah, um, so I think you're absolutely right, Jenny. It it does really start with people. Um, you know, when when I moved there, and I moved there in 2018, uh several uh kind of um uh camps that were doing three excellent 3D work. Um, and essentially it was doing a little bit of investigation on what kind of work's uh happening already, and then bringing people together under an umbrella and agreeing as a mission that we're gonna do this together. Uh, and I think that is not it's it's hard and it's not so hard. It's not so hard because people are extremely enthusiastic in this space, as you know, you've been a leader in this space for so long, and you don't need to convince a lot of people. We have the understanding that 3D technologies are better for patient outcomes. So when you get a group of leaders like that, then it's an agreement on okay, let's go for funding. We did go for funding to uh UCSF, we applied for a competitive grant and got some seed money. Uh, after we got the seed money, it was about 1.4 million. We're able to set up our operations, hired an engineer, bought a high-end printer, brought the rest of our printers and consolidated that. And then we just started. We started with our highest referring subspecialties, which were orthopedics, craniofacial, and cardiac. And we started making these models available. And part of my job as medical director was ensuring quality. That was one of the hardest things to have a very consistent set of operations. But because I come from the imaging world, as you know, uh, we have templates for that already. How do you get a great study? I basically applied uh imaging principles to 3D printing and set up uh checklists and workflows. So we put out a quality product and over time built a lot of coalition among our uh surgical subspecialists so that in a couple of years, when we predictably ran out of funding, I could go back to the institution and say, well, we've made X number of models, we've served X number of subspecialties, and now we need follow-on funding uh to be sustainable. And that's when UCSF essentially put us under the perioperative umbrella, and we brought all our operations in-house. So we exist under uh the surgical perioperative space, and our expenses now are underwritten by UCSF Health. So I don't worry that much about funding because that part is covered. Uh, we just focus on the work uh and the patient as much as possible.

SPEAKER_05 1:27:58

Yeah, it certainly is generating a lot of returns. Uh okay, one one question for you before we move on to the next uh presentation is um you show that you've approached you're approaching 1,000 3D models, and then maybe slightly less than that is actually 3D print. How do you decide which one is worth it, worth your time to do these?

SPEAKER_04 1:28:19

Yeah.

SPEAKER_05 1:28:20

In each case.

SPEAKER_04 1:28:21

Yeah. Uh so the the clinical models come to us from the surgeons, and you know, the surgeons get to decide if the print is gonna be helpful or a model is gonna be helpful or not. Uh, and most of the time they are. We've never had to turn away like a superfluous case, if you will. And if it does seem that we are not gonna add a lot of value, we just try to tell them up front and we don't even take the order, you know, you know what I mean? Um but if it's a case of model uh digital versus print, we always try to encourage our uh surgeons to think, do you really need to print? Because obviously we're producing plastic, it adds cost and time. And then we'll do digital modeling. And over time, our volume is broken down to about 50-50, 50% digital, 50% electronic. Although all of our models have an electronic version, as you know, so it's like 100% electronic, and half of that we will print.

SPEAKER_05 1:29:14

Do you have a rough idea of what kind of uh of cases require or people request to be 3D print?

Evidence, Accuracy, And Adoption

SPEAKER_04 1:29:22

Yeah, I do. For cardiacs, we've had much more success with 3D printing, and it could be because of our surgeon's experience. The tactile feel of the models plus the very complex intracardiac anatomy, uh, I think is easier to visualize. Um, for some of our um um bony cases, craniofacial, plastic surgery, etc., they're fine just going with an electronic model or orthopedics, unless they want to use it uh for cutting guides, uh bending plates, and things like that, in which case we do print out a model. And these days now we've moved on to sterilizable surgical guides, and obviously those are all printed. Because they're applying it and touching patients on the OR.

