AI tool scored follow-up brain scans more often correctly than the first reports
Original title: Agentic Automation of BT-RADS Scoring Using an End-to-End System for Standardized Glioma Follow-up Assessment.
How far along is this research?
- Lab cells
- Animals
- Review
- Tested in people
This was tested in people. That is the most reliable kind of research we share.
It was tested on patient scans already collected, not yet used in real care.
The short version
Researchers built AI software that scored old brain scans more accurately than the first clinic reports did.
What was studied. After treatment for a brain tumor called glioma: A tumor that starts in the glial cells. Those are the support cells of the brain and spinal cord. Gliomas are given a grade from 1 to 4. The grade says how the tumor is expected to act. See the glossary, doctors score follow-up MRI scans with a system called BT-RADS. Researchers built AI software to do that scoring and tested it on five hundred nine past scans from one hospital.
What they found. Four hundred ninety-two of the scans could be scored. The software gave the right score for 374 of 492 scans, or 76.0%. The first reports from the clinic were right for 283 of 492 scans, or 57.5%, a gap of 18.5 percentage points. The software did better on some kinds of scores than others.
What this means, and what it doesn't
What it could mean: A tool like this might one day help doctors score scans in a more consistent way. It could act as a second set of eyes on a scan report. Right now it is a research tool only.
What it doesn't mean: This does not mean the software is in use at clinics today. It was tested on scans that were already collected, not during real care. It only helps read scans. It does not treat a tumor, and it is not a cure. It also got many scans wrong. Much more testing is needed first.
Source: PubMed, October 21, 2026 · Read the original
This plain-language summary was written by AI and published automatically after passing our automatic safety checks. How we write.