Effect of 3-Dimensional Virtual Reality Models for Surgical Planning of Robotic-Assisted Partial Nephrectomy on Surgical Outcomes - Ketan Badani
July 27, 2020
Ketan Badani, MD, Professor, The Mount Sinai Hospital, New York, New York, USA.
Jaime Landman, MD, Professor and Chairman, UCI Department of Urology, UC Irvine Medical Center
Sumanta Kumar Pal, MD, Associate Professor, Department of Medical Oncology and Therapeutics Research, Co-Director, Kidney Cancer Program, City of Hope
Jaime Landman: Welcome to Kidney Cancer Today. My name is Jaime Landman from the University of California, Irvine.
Monty Pal: And I'm Monty Pal, the Medical Oncologist at the City of Hope Comprehensive Cancer Center here in Los Angeles.
Jaime Landman: Today we are really very fortunate. We are going to be talking to Ketan Badani, who is a urologist over at Mount Sinai Hospital in New York City. I've had the privilege of working very closely with Ketan when we were at Columbia together for a good long time. He is one of the world's most respected robotic surgeons and has published recently on some really interesting, novel aspects of partial nephrectomy. Welcome to Kidney Cancer Today, Ketan.
Ketan Badani: Thank you, Jamie. Thank you, Monty. It's a absolute pleasure to be here and an honor.
Jaime Landman: Well, let's jump right on in if you don't mind.
Ketan Badani: Yeah, absolutely.
Jaime Landman: I'd seen your work previously, and it finally came to fruition, the manuscript you published in JAMA Network. That was just last month on the Effect of 3-Dimensional Virtual Reality Models for Surgical Planning on Robotic-Assisted Partial Nephrectomy on Surgical Outcomes.
So, just as an introduction, we all use CT scans and MRIs. Each one typically has between 1,500 and 2,500 images. I did a little math, and if you spent three seconds on each image, it would take you two hours to go through any CT. VR could be the future, in that it takes all these images and literally spoon-feeds your brain the data. You took full advantage of that in this publication, didn't you?
Ketan Badani: Yeah, I think you put it in a interesting way, where you look at the massive amount of data on an axial scan, and then we interpret it in a way where we're looking at 2D images, and we're rendering it into a 3D image in our head, and then building that model just sort of in our head by looking at the pictures. But we've talked about this before. You and I both know that actually we're pretty bad at doing that in reality. The goal here was to actually take 3D models, and again, in this study, this was virtual reality models, as opposed to physical models. Some very interesting centers are doing this work where they're actually 3D printing models, but these are virtual models.
I'll just say that one of the unique and neat advantages of a virtual model is you can add and subtract layers of anatomy pretty easily. It's all software, so, unlike a physical model, where it takes a lot of time and effort to... Let's just say you want to look at the arterial anatomy of a kidney, then you want to look at the venous anatomy, and then you want to just look at the collecting system. When you do it virtually, you can easily just flip through these different views. This is the premise of saying, okay, well, this has been shown in the liver to be beneficial, in lung, but it's a perfect thing to study for kidney because, as we all know, kidney anatomy is highly variable, not just arterial but tumors, depth, invasion, etc. So could we somehow find a meaningful change in outcome by using actual 3D purpose-built models, as opposed to doing our own 3D rendering in our head or doing this basic 3D rendering that our radiologists would do on a CT or an MRI?
Jaime Landman: That's truly amazing technology. The study you're quoting is the... There was an NYU study where it showed only about 30 or 40% overlap where surgeons thought the tumor would be to where it actually was. We actually repeated that study with slightly larger numbers and found very similar results. That'll be at the World Congress in Dubai this year.
The software you used, that was commercially available, or is that still something that's investigational?
Ketan Badani: This software now is commercially available. This is through a company called Ceevra Tech, C-E-E-V-R-A. They build out the models through an online system, an online portal. So it is now commercially available.
Jaime Landman: Cool. So anyone can pick this thing up. And what does that cost us?
Ketan Badani: Honestly, I don't know a lot about the cost model. All of this work I've done has been through validation and research protocols. Your question is an excellent one, and not one that I'm capable of answering at this time. But I think that the premise, and not unique to this particular product but other 3D models and IT-based, is a subscription-type model where there is a monthly or an annual or some type of subscription fee, and then you just have access to that technology for that period of time. I've seen that in other realms, and I imagine this would follow the same suit.
Jaime Landman: And do you know, do they do their reconstructions with some kind of an algorithm, or is it manually done, or some combination?
Ketan Badani: It's a combination. There's definitely a manual component to it, but it's not all manually done. I think they're able to get a pretty good framework from the axial imaging, but, obviously, when you're doing regions of interest and honing in on actual arterial anatomy, et cetera, there's a manual component to it. So there's a little bit of labor involved.
Jaime Landman: I think this process is fascinating. So you just send them the DICOM of your patient in a de-identified way, and then they send this back to you?
Ketan Badani: That's right. Yep. It's de-identified DICOM image. Most hospitals can give you a de-identified disc, so it's actually a pretty easy process. Then you upload it onto a web-based platform, and I'd say it's probably about a week turnaround time, somewhere between five and 10 days, where you get the model back. The model comes back on an iOS or Android app, so you would open it up on your iPhone, your iPad, that sort of thing. You can view these models in 2D on your device, or you can use a VR headset, a gaming headset, basically. But you can use a VR headset that you put your phone into and view it, actually, in VR mode. These are the two ways you can view the content.
Jaime Landman: Is it one of these interactive models where you can use the platform to take the anatomy apart? Can you pull the tumor apart, the collecting system? Or is it all just integrated and you can float through it?
