A computer tool learns to find brain tumors on MRI and name the type
Original title: LIGHT: Learning Image-text Grounding for Hierarchical Tumor Localization and Subtype Classification in Multimodal Brain MRI.
How far along is this research?
- Lab cells
- Animals
- Review
- Tested in people
This was done on cells in a lab, not in people. It is a very early step.
This is an early computer tool tested on stored scans, not something used in real care yet.
The short version
Researchers built a computer tool that looks at brain MRI scans and tries to say where a tumor is and what type it is.
What was studied. Researchers built a tool called LIGHT. It learns from MRI scans paired with the reports doctors write about them. It was trained on 99,813 scan and report pairs, then tested on 11,034 brain tumor MRI cases from their own hospital, another clinic, and public collections.
What they found. The tool picked the correct broad area of the brain 72.1% of the time. For a more exact spot, its top answer was right 70.1% of the time. For naming the tumor type, it was right 76.0%, 83.7%, and 84.6% of the time on three different sets of scans.
What this means, and what it doesn't
What it could mean: Tools like this may one day help doctors read brain MRI scans faster and describe findings in the same words used in radiology reports. It could act as a second set of eyes for the doctor. It would not replace your own doctor or your scan results.
What it doesn't mean: This is not a treatment, and it is not a cure. It is early computer research. The tool was tested on scans that were already collected, not used in real appointments with patients. It also got a fair number of answers wrong. Nothing here changes the care you get today.
Source: PubMed, July 31, 2026 · Read the original
This plain-language summary was written by AI and published automatically after passing our automatic safety checks. How we write.