AI may show which brain scans need contrast dye
Original title: Predicting brain tumour enhancement from non-contrast MRI with artificial intelligence: a multicohort, retrospective, diagnostic accuracy study.
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 scans already taken, not used in real care yet.
The short version
A computer tool may one day cut how often brain scans need a dye shot.
What was studied. Researchers trained AI on brain MRI scans from ten datasets in four countries. The scans came from adults and children with several kinds of brain tumour. The AI had to tell which tumours would light up with dye, using only scans taken without it.
What they found. The AI was tested on 1109 scans it had not seen before. It spotted tumours that light up more often than the doctors did. Eleven expert doctors who read scans did the same task on 564 cases, and they did less well. The AI did worse on children, matching the tumour area in only nine of 20 cases.
What this means, and what it doesn't
What it could mean: Dye is not easy or wanted for everyone. Some people have kidney trouble or an allergy. Children often need many scans over time. A tool like this could help doctors decide when dye is really needed.
What it doesn't mean: This is not a treatment and not a cure. The team looked back at scans that were already taken. The tool was not used in real care with patients waiting. It still made mistakes, and it did worse on children. You cannot get this at a clinic today. The authors say more work with scan doctors is needed first.
Source: PubMed, September 1, 2026 · Read the original
This plain-language summary was written by AI and published automatically after passing our automatic safety checks. How we write.