Why ‘digital twins’ could accelerate drug discovery

This is the first part of a six-part series on how AI is changing medical research and treatment.

The heart in front of me beats and moves like a human organ, but it does not flow blood, nor does it reside in a human body.

It is a computer-generated heart, or digital twin, that is used to test implantable cardiovascular devices, such as stents, and artificial valves that, after being proven safe, are eventually used on real people. will go

But the creator of the heart, Adseliko, has gone beyond just creating an accurate model.

Using artificial intelligence and massive amounts of data, they have created several different hearts.

These AI-generated artificial hearts can be designed to reflect not only biological attributes such as weight, age, gender and blood pressure, but also health conditions and ethnic background.

Because these differences are often not reflected in clinical data, digital twin hearts can help device manufacturers conduct trials in more diverse populations than human trials, or trials involving only digital twins without AI.

“This allows us to capture the full diversity of patient anatomies and physiological responses, which is not possible with traditional methods. This use of AI to enhance device testing is the first step in the development of such devices,” says Sheena McPherson, chief executive of AdSilico. leading to developments that are more comprehensive and secure.

A 2018 investigation by International Consortium of Investigative Journalists It was revealed that 83,000 deaths and over 1.7 million injuries were caused by medical devices.

Ms McPherson hopes AI-powered digital twins can reduce those numbers.

“To make these devices really safe, you need to test them more thoroughly, and it’s not possible to do that in a clinical trial environment because of the costs,” says Northumberland-based Ms Macpherson.

“So you want to be able to use the computer-generated version, to make sure that whatever you’re doing, you’ve done it as well as possible before you try it on a human. Checked.

“Even a fraction of these deaths — and the associated litigation — could have been avoided with more thorough testing. You can also get more detailed results.

“You can take the same. [virtual] The heart and you can test under low or high blood pressure, or against the progression of various diseases, to see if that affects the device in any way.”

Ms Macpherson added:[Virtual] Testing provides a lot of insight to medical device manufacturers. It also means we can test in other sub-patient groups, not just the white males that clinical trials have traditionally focused on.

Adsilico’s AI models are trained on a combination of cardiovascular data, and data from real MRI and CT scans, including medical imaging with consent from patients.

The data is derived from detailed anatomical structures of the heart, to create an accurate digital representation of how medical devices will interact with the patient’s various anatomies.

Adsilico’s trials involve creating a digital twin of the device to be tested, which is then inserted into a virtual heart in an AI-generated simulation.

It all happens inside a computer, where the test can be replicated in thousands of other hearts — all AI-simulated versions of a real human heart. On the other hand, human and animal trials involve only hundreds of participants.

Perhaps the biggest incentive for drug and device makers to complete clinical trials with AI digital twins is how it reduces the time it takes, which also translates into huge cost savings.

For example, drugmaker Sanofi hopes to cut testing times by 20 percent, while increasing success rates. It is using digital twin technology in its immunology, oncology and rare disease expertise.

Using biological data from real people, Sanofi creates artificial AI-based patients — not actual clones of specific individuals — that can be matched into control and placebo groups within a trial.

Sanofi’s AI programs also create computer-generated models of the drug to be tested, simulating properties such as how the drug will be absorbed throughout the body, to be tested on AI patients. could The program also predicts their reactions – simulating the actual test process.

“With a 90% failure rate in the new drug industry during clinical development, just a 10% increase in our success rate using technologies like digital twins could result in $100 million in savings, because Given the high cost of conducting late-phase clinical trials,” says Matt Turpo, Sanofi’s global head of research platforms and computational research and development.

Mr Turpo, who is based in Boston, US, adds that the results so far have been promising.

“There’s still a lot to do. Many of the diseases we’re trying to tackle now are extremely complex. That’s where tools like AI come in. Digital twins with accurate AI models of complex human biology. Empowering the next generation is the next frontier.”

Digital twins can have weaknesses though, says Charlie Peterson, an associate partner and NHS service manager at PA Consulting.

He explains that twins are only as good as the data they are trained on.

“[Due to] With aging data collection methods, and the underrepresentation of disadvantaged populations, we may be reaching a point where we may still introduce some of these biases when we’re programming individuals’ virtual entertainment. “

Working with limited legacy data to train its AI is a problem Sanofi is aware of and working to address.

To fill in the gaps in its internal data sets — containing millions of data points from the thousands of patients who undergo its trials each year — it pulls data from third parties such as electronic health records and biobanks.

Back at Edsilico, Ms Macpherson is optimistic that one day AI digital twin technology will eliminate animal testing from clinical trials, which is currently considered an essential part of the drug and device testing process.

“A virtual model of our hearts is still closer to a human heart than a dog, cow, sheep or pig, which tends to be what they use for implantable device studies,” she says.