If a digital copy of your heart or another organ were stored on a hospital supercomputer and could evolve alongside changes in your actual health, a doctor could use this type of personalized model to make custom-tailored decisions about treatments in real time. This “digital twin” could help you prevent diabetes, cardiovascular disease, or even cancer.
While researchers are on the cusp of creating computer models for entire organs, precision healthcare applications using digital twins are still theoretical. Karen Willcox, an SFI External Professor and University of Texas aerospace engineer, is helping to convene a National Science Foundation-sponsored workshop at SFI, October 12–13, to make this new type of modeling technology a reality.
What sets digital twins apart from conventional models and simulations is the dynamic interaction of data between the physical and virtual environments. Sensor data or remote sensing from the physical system is assimilated into the virtual model, causing it to evolve and adapt. The updated virtual model can then provide recommendations on how to improve the physical system. This feedback loop creates a continuous cycle of optimization, which could ultimately be used to improve everything from airplane fuel efficiency to natural disaster forecasting and personalized medicine. In fact, since NASA coined the phrase “digital twin” in 2010, scientists have found a variety of ways to put the technology to use in improving drones, spacecraft, and other mechanical systems.
“The scientific community has been building mathematical models and simulations of complex systems now for decades and it has really changed our understanding of engineering systems, the natural world, and medical outcomes,” Willcox said. “What digital twins now enable us to do is personalize these models in a way that has never been possible before.”
However, several barriers impede the realization of digital twins for more computationally intense activities. Our computing capabilities and modern algorithms are nearing the ability to model an entire human organ, but they remain distant from modeling complete human beings or entire planetary ecosystems. Additionally, uncertainty and trust on the part of decision-makers and other stakeholders are pivotal concerns when deploying digital twins, particularly when informing critical decisions involving things like human healthcare and urban infrastructure.
Willcox said her hope is that the workshop, “Crosscutting Research Needs for Digital Twins,” will help to start addressing these challenges by combining mathematical and modeling expertise from the wide array of subject matter experts.