SPEAKER_05 1:30:02

Yeah, congratulations to moving to the East Bay, because I also moved to the East Bay. So I'm happy to visit physically if possible one day. All right, cool. So I am not seeing Jenny on this panel right now, which means her flight is delayed. However, I have some strategy and we already thought about it, given the complete shutdown of our government, uh all the airports are delayed. So that is uh one of the casualties. So what we're gonna do is we're gonna share uh she has a 15-minute presentation, but I'm only gonna share five minutes and I will put the rest of the recording on the Zoom event hub, and you can watch the complete uh recording after this event, just given the fact that we have so many live panelists here, and I don't want to play a video um for a long, long time. And hopefully she can join us for a um discussion at the end. So, what I'm gonna do is I'm gonna share her video for five minutes so you can get to uh have some FaceTime and have a context of discussion. And uh for meanwhile, I want you to think about questions for the speakers which uh have already spoken, and we're gonna start our panel discussion after this five minutes played. Okay. All right, I'm gonna start sharing. Okay, give me a thumb up if you can see this. Okay, good.

SPEAKER_07 1:31:38

Good morning. Um if you're seeing this version of the presentation, that means that I didn't make it on time for my flight. I'm flying to another conference, but I wanted to make sure to be part of this great webinar. So I record my presentation and hopefully I can log in on time uh for any questions. Um thank you, Dr. Chen, for the invitation. So I'll be talking about 3D imaging technologies and the role in congenital cardiac interventions. I'm one of the interventional cardiologists at uh the University of Colorado and the Children's Hospital Colorado. So, how did our program start? Um in 2018, we started to create 3D prints uh of the 3D rotational ideograms, which is uh 3D imaging acquired in the Gav lab. And we can get diecoms of those images and be able to segment them in virtual reality, but then get them to print. And I was really interested in bringing something to the families that they could use and understand the anatomy of their uh kid or their themselves in a nicer way. So you may have her already. Um Nick, he started with us uh collaborating and he helped me set up my first printer, and it was amazing. Uh so thanks for that help. And here's um our collaboration with also my colleague Dr. Gareth Morgan and some of the CFD printed hearts. So the goal for printing uh back then was for procedural planning, but especially for patient education and trainees' education. So we experimented with ways of how to print our

Cath-Lab 3D Printing For Families

SPEAKER_07 1:33:14

way with vessels so we could see the correlation without uh you know shifting anatomies. So from here, we can see how the 3D array works, uh goes 240 degrees over four seconds. We do the injection in a specific um place, and then we post-process the images. We use them in the CAFA for procedural guidance. We can overlay them, we can uh plan procedures with it. And our protocol in Colorado is that we don't repeat the angiogram, so actually helps with decreasing uh contrast and uh radiation during these procedures. Um, but then the images can be manipulated and export them as a CT-like image and uh 384 slices from the Philips system specifically, and then we process them for print. So this has been a great process, and since then we've printed over 800 hearts for patient experience and education. We can see one of our first. Uh this patient has the quartation before and after. This was from the CT, and the other one was from the 3D rotational angiogram or 3D array from the cat lab. This is how it looks. We do a little box, have a label uh with our names. Now, obviously, this change. We place a little sticker and has the four of our interventionists in our program. So it's grown with time, and we can see the differences in this case, an RVOT stent before and after, and the family can see this and understand what we really did. And we printed at scale, and we're using a formless print uh printer with draft material. So it's very quick. Turn around, it's like an hour. I have segmented the images and then it's uh post-processing thanks to a lot of the students from the modern human anatomy master program here in campus uh that they come and do projects with me. So we can see a lot of the publications we work with, and a 3D printing has become a great tool for education for patients, and it's part of our workflow nowadays, and it's uh something sustainable, as you can see, um, that we have specifically done it in the Cat Lab. This is uh few of the QR codes for our program and articles about this, and our our multidisciplinary uh work together um to get things uh going on this specific field. But then we have um how we uh evolve from 3D printing to 3D viewing, and I'm saying viewing because we uh get a lot of uh tools to do this, and um we have decided since 2020 to transition for uh procedural planning with virtual reality and then in augmented reality for procedural guidance. I'll show you more of that. And um, we have done a library of these devices that are pre-designed at scale and easily imported for planning. As you can see, we have VADs, Impella, multiple pulmonary valve frames. Um, we have plenty of devices and we have done validation to show how we can anticipate what device is gonna fit compared with the gold standard, specifically for self-expanding valves. Um, we have created uh plenty of these devices, just creating a 3D array of these devices and shaping them as needed and then importing them into our library. So, this is something sustainable that we have done ourselves. Some of the companies are able to give us the SDL files. So then that saves a lot of time on the process. But if you don't have them, it's easy to get them done. But then all these tools help us get measurements, annotations for future reference, but also for planning. In the left, we have the RTNS Avent Reality tool where we're, you know, we're working here in the workroom, and we are uh processing like this Pum Laiba planning. We can get a center line, plan what's gonna happen. We can have that projection on a

