Adsilico
AI can produce many versions of a digital heart
This is the first in 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 no blood flows through it, nor does it live in a human body.
It is a computer-generated heart, or digital twin, used to test implantable cardiovascular devices, such as stents, and prosthetic valves that, once confirmed to be safe, will eventually be used on real people.
But the creators of the core, Adsilico, went beyond just creating an accurate model.
Using artificial intelligence and huge amounts of data, they created several different cores.
These AI-generated synthetic hearts can be designed to reflect not only biological attributes such as weight, age, gender and blood pressure, but also health status and ethnicity.
Because these differences are often not represented in clinical data, digital twin hearts can help device makers conduct trials in more diverse populations than they could with human trials or trials involving only digital twins without AI.
“This allows us to capture the full diversity of patients' anatomy and physiological responses, which is not possible with conventional methods. This use of AI to improve device testing leads to the development of more inclusive and secure devices,” says Sheena Macpherson, Managing Director of Adsilico.
In 2018, an investigation by the International Consortium of Investigative Journalists found that 83,000 deaths and more than 1.7 million injuries were caused by medical devices.
Ms Macpherson hopes AI-powered digital twins can reduce these numbers.
“To really make these devices safer you need to test them more thoroughly, and that's not feasible in a clinical trial environment because of the cost,” says Ms Macpherson, based in Northumberland.
“So you want to be able to use the computer-generated version, to make sure that whatever you're doing, you've tested it as thoroughly as possible before you test it on a human.
“Even a fraction of these deaths – and associated lawsuits – could have been avoided with more extensive testing. You can also get more detailed results.
“You can take the same (virtual) heart and test it under low or high blood pressure, or against different disease progressions, to see if that affects the device in any way.”
Macpherson adds: “(Virtual) testing gives medical device manufacturers much more information. It also means we can test other subgroups of patients, not just white men on whom clinical trials are traditionally based.
Getty Images
AI can detect patterns that humans might miss
Adsilico's AI models are trained on a combination of cardiovascular data and data from real MRI and CT scans, which include medical imaging from consenting patients.
The data draws on detailed anatomical structures of the heart, to help create accurate digital representations of how medical devices will interact with different patient anatomy.
Adsilico's testing involves creating a digital twin of the device under test, which is then inserted into the virtual core in an AI-generated simulation.
It all takes place in a computer, where the test can be replicated on thousands of other cores – all AI-simulated versions of a real human heart. In contrast, human and animal trials typically involve only hundreds of participants.
Perhaps the biggest incentive for drug and device makers to supplement clinical trials with AI digital twins is how it reduces the time needed, which also translates into significant cost savings.
Drugmaker Sanofi, for example, hopes to reduce the testing period by 20%, while increasing the success rate. She uses digital twin technology in her specialization in immunology, oncology and rare diseases.
Using biological data from real people, Sanofi creates AI-based simulated patients – not real clones of specific individuals – that can be divided between the trial's control and placebo groups.
Sanofi's AI programs also create computer-generated models of the drug to be tested, synthesizing properties such as how the drug would be absorbed in the body, so that it can be tested on AI patients. . The program also predicts their reactions, thus replicating the actual trial process.
Sanofi
Using digital twins could mean big savings for pharmaceutical companies, says Matt Truppo
“With an industry-wide 90% failure rate for new drugs during clinical development, just a 10% increase in our success rate through the use of technologies such as digital twins could result in savings of $100 million, given the high cost of performing late clinical phases. trials,” says Matt Truppo, global head of research platforms and IT R&D at Sanofi.
So far, the results are promising, adds Mr. Truppo, based in Boston, United States.
“There is still a lot to do. Many of the diseases we are currently trying to tackle are very complex. This is where tools like AI come into play. Powering the next generation of digital twins with accurate AI models of complex human biology is the next frontier.
PA consultancy
AI is only as good as the datasets it is trained on, says Charlie Paterson
Digital twins could have weaknesses, however, says Charlie Paterson, associate partner at PA Consulting and former head of NHS services.
He emphasizes that the quality of twins depends on the data they are trained on.
“(Due to) outdated data collection methods and low representation of marginalized populations, we could find ourselves in a position where we could still introduce some of these biases when we program virtual recreations of individuals.”
Working with limited existing data to train its AI is a problem that Sanofi is aware of and working to solve.
To fill gaps in its internal data sets – made up of millions of data points from thousands of patients undergoing its trials each year – it sources data from third parties, such as electronic health records and biobanks.
Back at Adsilico, Macpherson hopes that one day AI digital twin technology will eliminate animal testing from clinical trials, which is still currently considered an essential part of the drug and device testing process .
“A virtual model of our heart is even closer to a human heart than that of a dog, cow, sheep or pig, which tend to be used for device studies implantable,” she says.
More business technology
Source link
