A technology has emerged that can view all past and present medical histories with just one blood test. Artificial intelligence (AI) analyzes the genetic sequences of immune cells in the blood to diagnose infectious diseases and health conditions.
Researchers from the University of Cambridge in the United Kingdom and Stanford University in the United States announced on the 21st in the international journal 'Science' that they have successfully determined the presence of diseases such as COVID-19, type 1 diabetes, human immunodeficiency virus (HIV), and autoimmune diseases by analyzing data from nearly 600 participants using AI.
The human immune system stores disease records in immune cells called B cells and T cells. B cells produce antibodies that bind to and neutralize viruses and harmful molecules, while T cells activate other immune responses or eliminate infected cells. When infection or autoimmune diseases occur, the number of B cells and T cells increases, and receptors that recognize antigens of specific pathogens are created. By analyzing the genetic sequences that encode these receptors, one can identify an individual's unique disease records.
Traditional diagnostics relied on physical examinations or patient histories and did not sufficiently utilize the disease exposure records of the immune system. Most have used data from only one of the B cells or T cells. The researchers developed an AI that analyzes information from both B cells and T cells to extract disease-related patterns.
The research team analyzed 16.2 million B cell receptors and 23.5 million T cell receptors from blood samples of 593 participants using AI. As a result, they were able to classify the medical history and health status of COVID-19, type 1 diabetes, HIV, and lupus with an accuracy of 98.6%. The genetic sequences of T cell receptors clearly indicated the history of type 1 diabetes and lupus, while the history of COVID-19, HIV, and influenza vaccination was better identified in B cell receptors, the researchers explained.
The research team noted, 'This technology could help address diseases for which there is currently no definitive testing method,' while also emphasizing that 'although the AI performs well in detecting diseases, additional validation is necessary for real-world medical application.'
References
Science (2025), DOI: https://doi.org/10.1126/science.adp2407