First-of-its-kind simulation of the heart, lung, brain and knee culminates in a transformation of healthcare processes.
Dassault Systèmes is looking to make a massive impact in healthcare technology. The software company is en route to creating virtual twins of human beings for a variety of medical applications. So far, the following programs have been established:
- Living Heart Project
- Living Lung Project
- Living Brain Project
- Living Knee Project
Engineering.com had the opportunity to speak with Steve Levine, senior director of Virtual Human Modeling at Dassault Systèmes, about the company’s vision for life sciences. Levine comes with over 30 years of experience in the development of computational tools for product innovation, and holds a PhD in Materials Science from Rutgers University. He was elected into the College of Fellows in the American Institute for Medical and Biological Engineering (AIMBE), which represents the top two percent of medical and biological engineers in academia, industry, education, clinical practice, and government. At Dassault Systèmes, Levine mentors startup healthcare companies within 3DEXPERIENCE Labs for Life Science, and participates on several advisory boards. He is also the founder and executive director of the Living Heart Project.
“It started when we began to change our focus as a company from primarily helping build great products, to building sustainable products that make the world a better place,” said Levine. “We called it harmonizing product, nature and life. Our CEO, Bernard Charlès, challenged us to see if we could apply our technology to these broader world topics, and I took on the challenge.
“At the time, I was leading the strategy for our SIMULIA division. I thought, ‘How do we simulate life?’ Having some exposure to the cardiovascular world, I set out with the idea—why can’t we build a human heart? I realized that the biggest challenge was not our ability to understand how the heart works; it was the ability to bring together all the pieces, because the heart is so complex. The body is so complex that we break it down into individual domains, and everybody has their specialty, and they study a very precise piece. And then they publish or gain their knowledge, but no one ever actually puts it together. I gathered doctors, researchers, regulators and industry from all over the world, to work together on a common model of the human body. That became the Living Heart Project, which seemed a little bit of a moonshot back when we started. We’ve been able to succeed and expand to the rest of the body.”
The Living Heart Project
The Living Heart Project involves the development of the world’s first commercially available electromechanical human heart simulation. In addition to aiding innovation of new treatments, by building virtual models of the human heart, Dassault Systèmes hopes to reduce death and disability from heart disease by helping patients make better decisions regarding their medical care through clearer understanding and monitoring of their condition.
“What I think is most profound, is that the heart is available for anyone to use on our platform,” expressed Levine. “Just like any other piece of software, you can simply go to the cloud and model what happens inside a human heart. If you have a new idea of how to cure it or want to understand how it works, you can just go and get it. It’s commercial-grade, and it works on our enterprise-wide platform. That didn’t exist for anyone five years ago.”
So, how did Levine go about creating a working 3D model of the heart?
“We began with the benefit of a really powerful platform,” explained Levine. “We had built enough technology to bring together an army of workers. For example, a commercial jet made by a company like Boeing has a thousand different companies working independently on the same model of the jet. They all have to fit together, while still having to protect their unique intellectual property. We have the infrastructure to be able to manage something of that complexity. I used that as a mental model to say, ‘What if I allowed each of the individual experts to maintain ownership of what they knew, but used our platform to bring it together into a whole, and then gave them back a virtual twin of, in this case, the heart?’
“We took their piece, embedded it into the virtual twin—which embodies everything we know about the function of that twin—and then provided it back to them to say, ‘Does this still truly represent the specialty that you had? Only now you can actually study it in a context of an entire human heart and do amazing things with it, rather than just the one specific thing that you could do before.’ By iterating that process, we rapidly came up with a model that keeps getting better and better.”
Steve first launched the translational project with motivation to support his daughter Jesse, who was born with a rare congenital heart defect where her left and right ventricles are reversed. Since very little reliable data exists for Jesse’s unique case, a true-to-life model of her heart can be used to run simulations of thousands of patients that don’t exist, in order to assess how the disease would progress and how any interventions would affect her heart. This ability to predict outcomes is critical for dealing with a health condition for which, by the time doctors see symptoms, it’s often too late.
“We’re working with clinicians to build virtual patients in clinical environment,” said Levine. “Rather than surgeons guessing with a problem they have never seen before, we can now build the models and apply what we’re learning about the human anatomy. It turns out that the body—when you start to put it into engineering terms—is not as complicated as people otherwise think. We’re getting pretty good at being able to predict how the body will react, at least acutely, meaning post-surgical. We can perform the surgery fairly accurately for congenital defects, reconstructions, and so on. We’re just starting to learn how to predict long-term outcomes, i.e. what happens two years later. That’s a whole other science that we’re beginning to explore, with accelerated aging studies, etc.”
The Living Heart Project is largely used by medical device companies for developing and testing new devices. The U.S. Food and Drug Administration (FDA) is heavily involved.
“The FDA has been a participant in the project now for over five years,” said Levine. “They’re very eager to be able to utilize it, and have funded us to use the Living Heart to create a virtual patient population and perform an in-silico clinical trial. The project is sponsored by Dr. Jeff Shuren, Director of the Center for Devices and Radiological Health (CDRH), who believes that this can be important to the future of their regulatory process.”