Transition To VR Planning And AR Guidance

SPEAKER_07 1:37:08

screen so other people can share, but we're collaborating in the same environment with another colleague. And in this side, we can see how we can plan cases, creating patches which are pretty dynamic in 3D. We can do measurements which are not uh necessarily like a straight line. We can fly into the uh defects and identify uh in this case uh these complex uh VSD. Uh we team up with the surgical team to create these baffles. As you know, for them is very important to have a sense of the size of things, but also if it's possible in the chamber um when it's patched, if we have enough um area, specifically for very complex biatricular repairs. You can see how there's a 3D involvement and interaction of that uh patch. So this is very great tools and is uh encourages multidisciplinary collaboration. Um, talking about that, we have here a case where we're collaborating with uh some of our colleagues in novelist centers using the same system, uh elucids from released medical environment using virtual reality. We have published this with my colleague Dr. Arash Salvitavar from Columbus, Ohio, and uh we are able to uh review these um images together. This is a case that we did here uh in the hospital, uh, where we have the eyes image. So we have a 3D echo, this is infracardia echo, where we see it uh dynamically uh overlaid under the CT. So this is the same patient, and we in uh imported um the echo images. The same thing you can see on this lower panel. We have the 3D echo, which we can import, and there they can be, you know, they're dynamic, and we can have a dynamic CT with a dynamic echo. Very important to see when you have connections in uh complex biventricular repairs again, to see if that's possible. You have codec crossing VSDs. You can overlay the 3D echo on the CT using these tools, and this is another way of viewing it. You can see the valve there, and you can fade the image over the CT and see how it's aligned. The 3D eyes in this case, you can use TEE as well. So it's critical to say that uh we can have have multi-imaging tools, um, not just CT and MRI in this, but also we can now incorporate uh our I's 3D I's 3D TEE, 3D TTE, transthoracic echo into our uh VR environment. And this brings an overall image that is otherwise very hard to get in a 2D image. Um, these tools are available on a desktop, but I think the full experience and understanding of the anatomy comes from wearing the headset and having the 3D um idea of the structures and relationship of multiple structures.

SPEAKER_05 1:40:04

Okay. Um, we're going to upload the video onto our event hub so you should be able to see it um after this event ends. And I want to use the last couple of minutes to do a panel discussion um of the things. And I think the video itself also kind of segue our first question is um, and I want to invite uh everybody on the screen, please. Um actually, Nicholas, do you know what VR software uh Jenny was using? Because it looks really familiar, and also I think it's one of the pitch 3D startups as well.

SPEAKER_03 1:40:40

I I believe she's using uh I believe she's using Elusis.

SPEAKER_05 1:40:44

Oh yeah, that's right. That's right. That's right. That is the company. It's a Canadian company, yeah.

SPEAKER_03 1:40:49

Yeah, with Elusis um and has worked really closely.

SPEAKER_05 1:40:53

But all right, so she just made it. Jenny, you're right on time. Come on, join us. Uh so we were thank you so much. Um, we're able to uh play half of the video uh just because of time constraints, but we will upload a video online. So, but we got most of the gist of what you're doing. Um, so I think right now it's perfect timing. Um the one thing that you mentioned in your presentation was quantification, and I'm starting noticing a lot of academic centers, especially for vascular works, quantification, just you know,

Multi-Modality Fusion: Echo On CT

SPEAKER_05 1:41:25

measurements and stuff like that, you know, both either 2D or 3D is becoming more prevalent and important. So I'm just kind of curious in your digital workflow, how do you decide which measurements are important and how do you, you know, make the workflow smooth? You know, yes, we, you know, what numbers are important for you guys? And how do you streamline the workflow to get these numbers?

