In August 2020, Gartner named “digital me”—digital models representing humans in both real and virtual worlds—a must-know technology that will significantly affect society and business during the next five to 10 years.
But in healthcare, this technology is already being put to meaningful use.
What is a virtual twin?
A virtual twin is a three-dimensional, analytic model built to represent the physical size, shape and behavior of the original, constructed and operated using data that often come from sensors and connected devices. Once built, the model allows for trial and error experimentation to determine how to achieve a precisely desired outcome.
Virtual twins are typically used for physical objects such as planes and cars, where operating conditions can be readily measured. Before those machines are built, even before one screw or piece of metal is soldered, the manufacturer first builds a virtual twin using software. These digital copies allow them to rapidly see how a design will perform in different real-world scenarios and make changes as needed.
Most importantly, these twins are ideal for solving problems that would be impossible, unethical or simply too expensive to explore in the real world such as space travel or, as we now see, the inside of the human body. If the virtual twin of a car can perform unlimited numbers of virtual crash tests and produce outcomes that replicate its behavior in a real-world crash, could virtual human occupants replace crash test dummies to take safety even further?
A few years ago, the manufacturers of this software began to think about the potential of a virtual twin of a human. If possible, it would be a game changer for industries familiar with virtual twins. But some went further, why not create virtual patient twins to guide complex medical treatment? The benefits are clear, but there are important differences. Machines are constructed using standardized materials and methods that are accepted across their industries. Nothing goes in that isn’t already understood.
The human body is very different. When a medical problem arises, there is no functional diagram or blueprint to consult. What we now know is the result of deconstructing massively complex body systems into smaller, functional parts and then running trial and error analysis of their behavior. Outcomes are reverse-engineered to imply function and updated as we learn more. This methodology has allowed us to make incredible medical advances, but as in other industries, we are reaching their practical limit. We now face unsustainable hurdles to continue to understand—and ultimately optimize—the human body.
Modeling the human heart
No organ is more important or the subject of more study than the heart. Yet our knowledge base is massively fragmented and largely unusable by the standards of the digital era. The “Living Heart Project,” or LHP, was launched in 2014 to bring this knowledge together with the hope of creating a virtual twin of the human heart. Using a novel crowdsourcing approach, the LHP called on an array of experts from research, industry and medical practitioners to bring together their data and knowledge to create a virtual twin of the human heart. Project members adapt the model for their research, systematically building its robustness with each of their findings. Shortly after the launch, the FDA joined the project as part of their mission to pursue in silico medicine.
Within a few years, the project delivered the first reference model of a functioning, beating heart able to reproduce virtually any cardiovascular condition and safely test treatment options. Today, the project includes more than 150 member organizations across 24 countries. Medical device manufacturers such as Boston Scientific test new devices while doctors at top hospitals including Boston Children’s Hospital are using virtual twins as a way to plan surgical procedures with more predictable outcomes. Pharmaceutical companies can understand how cellular level drug interactions will ultimately impact the function of the entire organ. Because it is virtual, it can be safely interrogated, shared and visualized to tell the complete story of its function and likely outcomes. As an open reference platform, technical details and knowledge that were once fragmented and isolated are centralized, actionable and sharable.
But, the LHP delivered more than a functional model of the heart, it also served as a proof of concept. The LHP demonstrated that by using real-world experience of medical practitioners as the basis of the model, computational biologists and biomedical engineers can replicate the iterative process necessary to produce a variety of functional organs in the virtual world.
Transforming healthcare and personalizing medicine
Over the past seven years, researchers and doctors around the world have realized the power of virtual twins to transform healthcare and medical research. The Living Brain is being used to learn more about patient specific treatments for epilepsy and the progression of neurodegenerative diseases. An innovative startup created a virtual foot for personalized surgical treatments that consider growth and remodeling during the healing process. A recent research paper showcases the advantages of a digital twin for knee implants with patient-specific motion walking and squatting data.
As COVID-19 has rippled through the world, researchers are now using these digital tools to understand the long-term effects of the virus—with a virtual twin taking center stage. A new Living Lung Project has kicked off, and researchers at UC Riverside and Pontificia Universidad Católica de Chile are conducting groundbreaking analysis on COVID-19’s impact on the lungs, improved treatments and accelerated aging. What began as a useful model of the heart is now recognized as the first step toward the complete virtual twin of a human.
Once it was considered impossible to build a twin of an entire commercial jet under all possible flight conditions. Now it is routine. The human body poses some unique challenges, but these projects demonstrate that if we collectively commit to developing them, virtual twins can deliver the holy grail for medicine; personalized, precise, successful medical treatments—not by themselves, but by providing, in parallel, a living, breathing medical record that combines the latest fundamental knowledge with the patients exact history and unique physiology. One day, these virtual patients could be the medical “crash test dummies” for new innovations, acting as surrogates to replace animal testing and lower the burden on human testing. This vision is so compelling, academia, medical device and pharmaceutical companies, clinicians and regulators are already joining together to share their piece of the puzzle. Their collaborative information sharing produces constantly improving digital models, more validated uses and more targeted treatments.
The recent COVID-19 pandemic taught us that if carefully protected medical data, collected throughout the world are shared, new and novel treatments can be brought to patients in 1-2 years instead of 10-20. The virus provided a singular reference for all to share. More broadly, virtual twins hosted on a secure, cloud-based platform can serve as that reference for any human biomedical process. Drawing anonymized data from medical records, research repositories and collating them with physician insights, industry, researchers and practitioners can use tools such as artificial intelligence to unlock the secrets to battle disease and speed healing.
Medicine should always be a balance of art and science. Physicians must carefully hone their art through years of study and practice while they rely on the medical community to deliver the best science in a form they can use. The virtual human twin is that form. It is the most efficient translational tool for research, the key to truly personalized medicine and the only way to equally serve the doctor, patient and entire innovation ecosystem. I’m happy to report, we’re closer to that goal than you think.
Steven M. Levine, Ph.D., is the senior director of virtual human modeling at Dassault Systèmes.