Creating a virtual twin of the human body
30-06-2021 | Posted by Joaquín Martí
The human body is composed of molecules organized in cells and tissues delivering physiological functions at the macro scale. But, unlike machines, there is no design blueprint to consult. What we understand is largely the result of deconstructing complex systems into smaller, functional parts and running trial-and-error analysis of their behaviour. This has allowed us to make incredible medical advances, but we are reaching the practical limits of traditional approaches.
We know that DNA encodes our inherited genetics but, while we can now identify the body’s coding, we still don’t understand its language. Also, health data is being collected at an unprecedented rate, but each clue is of limited value without context. Can we create a virtual representation of a human body? Could we apply each individual’s attributes and history to adapt this virtual human into a digital surrogate, reduce medical uncertainties and find the optimal approach for each person?
Dassault Systèmes is trying to make a virtual twin of a human body. It started in 2014 with the Living Heart Project, which gathers cardiovascular researchers, educators, medical device developers, regulatory agencies, and practising cardiologists on a shared mission to develop and validate highly-accurate, personalized digital human heart models.
The project involves 67 leading research institutions, 46 industrial organisations, and 15 clinical establishments, apart from two regulatory bodies. The collaboration with the U.S. Food and Drug Administration was recently extended for another five years to spur innovation in the design of devices. This second phase uses virtual patients based on computational modelling to improve efficiency of clinical trials for new designs.
Encouraged by the success of the Living Heart Project, Dassault Systèmes has embarked in other virtual-twin initiatives, like the Living Brain Project. Having a virtual twin of the brain is promising in neurology, where uncertainties are high and errors catastrophic. Indeed, modelling the brain can provide critical guidance for mechanical or electrophysiological interventions.
In acute brain diseases such as epilepsy, personalized virtual twins are already helping to guide diagnosis and treatment. At the onset of seizure symptoms, the signals from the brain can be captured, diagnosed and tracked as care is administered. Since 2019, a clinical trial is using virtual twins of the brain constructed from patient imaging and electrophysiological data. The trial will validate virtual brain technology for surgical decisions and to suggest improvements.
With the lessons learned about the heart and the brain, digital twins are being used to model other parts of the body. While having virtual twins of organs such as the lungs and kidney, and body parts such as feet, appear frequently in the scientific literature, three nascent uses for virtual human twins are the skin, the cells and the gut.
Dassault Systèmes develops solutions that provide a multiscale model of skin, based on its fundamental molecular properties, that can accurately predict the penetration of chemicals through skin layers. Virtual models can reflect individual skin types and properties such as age, hydration levels or damage from exposure, and can test cosmetic creams and lotions.
The myriad of cells in the body are ultimately responsible for its functions. Modelling can provide critical insights by reproducing cells that can be studied for a variety of diseases. It is not yet possible to map all cells and personalise them in a virtual twin, but it is feasible to model some cells known to impact patient health.
Finally, over the last three years, Dassault Systèmes has integrated existing knowledge of the interaction between the gut microbiota and the human body into a physiologically-based kinetic (PBK) model describing the cycling of key metabolites through organs of the digestive system. It is used to understand better the role of microbiota, enriching the human twin with a functional, physiological layer, completing the 3D physics, fluidics and electrical parts.
Harnessing data from medical records and collating insights and intelligence from various sources – industry, research, practitioners and even patients – can help to prevent and battle disease. The data must be stored on a secure, cloud-based platform accessible by anyone, anywhere. With it, we can create virtual twins of the human body to visualize, test, understand and predict, from the way drugs affect a disease to surgical outcomes, before a patient is treated. The virtual twin can initially be built to represent most, then personalized for each patient. The time has come to act.