SPEAKER_08 1:41:51

Yeah, I I think that's a critical uh question. Um thank you for having me and sorry that I'm rushing, my flight was delayed, but I appreciate your time. Um for us, it's important to determine first what's the question, what's the procedure being expected. So the workflow has been um most of the time, I am the one from the task perspective and the surgeons. Uh I download the images, go through the process of uh you know, identifying the procedure that could potentially be being done, and then identify the measurements that are key. I think it's important to not measure on the 3D rendering, as you know, that we can change that and make it very inaccurate. So a lot of the measurements are most of the time in the blood pool to make sure that they're uh you know the most accurate that we can do and in different views. And then we have most of the planning goes along with uh sense and may make uh a match for that area and make sure that it makes sense. Um so I think the technology has been so good that there's so many different tools to uh double check those measurements, and of course, we rely on our radiologists to go through the traditional uh routes, and if we have a discrepancy, discuss this to make sure that we keep quality as the first thing.

SPEAKER_05 1:43:09

Thank you. Um Shavkott, uh, Dr. Um Ashwath, do you guys have any comments and Sarah? And do you guys have any solutions for uh these quantification um requests and stuff like that?

SPEAKER_04 1:43:24

Um yeah, I guess you could call me a traditionalist in this space, right? Like, you know, I grew up with traditional 2D imaging, and I still rely on 2D imaging quite a lot. Um, and that's how I train my fellows. They have to learn 2D in order to advance to 3D. Although I agree with Dr. Ashwat, sometimes you can back train, you learn on the 3D models and then look back to 2D. So when it comes back to quantification, obviously there's traditional methods, um, double obliquing in parallel planes. That essentially comes off of a good 3D data set, but on traditional things like multi-planar reconstruction. I do have some pause when we're doing fancy quantifications off of 3D models, especially from blood pool models. Like to Dr. Zapla's point, uh, especially with volume rendering.

SPEAKER_08 1:44:10

Give me my stuff head up there.

SPEAKER_04 1:44:12

Yeah, I think so. Uh I think it's it's uh you have to be a little cautious, uh, measuring off of 3D models, unless your segmentation is very accurate. And then we will routinely quantify. Uh, I do my quantifications all on um materialized mimics uh with uh with the threat segmentation, not in virtual reality. So there I have you know pixel level resolution, so I I can make any measurements in the 3D space. So I think the point I'm trying to make is just be very, very OCD uh about data quality uh and your segmentation. And then if you have good segmentation, I think it's fine to measure on the 3D models. Okay.

SPEAKER_01 1:44:54

I'm uh I'm on the same bus as the Chefcut's OCD bus. Uh we're all on the same bus. But uh for me, I think uh, you know, if it's the old tradition,

Panel: Measurement, Accuracy, And Materials

SPEAKER_01 1:45:05

I'm older than that. I think the 2D raw images, as you being a radiologist, you know that raw images is what I want to go by. But in cardiology, the other thing that people have realized is what phase of the cardiac cycle we are trying to get measurement is very, very important. If we're doing volumes and we are getting STT in one phase, which could be ancestral, it's gonna be totally off of what the surgeon is looking for. So raw images, phase of the cardiac cycle, but at the same time, I have the uh acceptance or at least I have the openness to see if the blood pool model and this, how much they're correlating every time for us to make a decision that yes, you can go by this, you know, it's gonna be this much. I don't know whether we have done a study. Maybe it is good to do a study to, you know, compare, you know, the 3D blood pool volumes versus the regular ones and actually go from it. So I think being very particular and also doing disclaimers is very important of when you measured it and what were the difficulties in measuring it.

SPEAKER_05 1:46:08

Sarah, do you guys have any solution to automate this process or make it easy for the doctors?

SPEAKER_06 1:46:15

Yeah, I mean, I guess it depends on which aspects would be the most important to automate. I mean, right away you would think the being able to bring in the multi-image um multi-phase data easily and be able to segment it all as quickly and easily as possible. Um so there are some tools that I know we have in Mimix where you can bring in multi-phase data, you can run an algorithm to output the segmentation. Um it might, I think it might be a bit limited to just the left part at the time being. But um I think it would be interesting to to explore and even get some feedback for what would be the most important aspects to to automate, to make as efficient as possible and see what we could come up with on our end.