By conducting a clinical trial using virtual patients, more information can be gained before running trials on actual humans. The cost and number of human patients can be lowered, and the entire process can be expedited. The most relevant patients can be studied when assessing therapeutic solutions for efficacy. Additional populations (such as children) can be included in simulations, accelerating the availability of treatments to a broader audience while still continuing to adhere to patient safety. A large portion of animal testing can be circumvented, leading to a far more sustainable and humane approach to clinical trials.
“Interestingly, when we conceived the project with Dr. Shuren two years ago, his challenge to us was to see if we could get the regulatory process for a new device from 9-12 months today, processed down to two weeks,” revealed Levine. “He knew you couldn’t do that by just tuning the current system. You need a completely new way of doing it, and he thought maybe the Living Heart was that new way. People didn’t think that two weeks was even possible—except then that the pandemic hit and it became a requirement. The vaccines that we’re all now getting the benefits of, were actually approved in two weeks. What—two years ago—seemed like a dream idea, I think now will eventually become the goal and the norm. And we’re kind of at the center of it, working with the FDA on that.”
The Living Lung Project
After the success of the Living Heart Project, Dassault Systèmes began work on the lung.
“Using the Living Heart Project as a template, we were hoping that other communities would naturally come together and try to do the same thing,” said Levine. “Turns out that we’ve been more successful than I had hoped. Communities have come to us because we already have the formula, technology foundation, expertise, and support from leadership.”
The Living Lung Project was inspired by needs coming out of COVID-19. Early in the pandemic, ventilators were found to be causing damage to lung tissue and exacerbating COVID’s lung inflammation problem. Researchers from University of California, Riverside, Illinois Tech and Pontificia Universidad Católica de Chile (UC) have joined forces with Dassault Systèmes to build virtual lungs in order to enhance their understanding of pulmonary biomechanics and establish oxygenation strategies that would avoid injury. The structural mechanics of lungs are currently being investigated through computational modeling at the organ, tissue and microscopic scale.
“We are working to bring together experts from around the world who understand lung tissue,” described Levine. “The way lungs behave, their function, their connections to the heart. The heart and lung work as a system. It was very much a challenge with COVID, where a lot of people died simply because of the pressure that the lungs put back on the heart. To try to treat that, you need to understand how the two organs are working, both individually and together. In the 3D world, these models fit perfectly together.”
According to Levine, they aren’t at the lung diagnostics stage yet, and are currently more focused on building accurate models. While billions of dollars have been spent on cardiovascular research, not as much has been invested in understanding respiratory problems at the level of the mechanics of the heart. The Virtual Human Modeling team have had to work from the ground up in understanding exactly how the lungs work, and building them up part by part. Early work has been completed on animal lungs, after which it has been converted to human models.
“It became rather difficult to get human lungs to validate on during COVID, because the labs that were studying lungs were at a different level of productivity,” said Levine. “So, that’s where we’re focusing now. Also, since recognizing that respirators actually work against the natural function of the lung, a whole new generation of respirators need to be created. We’re now surfacing our understanding of respirators, and I think we’re going to really change that whole industry.”
The Living Brain Project
Dassault Systèmes has been applying simulation for a range of Living Brain applications.
For instance, the Virtual Human Modeling team is working with the Living Brain community to study damage stemming from chronic traumatic encephalopathy (CTE), which is prevalent in sports. Augmented surgery is being performed on the virtual brain so that clinicians can learn how to avoid sensitive areas such as speech centers. Tumor locations are being investigated along with treatment protocols. Electrical signals are being identified for health conditions such as epilepsy. Methodology is being developed for the early detection of neurodegenerative diseases.
“The challenge with neurodegenerative diseases is that it takes a long time for them to develop,” said Levine. “We need to identify the biomarkers with a degree of reliability—and then we need to wait 10 years before we know whether we’re right. We’re developing those markers and tools to be able to identify them, but we can’t say with confidence that we have them correct. We will over time, though. I think more importantly, we’re building up the knowledge base—so that as more longitudinal data becomes available, we can filter through all of our hypotheses.
“Pharmaceutical companies can still use the technology to track the early impact of new treatments. We’re building the connection between the brain and the muscles by reconstructing brain paths and the muscular skeletal system. Then we can measure the forces on each of the muscles, and the signals coming from the brain to the muscles which get deteriorated in neurodegenerative diseases. After measuring those, muscle by muscle, we can track them over time to see if the forces on any individual muscles are improving based on the treatment. It could be a macroscopic change in the clinical measure, such as the speed at which you can walk 20 meters, or how many squats you can do. So, we’re reverse-engineering the human body to be able to guide better treatments and get them on to the market in a faster way. As opposed to waiting 5-10 years for a drug to assure its effects, which is a long time to wait.”
In order to create a working model of the brain, the Virtual Human Modeling team is coordinating with startups that have monitors.
“A lot of the time when you go to the hospital for a measurement, if the problem doesn’t happen during that one hour of you being tested, the doctors don’t have a good map,” said Levine. “With wearable sensors in a cap, you can get that map. We translate that map into the virtual brain.”
According to Levine, imaging of the brain is more straightforward than that of the heart.
