Page 56 - 23_EETE_03

56 EE|Times EUROPE

DIGITAL TWINS
Human Digital Twins Are Set to Revolutionize

Medicine

By Rebecca Pool
Digital replicas of entire humans are under development and could be personalizing

healthcare sooner than you think.

e it for a Siemens motor, a Tesla car
or the entire city of Shanghai, digital Dr Caroline Roney
twins are being used far and wide to Imaging Atrial fibres Atrial fibrosis AF simulation Critical regions
Bmimic real-world objects, systems
and processes. A new digital twin is now
entering this tech space: you.
To many, real-time virtual models of
humans will sound just a little too sci-fi, but
the science and engineering exist to create
them, and healthcare companies are already
hooked. For example, Neuroelectrics
Barcelona is heading up an EU-funded,
multimillion-euro project, Neurotwin, to
develop personalized brain models based on
neuroimaging data from Alzheimer’s patients,
looking for ways to restore healthy brain
dynamics. And as part of a huge international
endeavor called the Living Heart Project,
France-based Dassault Systèmes and many
partners are developing digital human heart Atrial digital twins from Caroline Roney and colleagues: Personalized models
models to provide personalized cardiac- are constructed from MRI images (far-left column), and structural and electrical
patient care. personalization is applied (subsequent columns). Patient-specific computer simulations
As virtual models of human organs are used to investigate how the researchers think atrial fibrillation might take place in the
proliferate, one researcher who is already patient and to test different treatment approaches. (Source: Caroline Roney)
connecting the dots is Gunnar Cedersund of
the Biomedical Engineering Department at
Sweden’s Linköping University. He believes digital twins will be critical flow to form multi-organ models that can analyze the organs as well as
to the future of healthcare and is certain that the time is right to bring the related biochemical processes.
computer copies of people to life. One key process under study is metabolism. When you eat a meal,
As Cedersund points out, care of the elderly is set to explode in most sugar reaches your bloodstream from the intestines. Your pancreas
Westernized nations, with Type 2 diabetes, cardiovascular disease and detects this process and produces insulin that will reach other organs,
obesity-related illnesses stretching health services to the limit. At the which may change their metabolism in response. Cedersund and
same time, the world is witnessing an explosion in data as myriad con- colleagues formulate equations to represent these processes, which are
sumer biotech and digital health tools come online and as remarkable simulated in their models and compared against actual experimental
developments in modeling and machine learning proceed apace. data from clinical studies to deliver the likely human response. Com-
“We’re seeing this increasing need for digital twins alongside pared with machine learning, such a mechanistic model yields a more
increasing possibilities, and we believe we can make [models] that are general artificial intelligence, as it can incorporate a wide variety of
appealing, intuitive and useful,” Cedersund said in an interview. data, including data not used for developing the model.
With digital twins, he said, “we should be able to analyze this data For example, Cedersund said, “we’ve put a detailed liver metabolism
and find the best way to either prevent disease or treat a person. Inter- model together with a meal model that also contains data on metab-
active computer copies of people will allow you to look inside and see olism from other organs. Earlier models haven’t been able to describe
what’s going on but also take a look at what will happen if they take a the transformation from proteins to sugars and have therefore not been
certain medication, change their diet or start to exercise.” able to simulate the effects of changing diets.”
Pharmaceutical-industry giants are interested. As part of his
MAKING MODELS research, Cedersund has been working with AstraZeneca to under-
Earlier learning-based models developed by Cedersund and colleagues stand how Type 2 diabetes drugs can reduce the risk of cardiac disease.
use data on height, weight, blood pressure and fasting glucose levels Using data generated by the British-Swedish multinational’s in vitro
to predict a person’s risk of stroke. However, the researchers have also organoid experiments, he creates a computer copy and runs his models
created sophisticated mechanistic models of most of the main organs to understand processes and predict outcomes. Traditional drug testing
in the human body, including the heart, liver, brain and pancreas, as costs billions of dollars, but the mathematical models may provide a far
well as adipose and muscle tissue. And intriguingly, they have con- cheaper investigation route and could even form the basis of future—
nected these individual organ models with existing models of blood and faster—certification processes.

MARCH 2023 | www.eetimes.eu