SPEAKER_05 1:47:05

Yeah, uh well, I'm sure all the doctors on call right now can tell you, based on their own experience, what is uh more important. Now, Shaft Cut, I remember that you did a survey and you know, how accurate your surgeons think your models are, and it was 4.5 out of five. You know, if if you know I'm a typical tiger mom, I'll be like, why didn't you get to 100%? Um, so I'd like to post that question to everybody. Um, you know, what do you think can need to take place so that our accuracy is gonna be 4.9? You know, that's like a reasonable Uber score for you. So if you're not gonna take an Uber, if it is not 4.9 plus, why would you want to go under surgery for your heart with a score of 4.5, which is not really bad, but you know, not perfect, like I said, being a tiger mom.

SPEAKER_04 1:47:56

Right, I'll take that first. So it was 4.7. 4.7, okay, not bad. So you know, I think you're really talking in terms of um sub-millimeter levels here, right? So what the bar that we set for that survey that I showed was that if we're off by a millimeter, one millimeter, they get to knock us down a whole point. So instead of giving us a five, they give us a four. Uh so first we set the bar pretty high because we wanted to capture really like how accurate we were being, because we were scanning at 0.6 millimeter resolution. So I thought we should at least be um printing at a millimeter resolution. That's that's reasonable. And where we missed points were more matters of perception or lack of um attention, I would say, to what the model was showing versus what the surgery showed. So, for example, we had a model where we showed two VSDs, and one of the VSDs were quite small, and we had shown it in the model. Now we review our models maybe two weeks before the actual case, and the surgeons focused on the big VSD. And thankfully, I was doing the TEE for this case. So when I went back in the OR and I discovered that the other VSD was there, but now it was larger because it was getting more flow through it. I pointed that out to the surgeons, to which the surgeon said, Well, you didn't show me this. I was like, Yes, I did. You just didn't pay attention to it. You know, so some of those uh differences in the scores is what the surgeons uh think they saw in the model versus what they overlooked. And some of those scores need to be revised. But it is true. Uh, we were a 4.7 uh out of uh eight years and two centers, uh, which I think it's fine. That's like what closer to 95% accuracy rate. And do we ever get the same data for MRIs or CTs or Echo? Like, do we ever go back and ask the surgeons to grade us on our Echo reports? We have a QI process, but I think we set the bar really high for 3D modeling, and that's self-imposed. But I wonder what our scores would be for our traditional imaging if we actually went back and asked.

SPEAKER_05 1:50:02

Yeah, that's a good question. It may never have an answer, actually. Well, what do you think, uh, Dr. Ashwat, you know, in terms of what can be done or what technology is needed uh in the future to make it more accurate?

SPEAKER_01 1:50:17

I think the first of all is the the surgeons being the gold standard over the years, you know, working in centers with different surgeons. I'm not sure why we have even benchmarked surgeons as the gold standard. Hopefully there are not many on the call, but uh I think we have sent cases to different centers and we have the diagnostics come back uh with uh several different diagnostics on the same patient with surgery. So comparing tough cut study to a surgeon and objectively measuring these, I'm not sure whether it is uh actually the right thing to do on this particular uh pediatric cardiology with so much variation of nomenclatures and all that stuff. So I mean I think your question is uh well taken in terms of what are we doing to make it a five. But my question is uh what is five when there's uh different surgeons telling different things. So the question is is it going to be a five for one surgeon versus not? So in my small field of pediatric cardiology, I think the best thing we have found is uh to actually explain to them exactly what you see and what

What We Need Next And Closing

SPEAKER_01 1:51:20

are the things that they can encounter and then come back and ask them if there was a little difference, you know, was it a difference that actually mattered for surgery for the output? Whereas it makes no difference if it's a five millimeter VT found a six millimeter VST, makes no difference in any of those things. So I think one of the points that can be added to this 4.7 is in the end, did it make a difference to you for approach and the results and the outcome for stuff? So I think based on that, I don't see a whole lot of uh coming back saying it different. So we will try with the engineers to make it phi, but I think both have to be educated on the same page, and the gold standard may have to change.

