Tthe Food and Drug Administration on Friday issued an emergency permit for a new test to detect Covid-19 infections – one that differs from the hundreds already authorized.
Unlike tests that detect pieces of SARS-CoV-2 or antibodies to them, the new test, called T-Detect COVID, looks for signals of past infections in the body’s adaptive immune system – especially the T cells that the body uses. helps remember what his viral enemies look like there. It was developed by Adaptive Biotechnologies in Seattle and is the first test of its kind.
Adaptive’s approach involves mapping antigens to their corresponding receptors on the surface of T cells. They and other researchers have already shown that the casting of T cells that float in an individual’s blood reflects the diseases they encountered, in many cases years later. The next step is to unlock the information to diagnose the infections in the past.
advertisement
The challenge is extremely data-heavy. “When you think of a patient level, we look at an average of 300,000 to 400,000 T cells,” says Lance Baldo, Adaptive’s chief medical officer. ‘If you look at a population level, we look at hundreds of millions and ultimately billions of T cells. It is therefore becoming a problem on the internet. ‘
Enter Microsoft. In 2018, Adaptive developed a partnership with its technology giant neighbor to build the cloud infrastructure and machine learning models needed to handle the amount of data – in particular to build a complete map of which T-cells to which antigen bind.
advertisement
“Microsoft wants and wants to get into healthcare,” Baldo said. ‘Adaptive requires expertise in cloud computing, machine learning and AI. So it was pretty ideal. Teams from both companies worked together one day a week in Adaptive’s office in Seattle or Microsoft in Redmond.
When the virus began to increase rapidly, they quickly turned a large part of the team to work on Covid-19. By June, they were able to access blood samples from people infected with the coronavirus and sequence the genomes of the T cell receptors therein. Then they were able to compare the data set with their control group – the database of T-cell receptor sequences they had been working on for years – and within two months they had collected enough data to publish their first results.
The machine learning models needed to develop the T-Detect test were ultimately relatively simple. “For me, it’s actually a big plus,” says Jonathan Carlson, senior director of immunomics at Microsoft and leader of the partnership with Adaptive. ‘It’s a viral infection and causes an angry T cell response, and it turns out you can find exactly the same T cell receptors in many people. And it allows you to use a fairly simple statistical approach. ‘The test reported a sensitivity of 97.1% and 100% specificity.
The EU-issued EUA reflects the first approach – but it is not the end of the evolution of the test. “When we submit to the FDA, we do something called ‘closing the classifier,'” says Baldo, the algorithm that determines whether the T-cell receptors of a blood sample say ‘Yes Covid’ or ‘No Covid’. “
However, the reactions of the T cell may depend on the version of a virus to which you have been exposed.
“We’ve already discussed this with the FDA,” Baldo said. “You have mutations and other variants coming.” Adaptive and Microsoft therefore continue to improve the classifier. “The models are getting better regularly,” Carlson said. “Weekly, monthly.” The question that then remains is: ‘when is it better? enough? This is where Adaptive spends a lot of time thinking. ”
At some point, when the test reaches a new threshold of sensitivity and specificity, they plan to submit a second version of the test for the FDA’s review.
This is just one of Adaptive’s three focus areas in the coming months, Baldo said. “One of the pillars is to improve the current algorithm and make sure we’re still going to keep a good test, as the virus is still changing,” he said. The second is to focus the expertise of the company’s T-cell on other questions surrounding Covid-19, including the effects of long Covid and the efficiency and durability of the immune response elicited by different vaccines.
The third is to continue work on other diagnostic diseases for celiac disease and multiple sclerosis. Prior to the pandemic, the company was focused on developing evidence of conceptual diagnosis for Lyme disease, which it announced in November 2019.
The diffused focus will force the company to build not only its biological capabilities, but also its machine learning approaches. Although the approach to Covid-19 screening is relatively simple, Carlson said, “I do not expect it to work for every disease.”