A decade. That’s how long it took for Dassault Systemes to build and create a ‘virtual twin’ of the human heart as part of its Living Heart project. As the company showcased at the recently-concluded CES in Las Vegas, this ‘virtual heart’ is now saving lives thanks to advances in Artificial Intelligence (AI) and machine learning models.
For the French company, creating several ‘virtual twin’ versions of complex machines such as automobiles, aircraft, etc is a routine task and it has been doing this for decades. But creating a virtual twin of the human heart was a different and complex challenge.
“The way that treatments are developed, they don’t have the equivalent of a virtual twin of a patient to be able to say well, let me test it on the virtual twin. So, we took on the challenge to create virtual twins of humans, which is a very complex topic. We chose to start with the human heart for a number of reasons, not least of which is that heart disease remains the number one cause of death worldwide,” Dr Steve Levine, Senior Director, Virtual Human Modeling, Dassault Systemes, explained to indianexpress.com over a call.
Dassault’s project brought together researchers, experts, and doctors in the field to help create this 3D ‘virtual twin’. The project now has over 150 different organisations in 49 countries working together. The 3D model of the human heart is “a physics-based” one, and “operates the same way a real human heart.” Doctors can even introduce a heart disease and run simulations on this ‘virtual twin’ and see how it would behave to the complications. With virtual reality now catching up and getting more powerful, researchers can even go inside and experience this heart and look for any defects. The researchers have “complete flexibility to introduce anything that might happen in the heart,” Dr Levine pointed out.
In India, companies like Sahajanand Medicals are using Dassault’s technology to help create personalised stents at less than half the price. The ‘virtual twin’ heart is also being used by drug companies and medical device manufacturers to see how the heart would react to their medicines and instruments.
He gave an example of how the 3D ‘virtual twin’ is used by the Boston Children’s Hospital to treat complex cases and how the model has correctly predicted the kind of surgical procedures that doctors need to undertake. In one such case, a toddler was dying because all the oxygen in the blood was flowing to one lung and doctors were not sure of the procedure to follow.
“Dr David Hoganson at the hospital was able to build an exact virtual twin of that patient. They actually used the twin to design an experiment or a procedure that would work. The model found the optimal surgery and then when they did it, it worked perfectly exactly as predicted,” Dr Levine explained.
“The real challenge is to be systematic and measure the right data. The heart is essentially an electro-mechanical pump. We understand the electrical and mechanical system, the pumping, but the mechanics of how nature has constructed it very carefully have to be replicated otherwise the behaviour won’t be the same,” Dr Levine said.
More specifically, the ‘virtual’ heart can be customised for each patient to ensure optimal results. “The Artificial Intelligence algorithms can now know about what’s happening to the patient, as well as what’s happening to the clinical measurements that they get,” Dr Levine said. Also, the nature of these cloud-based platforms is that they can actually communicate and work anywhere. So, technically doctors in any country can benefit from the data and procedures created using this system.
Dassault is also working with the US FDA to build entire clinical trials on virtual humans, which could allow treatments to be screened on the virtual body and see how the medication works early on. But Dassault is not just limiting the ‘virtual twin’ creation to the human heart. It is also trying to do the same with the human brain and liver, Dr Levine said.
However, there are challenges. “Other organs like the brain, there are things we know very well, but other parts are still a complete mystery. So, it’s very difficult for us to model something we don’t understand because we will never know if we get it right,” he said.
The liver is another organ the company wants to create a virtual twin for, especially because it would help pharmaceutical companies understand how a drug could impact the liver since drug toxicity is a major challenge. New tools like AI are also helping them “make sense of very complex datasets.”
“We don’t need a million patients to know how something functions. So, we only need a handful that is representative. But we have to find those exact right patients and collect data, then we can test them against a million patients. That’s important,” he said.
While Dr Levine is confident they will have a working ‘virtual liver’ within a year, he predicts getting the full functions to work could take another decade.
Of course, there’s also a significant cost involved when such systems are being developed. For Dassault, the ‘business challenge’ is also convincing hospitals and insurance companies that such technology can help bring down healthcare costs in the long run.
“First, you have to spend the money to prove you’re going to save money. We’re at that stage now. Now in Boston, because they’ve been doing that for three years, they have enough data The outcomes are universally better, the teams are more prepared, and repeat surgeries have gone down dramatically,” he pointed out.