SPEAKER_04 1:52:13

Absolutely. If I may add to that, Kenny, one thing that we probably don't do as good a job on is AV valves, atrial ventricular valves and cordae, because these are mobile structures and they're very thin. Um, but you know, what's exciting to what Dr. Zablo was showing is co-registration of technologies or at least overlays of complementary technologies, because CT and MRI may not show AV valves that well, and part of our you know imperfection, if you will, was because um we show try to show the valves well, but from CT we can't show it all that well. But what Dr. Zablo was showing is you can take echo data, which does show valves really well, and co-register that information to make a more comprehensive model. So I think multi-parametric imaging is important to make models more accurate.

SPEAKER_05 1:53:01

Yeah, that's really well said. Jenny, do you have any input on those?

SPEAKER_08 1:53:06

Yeah, and I think uh I agree with both. Uh the key is a few things is trying to combine technology. I think uh moving some from 3D printing to three like virtual modeling has helped to go in detail to other aspects of the planning. But I think uh same thing. The 4.7 is very uh, I think I agree with that number, but it can be uh, you know, depending on the case, it's very difficult in considerable heart disease to give an accurate diagnosis, as uh Dr. Ashwat was saying, mostly because of the different nomenclatures. But the key is that the surgery that is done is the right one, and also that we can save time under an institution for the patients. And many times just having an alternative approach, we tend to collaborate a lot with our surgeons. So just choosing to go to the cat lab for us to do a procedure or do a surgical uh piece is also critical, and also having the right tools. So I think those things have built a lot of resources that despite the technology is not a five, yeah, the change that is done, if you compare that from before the times of 3D anything, um, is already a five in my mind. It's a win. What is coming, I'm very excited to see. Uh and again, I think it's a lot of uh having all of our heads together and seeing the need with the day-to-day uh applications so uh everybody else can start creating and building from that. But I think we have a good technology already to build from.

SPEAKER_05 1:54:35

Absolutely. I think one thing that people didn't talk about is the overall improved quality of pre-surgical planning, uh, it reduces human error. And actually, the fact that we say less in the hospital from what, 22 weeks to one day, just imagine the number of human errors you're gonna avoid. And in fact, medical errors kills equivalent as accidents per year in the United States. I just put up that data actually, in case you want to look. And so so that in itself, you know, reduced operating time, that also reduces and seizure errors and nursing errors and all that stuff. So it's is is a huge lifesaver if if you believe uh what I just said. So yeah, no, I absolutely agree. Uh Nick and Sarah, do you guys have any yeah, Nick, you're raising your hand. Just speak up.

SPEAKER_03 1:55:23

There's one thing that I would add to that is that it's not always just about being able to visualize CT and MR, but there is such a wealth of data out there. If I put on my computer science hat, um uh every day I go into the UR, I see new tools, new forms of data acquisition that gets so lost. It's so difficult to be able to see these. If it's difficult to be able to visualize anatomy from a stack of two-dimensional images, the vast amount of data out there, we need new forms of 3D that can merge all of that data so that way we can more accurately comprehend and understand that data to take full advantage of it. So it's it goes so far beyond just those simple planning models, but the integration of all the forms of data that are coming down the line.

SPEAKER_05 1:56:16

Absolutely. It's like learning of new skills, except in this case the computer is learning the new skills more quickly than us. Um we have one question in front of the audience in terms of the three printing materials that you guys prefer to use. Uh anyone wants to jump in on that?

SPEAKER_04 1:56:36

I can take a swing of that. Um, we're fortunate because we have uh kind of a high-end printer that can produce materials that uh mimic different densities. So we uh use the polyjet technologies anytime we want to make cardiac models. Um the drometers can go anywhere from as dense as bone to as soft as skin. Um, it's okay for simulation. I mean, out of the print technologies that I've experienced, I think probably that Stratasys uh makes good uh models that kind of mimic tissue, but probably the gold standard would be silicone. Um silicone casting is uh difficult, but I think the group at Toronto Sick Kids has shown that it is possible to use silicone, it's just a much more expensive and time-consuming process. So without that, I I still think that the polyjet's good. We've got good feedback from our surgeons about it. Okay.

