New Tool Detects Radiation Using Routine Blood Tests
Researchers developed a machine learning model that identifies radiation exposure through complete blood count data.
Updated on Oct. 20, 2026 in Life Sciences
Live Poll
Do you believe new medical screening technologies significantly improve safety for the average person?
Scientists at the Lawrence Berkeley National Laboratory have created an open-source tool called Sparse CBC to detect radiation exposure. The system uses machine learning to analyze standard blood test results, offering a faster alternative to traditional biodosimetry methods.
Why it matters
Current radiation detection methods are time-intensive, require specialized facilities, and often struggle to produce clear results for low-dose exposure. This new tool could streamline the process by utilizing data already collected during routine medical screenings.
The researchers validated the model using a total of 1,582 mice exposed to radiation levels ranging from 0.05 to 4 grays. Blood samples were collected for analysis between one and 150 days following exposure to calibrate the algorithm.
The players
Lawrence Berkeley National Laboratory
This Department of Energy national laboratory is a leader in scientific research and advanced technology development.
Intelligence Advanced Research Projects Activity
This agency within the Office of the Director of National Intelligence invests in high-risk, high-payoff research projects.
The details
The Sparse CBC tool employs unsupervised sparse representation learning to identify signals in complete blood count data that correlate with radiation dosage. By testing the technology on genetically diverse mouse populations, the team ensured the model could account for varied biological responses to radiation.
Timeline
2022: The TEI-REX program was established by IARPA.
October 20, 2026: The research results were published in the journal Communications Medicine.
The Big Picture
This development follows the research objectives set by the IARPA TEI-REX program to improve biodosimetry capabilities. The transition from legacy lab-intensive methods to machine learning analysis represents a paradigm shift in how exposure events are managed.
If validated for human use, this tool could allow doctors to screen patients for radiation exposure using standard blood work during routine visits. This would eliminate the need for specialized laboratory testing, providing faster triage in potential exposure incidents.
The takeaway
The development of Sparse CBC marks a significant step toward making radiation detection as simple as a standard physical exam. By utilizing machine learning on common blood count data, researchers aim to remove the logistical barriers that currently limit effective biodosimetry.
Further reading
For more advancements in the field, explore the Life Sciences section.
More information
Access the Open-source model repository to review the project documentation.
Live Poll
Do you believe new medical screening technologies significantly improve safety for the average person?