Ketan Badani: You can definitely add and subtract anatomy. You're not quite doing it with your hands, so to speak. You'd have to check box what anatomy you do and don't want to see. You can take out venous arterial collecting system, parenchyma, tumor. You can just take out the tumor and see what the defect might look like. There's all these different layers of anatomy that are kind of neat, but you'd select those just like you would on any app.
Jaime Landman: In terms of your personal experience when you were using it, was it the kind of thing where you looked at the CT, and then you looked at this separately?
Ketan Badani: Exactly. What we did in the study was we recorded... These are difficult things to do because there's very tangible things that you do in the operating room that benefit you, and there's very intangible things too. I think the benefits of this type of technology also fall into an intangible category. So yeah, we look at the CT, trying to take note of how long we're looking at it, how many times we opened it, get up out of our seat when you're doing robotic surgery to look at the imaging and scroll through it, and then use the VR and do it that way. You find that you actually do end up scrutinizing anatomy on a CT scan quite a bit. Without paying attention to it, you don't really notice it, but then when you're actually recording this stuff, I found myself on average getting up three, almost four, times throughout the course of the case to fire up the computer, to get out of sleep mode to look at the image and scroll through it and look at the various anatomical details, where the arteries branch, et cetera, et cetera.
So when you start looking at, you say, "Wow, I'm really kind of disturbing the operation a bit by getting up and getting down to keep looking at the anatomy." So these intangible things, where I really was doing that far less when I had the VR model... Hard to capture that, but that's really where you start seeing the benefits of this type of technology.
Jaime Landman: Essentially, I think what you're telling me, you just had a better overall understanding. At least that was your gut feeling.
Ketan Badani: Yeah, yeah. I would say if I had to point to the most impactful thing that a 3D model... It doesn't even have to be virtual. We're saying VR model, but quite honestly I'm simply just talking about the concept of building this 3D model. You don't have to look at it in a 3D headset to benefit from this. But the most impactful thing was really just looking at the arterial anatomy. If you know how many arteries, where they branch, where they are in relationship to the vein, it just speeds up the entire process of dissection, the operation. A lot of the op time, in my mind, and this wasn't clearly noted because it wasn't captured in the data, but a lot of the op time benefits in the surgery in the VR cohort, I think, was because the understanding of that vascular anatomy was so crystal clear that there was no time spent trying to figure it out.
Monty Pal: Got it. Really, this is just fascinating technology, Ketan. I've got to ask you, this must have some implications for blood loss. Did you notice that in the context of your study?
Ketan Badani: Yeah, the study did notice a decrease in blood loss. I don't have a great explanation for why that would be. I guess shorter op time and shorter ischemia time would lead to less blood loss. That inherently makes sense, but I'd say it was more of a secondary benefit from these other factors.
Monty Pal: Got it.
Jaime Landman: Where does the virtual rubber hit the virtual road? You did about 80-something cases between the group of investigators you had. What do you think the real benefits are?
Ketan Badani: Well, Jamie, I think we're just at the tip of the iceberg here. The 3D modeling, the potential artificial intelligence that'll go into guiding surgeons, anatomy landmarks this is just the beginning. So although this technology is very neat, and it's actually very practical and useful right away, because, again, right away you can understand the anatomy better and... It's not changing the fact that patients overall did well, right? We're not saying that, "Hey, if you didn't have a 3D model, those patients had poor outcomes." They had great outcomes. These were just a little better. But ultimately, as these types of technologies improve, it will make more and more meaningful benefits.
I think we're just at the beginning, where we have an opportunity to get 3D models in our hands, which we just didn't really have. You've done such tremendous work with 3D modeling, but you have a team, right? You have infrastructure. It's not something that was easy for the guy 200 miles away from you to get. At least now we have the ability to put these in the hands of surgeons to help them and then continue to refine it and make it better.
Jaime Landman: I couldn't agree with you more. The key is here to automate this, and we're working with our machine learning team to do it as well. That's going to be the final outcome where it's just kind of a standard thing.
The only other question I really have here, because I think you really have summarized the state of the art beautifully, is did you try this on your patients?
Ketan Badani: How do you mean?
Jaime Landman: Did you let your patients see the models?
Ketan Badani: Oh, oh. Excellent, excellent question. And I'll tell you, absolutely. The answer is absolutely, and every single one of them loved it.
More important to your question is enrollment. When the idea was put forth and we're designing the trial, one of the sites said, "Well, are we going to be able to accrue patients to this?" My initial answer was 100%. 100% of the time, if you show someone a 3D model and tell them you're going to have this for their operation, they're going to want to do it. And that's what happened. Every single site had virtually 100% accrual because patients love this. They love looking at this.
And you know what? You can explain the anatomy to them better. You can help them understand what you're doing. You can do it on a CT scan, but if you've never seen a CT scan, you're just lost in livers and lungs and all the other stuff in the abdomen and in the chest. But a 3D model is just a nice clean way to actually have a meaningful conversation with the patient. I'm glad you brought that up.
Jaime Landman: I actually think that at the current time... Because you said it. This is an incremental improvement that will take good outcomes and make them just a tad better. But I think that when you show patients CTS and MRIs, they really don't get a grasp of the anatomy. And when you show them this, which essentially just spoon-feeds the anatomy in a very intuitive way, it really has great value there too.
Monty Pal: Yeah, agreed.
Jaime Landman: Well, that was really a great summary. Ketan, on behalf of Monty and myself, thank you so much for taking the time to enlighten us with regards to the state of the art of VR and partial nephrectomy.
Ketan Badani: Thanks for inviting me. Tremendous honor, and always nice to speak to you.
Jaime Landman: Oh, likewise. This is Jamie Landman, signing off for Kidney Cancer Today.
Monty Pal: And Monty Pal here, signing off as well. Thank you for listening.
Jaime Landman: Thank you.
Ketan Badani: Take care, guys. Thanks a lot.