SPEAKER_05 1:57:27

Uh Dr. Ashwas, you have any input? Uh you're muted.

SPEAKER_01 1:57:31

No, I think I've uh enjoyed working with Agilis for the most part, and I stick with that for uh here. Um but I've done with other printers to other uh materials too.

SPEAKER_05 1:57:43

Jenny.

SPEAKER_08 1:57:46

I think I'm gonna give Nick the word here because as we've shifted more for using prints for education, I'm not gonna have a favorite. So Nick, what's your favorite?

SPEAKER_03 1:57:57

Uh there's a saying in Wisconsin that says horses for courses, use different horses for different courses. So there's so it depends on the application. But I will say I'm really excited. There are there's a new uh a couple new uh silicone 3D printing groups. There's one that just spun out of MIT that uh um I've got some samples for. Uh just got some samples last week, um, and it's biocompatible um silicone platinum that uh um has a varying level of barometer that becomes really exciting. Um so we live in a wonderful time right now. You know, maybe six years ago before COVID, there was a small handful of um very fatigue, very fast degrading materials. And since then there's been this explosion of materials. Um right now, silicone uh 3D printing has my attention.

SPEAKER_05 1:58:48

Cool. Uh we actually also had a pitch 3D startup called 3D US that's in uh France that actually creates um silicone-based models as well. Um, but thank you guys so much. Um, so we're approaching the end of this webinar, and somehow we cobble together this panel that's just incredible, um, even though people are having logistic challenges. But thank you so much for making it. Okay, so want to make this webinar worth it to you. What are your immediate requests for the next year or so uh that you wish you can just shout out to the universe and hopefully somebody's listening? Uh Nick, I'll start with you because I know you have a long list. But just one though, just one. Um pick one.

SPEAKER_03 1:59:31

Um uh well uh advanced imaging. Um, you know, I'd like to see some of the latest in imaging technology and what people are doing with that.

SPEAKER_05 1:59:42

Absolutely. Bad data is equal to bad print.

unknown 1:59:47

Dr.

SPEAKER_05 1:59:47

Ashwaz.

SPEAKER_01 1:59:49

I want to see how much uh valve anatomy that we can include with Dr. Zabal working on stuff, things like that. That's where I think we have not done a great job for the surgeons with the valve. So I think including the 3D dataset from the echo cardiograms integrated into the CT and MR, I think is what I want to say to make it very seamless.

SPEAKER_04 2:00:12

Thank you. Shavcat? Uh mine is probably less technical. Uh I would love to know how we can disseminate this technology more and make it more sustainable within our medical centers. Uh there are some centers like yours that have now made this technology kind of standard of care, but uh it's n it's hard to disseminate. So I'd love to learn more about that. Sarah?

SPEAKER_06 2:00:39

Yeah, I agree with the sustainability aspect. That would be, I think, really interesting to get a lot of different kind of points of view on what's been successful, what is predicted to be successful, and and where all of we can find like all these consolidated resources essentially. Um yeah, that would be very interesting. And I think also, I mean, just continuing to see how AI is impacting the space.

SPEAKER_05 2:01:03

Yeah, curious to see as well. Jenny.

SPEAKER_08 2:01:07

Well, I think I'm gonna go back to the technical, and I think for us at the Holy Grail is having compliance on the simulations virtually, so we can anticipate compliance of the tissue. Um with the simulation, I think that's gonna give us more accuracy. So looking forward to see what we can do with that. I know there's things out there, but something easily available and updatable.

SPEAKER_05 2:01:28

All right, you know what? Um, I'm gonna ask my assistant to make this final segment of video public so that it indeed will be to the universe, and hopefully, some collaborators or partners can show up for you. Thank you so much again for donating your precious time to join us today. The recording will be online for a couple weeks. Uh, so you can share the link, it's free for anybody who is uh interested. Thank you so much and have a wonderful day. Bye bye. Good seeing everyone.