Imagine if a simple blood test could provide accurate and early diagnosis of thousands of different diseases simultaneously? That is no easy task, but it may be possible using new technology that can leverage how our body naturally detects diseases using nature’s most finely-tuned diagnostic—the adaptive immune system.
At Adaptive Biotechnologies, we just launched T-Detect COVID, the first-ever T-cell test for individuals that can detect whether a person has had a recent or past COVID-19 infection. The development of this test was made possible by combining Adaptive’s understanding of how the immune system works with Microsoft’s AI and machine learning capabilities. Importantly, this is the first of many potential diseases that we can detect by looking at the T-cell response and represents a paradigm shift in how we diagnose, and ultimately treat, many different illnesses based on how the immune system naturally does this. We call this Immune Medicine.
The adaptive immune system: a massive but solvable machine learning problem
T cells are the first responders of the adaptive immune system and activate the antibody response. These cells have specialized receptors which must be extraordinarily diverse to recognize one or a small number of the millions of antigens to which our bodies are continuously exposed.
To understand how the immune system responds to disease, we are translating the diverse genetic code of these receptors into data so that it can be analyzed both at the individual level and across entire populations. This is no easy task, and that is why we are collaborating with Microsoft. With their advanced AI and machine learning technologies, we have the ability to comb through this massive data set, effectively creating a “map” of these immune cells to most diseases. Ultimately, we have translated the complex biology of the adaptive immune system into a machine learning problem which is massive, but tractable.
When we partnered with Microsoft, we recognized that solving a challenge of this nature had never been done before. However, the result would be a true breakthrough—a simple blood sample that would reveal what diseases the body is currently fighting or has ever fought.
Rising to the challenge: COVID-19
When COVID-19 hit in 2020, we knew we could help those impacted by COVID-19 by applying these capabilities to the pandemic, collecting thousands of blood samples to map the T-cell response to SARS-CoV-2, the virus the causes COVID-19.
One of the limiting factors of traditional antibody tests—blood tests that look for antibodies to SARS-CoV-2—is that antibodies wane over time, meaning after a while you may not be able to detect whether you have previously had the virus. While T cells also wane over time, they stay in your body considerably longer than antibodies.
Testing T-cell levels in response to COVID-19 may therefore be an important complement to antibody testing to identify recent or prior infections. And, importantly, T cells may also help us better understand one’s immunity and protection either from natural infection or from a vaccine. This is especially important as new variants appear and the population starts to get vaccinated.
Just the tip of the iceberg
T-Detect COVID provides proof for our ultimate vision to enable early and accurate detection of many diseases from a single blood test – in fact we are already validating T-Detect for many other diseases. To date, we have confirmed two other signals: Lyme disease and Crohn’s disease, which are at advanced stages of development in our pipeline. We are also evaluating multiple tumor types for early and accurate detection of cancer in our T-Detect pipeline. As we continue in this endeavor with Microsoft, we are rapidly learning more that will enable us to use T cells to diagnose thousands of other diseases. Understanding the adaptive immune system is not only an important step in diagnosing many different diseases but also may ultimately help to develop novel, immune-based treatments.
T-Detect COVID is available for research use only. T-Detect COVID is not available for clinical use.