“With a CT scan, you can get the physical structure of the brain in great detail as opposed to the heart, which is very deep within the chest,” explained Levine. “The heart is very well-protected, for good reason. It’s also moving; there’s lots of fluid. It’s dynamic, so its shape changes over time. The brain, fortunately, is very accessible. It doesn’t move and we’re able to reproduce it quite effectively. If there’s a brain tumor, the CT scan will tell you exactly where it is, where the edges are.
“However, mapping the brain function requires other techniques. There are things like a functional MRI, which follows the electrical signals and allows you to map the connectivity of the brain. That is a very complex model and impossible to validate—and there is our biggest problem. If you don’t have the ability to validate a model, it’s hard to know how good it is. So I would say, the mechanics of building a brain is not our challenge. It’s understanding what the reference state is, and validating how the brain really works.”
The Virtual Human Modeling team is beginning to connect with scientists who build connectomes, i.e. logic circuits that make the brain function as a brain.
Levine is cautiously optimistic about being able to finalize a complete map of the brain network over the next 5-10 years, in order to end up with a functioning electromechanical model of the human brain. Significant advancements are already being made in brain research—such as Berkeley University’s recent breakthrough in following an individual thought through the brain.
“If we’re already able to map the specific paths of a thought through your brain, I think we’re on the cusp of creating maps for building patient-specific connectome models,” said Levine. “However, today, we can only build representative brains when it comes to how they function.”
The Living Knee Project
The Living Knee Project has evolved in a different path, with bone being easier to model than soft tissue from an engineering perspective.
“The heart has computational fluid dynamics with the blood,” said Levine. “It has structural mechanics for the tissue, and electrophysiology for the signals. You’ve got to marry all these physics together. The knee, on the other hand, is pretty straightforward. It’s just a structural analysis, and doesn’t have the complexity of as much physics.”
Dassault Systèmes has accomplished a considerable amount of work in reverse-engineering the knee’s ligaments, conducting virtual implant surgery, and carrying out wear testing and accelerated aging studies.
“A wear test that would ordinarily take three months on the bench, we can do overnight,” asserted Levine. “You no longer have to wait three months every time you have a design iteration, to see how it would perform.”
The knee replacement industry has recently been rejuvenated due to a new focus on robotics in surgery. Rather than performing surgery by eye, surgeons are using robots to improve precision.
“You have something that can understand the digital plan,” said Levine. “We produce a surgery on the computer beforehand, that minimizes all of the residual stresses that will be experienced by the knee, minimizes the possibility of pain, and maximizes motion. We then feed it to the robot. Now you have a system that changes the game for knee surgery. It also opens the world up to much more sophisticated knee implants, possibly including active knees that provide additional force for people who don’t have strong mobility. You can really think about a whole new generation of knees simply because we now have the ability to design, test and install them.
“We’re currently focused on connecting the R&D to the clinician and patient—and so, creating that digital continuity which is better for all concerned. Most specifically, this brings the patient into the process because they can now understand if you show them a working 3D model of their knee. You can 3D print it and say, ‘Here’s your knee. Here’s where it was broken. Here’s where we predict you’re going to have pain because this ligament seems to be under the most strain. Let us know how that’s behaving.’ We can provide feedback to the clinician and back to the industry in order to produce better products by getting feedback directly from patients.”
3DEXPERIENCE Platform Supports Workforce of the Future
As mentioned earlier, Levine is working with startup healthcare companies within 3DEXPERIENCE Labs. In one instance, a digital orthopedics startup has created a foot model going off of the Living Knee Project.
Similarly, Dassault Systèmes is leading initiatives to train the next Workforce of the Future in Life Sciences through the use of advanced digital tools.
“We’re installing labs in universities—for example, Illinois Tech,” said Levine. “We’re training biomedical engineers on the commercial tools they’ll be using when, hopefully, they’re sitting hand in hand with clinicians. It’s important to be able to create those pairings, teach them about anatomy, and do it in a way using virtual twins so that they understand [the human body] at the same level they understand engineering problems. I’ve been spending a lot of my time recently on training the next generation, and I’m really blown away by a number of groups. I’m working with maybe 10 universities in mentoring programs of students, including high school students.”
One high school student wants to be a pediatric cardiologist, which is particularly exciting for Levine.
“Inclusion of the patient and family in the process is important to me, particularly for pediatrics,” expressed Levine. “As a parent, your primary responsibility is to protect your children. When you have a problem like this, you’re incapacitated. You feel massive anxiety about not just your own child’s mortality, but your own ability as a parent. The student, Siya Kulkarni from Morris County New Jersey, is helping us find a whole new patient experience using these models, which are realistic and very easy to understand. They enable people to understand these technical problems and talk to their doctors in a knowledgeable way, creating a completely different relationship.
“The patient often gets lost in the process of talking about healthcare, and we’re really excited that these virtual twins basically give the patient more control over the process. Down the road, they can play their own what-if games and make decisions based on actual data, rather than just, ‘My doctor thinks it’s a good idea.’ This is all about putting the patient in the middle—hopefully starting with their twin, and then themselves.”
To learn more about initiatives like the Living Heart Project, click here.